Preface
Hi folks, it’s good to be back! Last time you saw us was when we toured the Audeze factory, and sat down with Sankar Thiagasamudram to discuss the development of high performance planar magnetic headphones. If you haven’t already seen it, please check it out (https://bit.ly/2yLcHHb). We were thrilled with the response we got from the Headfi community, and we hope you enjoy our next factory tour!
Introduction
After covering the development of cutting edge headphones at Audeze, we brainstormed what to cover next. All of us on the forums participate in the final leg of the music’s journey, its transmission from the source to our ears. The sounds we enjoy from our favorite pair of headphones or IEMs is the result of meticulous craftsmanship and purpose so we, as listeners, can achieve the most pleasure from it. After enjoying countless albums, I find myself asking this question, what is involved during the process of music creation? If audiophiles are so exacting in DAC/amp requirements for our gear, what is it like for the creators and professionals? How can technology be used to bridge the gap between the recording studio and the artist? A small company in Santa Monica may hold the answer to these questions.
Apogee Electronics
Apogee Electronics has a history that spans back to the 1980’s, and found its success in the Apogee 924 and 944 digital filters that were designed for top flight studio recorders such as the Sony PCM 3324.
Apogee’s filters were so popular in fact, it became almost impossible to find a studio without them and equipment distributors would bundle Apogee filters with nearly every recorder sold. Since then, Apogee has innovated its way through the past 35 years with the debut of the first standalone A/D D/A converter in 1991, the first multi channel 24 bit audio interface in 1997, and an extensive array of professional grade music production equipment. Apogee’s product line has evolved over the past few decades to reflect a shift from onsite recording equipment to the increasing market of portable mixing solutions with products such as Duet, Quartet and Apogee’s portable DAC Groove. While hifi consumers may not use Apogee’s professional grade equipment, many of us most likely listen to albums that utilized Apogee’s equipment in some way during the production process. The opportunity to understand how professionals use such equipment can give us a perspective outside world of consumer hifi.
When my father was interviewing industry insiders at CanJam SoCal (https://bit.ly/2KwWmdG), he was introduced to Ambeo through a Sennheiser representative. The Ambeo headset is an interesting piece of technology, as it allows consumers to recording binaural tracks without needing to purchase professional grade equipment which is not only expensive, but is quite bulky. The Ambeo headset works seamlessly with iPhone, and allows the user to capture a convincing 3D recording of any space they reside in. This product is a unique collaboration between Sennheiser and Apogee, with Apogee’s logo adorning the control module of the headset. Curiosity got the better of us, and we made inquiries to Apogee if we could discuss this amongst other topics with a representative. We were lucky enough that Apogee’s Product Evangelist Roger Robindore was able to speak with us. Roger really knows his stuff, as he’s been with Apogee for 20 years and we had a meaningful discussion across a breadth of topics in the audio industry.
Tour & Interview
Apogee’s HQ resides in a nondescript brick face building, with the Apogee logo emblazoned on the front. They offer two doors at the entrance, one for artists (right) and one for visitors/employees (left). Why do artists have a separate entrance? Stay tuned, we will get to this later! We were greeted Roger by reception, and asked him what was the deal with separate entrances. Roger smiled and said “There’s no better place to begin, let me show you Apogee’s crown jewel”. Walking through a separate entrance, we arrive at a nondescript door, which leads into an expansive studio. At the center of the studio is massive mixing console, which I am told is a vintage Rupert Neve Model 8068 console, there are only a few of this specific model in the world. Roger began to explain to use the purposes behind the studio, and the discussion quickly transitioned into a technical one, exploring different aspects of the audio field.
Roger: This studio represents a key aspect of Apogee, our relationship with artists and how their feedback informs our design decisions. One of the most interesting things about Apogee is we produce incredibly high end professional systems that reside in top studios around the world.
We have three divisions of products. First, is our professional products, this is our bread and butter and how we got started in the first place. Our second product category is the semi professional, such as the home studio. Lastly, is our consumer portable, this is reflected in our iOS focused product line, such as Duet, Quartet and others. You mentioned the Ambeo headset, and it is in part result from the movement of consumers and professionals using iOS products as part of their workflow. This whole idea of recording audio on your iPhone has permeated our industry, and we think introducing quality 3D audio recorders for iPhone users can give our customers a new creative canvas.
Kishore: Can you tell us about this mixing console?
Roger: The beating heart of Apogee is this studio Rupert Neve Model 8068. Though it’s Bob’s [Bob Clearmountain] studio, employees have access to it. We all get to share this wonderful gear during jam sessions, for instance we took delivery of some brand new recently, but we also have a selection of vintage mics that impart that give us that unique quality to recordings. While we are working on developing gear, we also come in this studio and put it to the test. I can’t tell you how much fun I’ve had in this studio jamming with the people here. We also have a unique partnership with KCRW who make use of this studio as well.
Kishore: When you develop new gear, how often do you go back to older gear for inspiration? There is some tremendous vintage stuff here, or do you start with a clean sheet design?
Roger: Hahaha that’s an interesting question. The legacy of Apogee is our A/D D/A technology. When Apogee started, the language, technology and science of conversion was an esoteric idea at the time. We had specialized terminologies such as Nyquist rates, convolution, filtering, and others. Whereas now, these ideas are more widespread. You folks understand them and I imagine many Headfi members read it for fun in their spare time.
Let’s go back to the days when the CD first came out. Do you remember the tagline for CD? Perfect sound forever? So technically, CD was supposed to be absolutely perfect. But there’s an old saying ‘In theory, there’s no difference between theory and practice’. In reality however, there’s a Grand Canyon of difference between theory and practice. So the digital audio theory is very sound (pun not intended) and everyone thought they had perfect sound forever with CD, but of course in the ensuing years, people were thinking “there’s something wrong with CD audio”. It took a lot of effort to determine what is wrong with the theoretically perfect process that wasn’t working. In the early days, it turned out to be the filtering. The fact you need to limit your bandwidth before you go into the digital process.
Bharat: What do you mean by bandwidth?
Roger: Let’s start with a quick visual representation of the wagon wheel optical illusion (https://bit.ly/2Ryf1bC). When you spin a wagon wheel past a certain speed, the spokes appear to move backward. This visual ‘trick’ occurs due to a camera’s sampling rate. Film, for instance 24 fps, is 24 individual pictures moving fast enough for our eye to perceive it as a movie. The film rate and the wagon wheel rate are interacting so you get this illusion of the wagon wheel spokes moving backward (illustrated below).
You can also see this effect if you film a helicopter where the rpm of the rotor blades are synchronized perfectly with the cameras capture rate [https://bit.ly/2n2v8Pv]. What does this mean in the for digital audio? Well, you need to limit your frequency range for your audio to half of the sample rate. If you don’t do this, then you introduce aliasing to your audio sample. Aliasing is where you get these false frequencies that get fed back into the audio range.
Let’s take a simple example, let’s take 48kHz audio. If you have a sine wave that has an amplitude higher than 24kHz, any audio at 24kHz will get sampled at different places so when it gets reconstructed it appears as 8kHz. The basic functionality of a digital system is, before you start sampling, you need to limit the frequency response. Now, for bizarre historical reasons the very first sample rate was 44.1 kHz, that means half of that is 22kHz which means you can have no audio beyond 22kHz, but up to 22kHz you want to have full frequency response right? Now, that filtering is crucial, nothing works without that filtering in place. If the filtering isn’t good enough, then the whole theory of digital audio goes out the window.
The way how Apogee got started was providing the audio industry with audio filtering technologies that corrected many long-standing issues. Before Apogee, the way the filters were being implemented, while technically fulfilling the requirements for digital audio, rendered audio so sharply that there were resonances being introduced into the audible high end bands. Apogee was one of the first companies to carefully examine the science of how filters affect audio quality. After determining this as a problem that needed to be solved, Apogee introduced a set of filters that were designed for the professional multi track tape machines of the 80’s. So we were approaching the problem from the production side as opposed from the consumer side. At the time, Sony pushed out most of the studio tape machines along with Mitsubishi and a few others. We made drop in filter modules which addressed those resonance issues among others. These filters became immensely popular, for instance in the United States there were 100 multi track machines, these went for $250,000 at least. Well, 99 of them had Apogee filters, and at a certain point Apogee converters became a stock configuration option when a studio bought these multi track machines. And that was Apogee’s starting point.
Kishore: Speaking of studio equipment, can you tell us more about the Apogee Studio?
Roger: Sure! First and foremost this space is our universal jamming space. Sometimes we have performances here with top artists, and it’s really fun. I mentioned earlier nearly all of our employees play instruments, and this space gives them a chance to jam and have sessions. This really helps bring the workforce together and bridges our passion for music to what we do at Apogee. However, I am most proud of how this studio plays a role in Apogee’s philanthropy. This is a private studio, and sometimes artists will rent it out for private use but when they do, all the proceeds this studio generates go to an orphanage in Mexico, Corazon de Vida.
When we moved into this building, all the other spaces were developed, and this area was one big empty space. So Bob has always had a vision of a live performance space in which great recordings could be made. So a lot of hard work later, we got it done, and on top of creating an excellent recording space we are able to donate proceeds to charity. The success of our studio has exceeded our expectations in every way. The caliber of artists that have recorded here, it’s just incredible.
Kishore: Can I ask who?
Roger: Sure, we had a show with John Legend in here with a piano and a string quartet, just incredible. Also, Queens of the Stone Age, Beck and a lot of folks with KCRW who we partner with. John Mayer rented our studio for a month, it was a blast to work with him. We’ve had such a wide variety of people come in and record here, this studio has just exceeded our wildest dreams.
Realizing we still had much to see on the tour, Roger asked us to walk through the studio and into the main offices. The aroma of toasted bagels and coffee drifted towards as we whizzed by an array of cubicles. We made way towards the back of the building and entered what looked like a workshop. Roger led us to the corner of the workshop area where the scent of warm electronics wafted from a rack with numerous wires probing the innards of Symphony I/O’s. The top panels of Symphony’s were off and we could see inside the intricate workings of Apogee’s flagship systems.
Roger: So this is the Symphony I/O assembly room. We receive the parts, circuit boards, panels, and everything is assembled by hand onsite. Our team goes through a multi stage testing process with many hours of burn in required. The Symphony I/O range (https://bit.ly/2r9osTC) represents a mission critical product and needs to empower artists and their audio professional teams to create the best music possible. Even after all the audiophile testing, we perform a final set of mission critical quality checks because not only is our flagship equipment going to audiophiles, but it also finds itself in top studios around the world and we need to not only meet, but exceed those standards. An audiophile can tinker all day with R2R DAC’s, mono blocks and spend five thousand dollars in cables, but if the recording studio doesn’t have high caliber recording equipment then it’s pointless. We want to deliver the best audio experience, and our equipment allows this for both the artist and the listener.
Kishore: How can studio equipment serve both the needs for studio engineer and audiophile?
Roger: I’m glad you asked! The Symphony I/O is modular by design, you can configure it via different card modules we offer. These I/O cards can configure a Symphony for a large number of I/O, you can configure a Symphony for up to 16 inputs and 16 outputs, giving you 32 total I/O. We have recently released the new Symphony 2X6 SE which began as a passion project of one of our analog engineers. The goal of the 2X6 SE is to offer the absolute best conversion possible, with all the latest components which are carefully implemented to our custom design. What’s interesting is that our Symphony is selling quite well to audiophile customers in China, many of these units you see being burned in here are going to China. We didn’t anticipate this sort of demand, but there it is.
Kishore: Which product would you recommend to an audiophile who’s looking to get studio grade performance?
Roger: I was actually thinking about this haha! Apogee Groove is our portable DAC/headphone amp which is trusted by musicians around the world. If anything, this product is a bit too modest considering the technology in it. It looks unassuming, yet it packs a high power amplifier which drives industry standard reference monitors from companies such as Sennheiser and Beyerdynamic with authority. These include T1’s, HD800’s and HD650’s. Of course, you can use nearly any headphone you like, but we’ve found most music studios utilize these headphones.
What makes the Groove special is the fundamentally different amplifier topology. Nearly every amplifier operates with constant voltage. We have an adapted constant current output driver in Groove. There’s only one other company, Korean I think, that does something like this. Why is this adaptive constant current needed in amplifiers? Well, today there are a litany of headphones for a consumer and professional to choose from, each are designed with different specifications, materials, technologies, and amplifiers that are expected to work perfectly together. However, if your headphone impedance is too close to your headphone output impedance, then you’ll get variations in frequency response. Clearly, this one size fits all approach isn’t good enough for us, so we designed an amplifier topology which can deliver performance that is headphone specific. The end goal is to deliver absolute consistent performance with whatever headphone regardless of the headphone impedance differences. For an audio enthusiast, this translates to a consistent flat frequency response across a variety of headphones. Many music professionals use different headphones, some swear by Sennheiser, others by Sony, and others use Beyerdynamic. We wanted to provide a consistent experience for both the music professional and listener and let people hear music as the artist intended.
After taking a few pictures, we went upstairs and headed into a conference room to continue our discussion.
Kishore: In our haste we forgot to ask about your background, how did you come to work at Apogee?
Roger: Sure! I was a recording engineer since my early twenties. I recorded classical music onto tape, then edited it using a razor blade on to a score I got from the conductor. Then I got into the pop world, I was on the recording team of the last Pink Floyd Live Album, and I also was able to work with Prince when he went to France and recorded there. The technology side drew my interest especially, and I started working for the Paris Apogee distributor. Soon after I met the team here in Santa Monica, and I joined Apogee full time here in 1999. I’ve had different positions within Apogee since then, and it’s been a blast so far.
Kishore: What is it like working here?
Roger: It’s pretty incredible, because nearly all of us are practicing creatives and due to that we are able to have amazing collaboration and create value based products.
Kishore: How do you create a great product at Apogee? What is the criteria?
Roger: When we start with a new product, we need to distill the idea down to its core principle. If one can summarize the product into two or three sentences which communicates the need, purpose and why clearly, then we will begin development. Generally this has served us quite well over the past few decades. In every instance of product development, we need to deliver a product which is far and beyond the current gold standard. The Duet is a great example, the market for this product was quite crowded and competition was fierce. But Duet offered such a compelling value for artists, the market has accepted the Duet as the de-facto gold standard. Duet has such a simple interface, very durable, and offers quality at such a high standard that large studios and artists use this for final recordings.
For instance, we make a microphone for iPhone and many of our customers use it for recording on the go. Now, one of our customers is a voice for Fox broadcasting. When he’s onsite at Fox studios they can offer him multi million dollar recording suites and so on. However, when he’s vacationing in another country and needs to do call backs, he uses one of our Apogee portable microphones and uses it with his iPhone. He can email these recordings to the studio, and they use them the next day for broadcasting. We loved how our customer used our portable product and used it for endpoint solutions such as final voice overs for broadcasting.
Bharat: That’s quite impressive! We can see Apogee has made significant contributions to iOS/Mac. But let’s turn back the clock. Earlier you mentioned Apogee created filters for studio equipment in the 1980’s. What did Apogee do after creating filters?
Roger: Ah yes, thank you. After we became successful in the area of filters, we started to address the idea of clocking.
Bharat: Can you clarify the concept of clocking?
Roger: Sure. The analogy I like to use for clocking is, imagine you have a wave form. And you overlay a piece of graph paper on the wave form. Using the graph paper, I can then assign coordinates to several points. If you then have a blank piece of graph paper, in theory you should be able to reconstruct the waveform perfectly using just the coordinates I mapped earlier. This is a good analogy for digital audio and represents reconstruction process I mentioned earlier.
Kishore: This is true only if the squares on the graph paper are small enough and allow enough resolution correct? Does that mean smaller graph paper squares mean better reconstruction?
Roger: Ah ha! I’ll get to that in a minute. But first, I need to cover the concept of jitter. Imagine if you have faulty graph paper that was unevenly spaced. You can deduce if you plotted the coordinates of the sine wave I measured earlier on to the uneven graph paper, the wave form wouldn’t be accurately produced. This is exactly what jitter is. You see, digital audio makes assumption, chiefly that the sample is incredibly regular, that is the interval between each sample is precisely the same. As soon as the spacing is not the same, you get these distortions of your wave form, and this is jitter.
Now we have this understanding of jitter, we can cover the concept of clock in digital audio. Moving to the analogue side, we have this clock that generates the coordinates on the graph paper. On the digital side, we will have a clock that reassembles the coordinates and the tiniest variations happen to be audible, which can cause music to sound unnatural or ‘artificial’. By the way, this happens in both the A/D and D/A side.
Bharat: So you have two clocks that work together?
Roger: Correct, but we’ll have to get real deep into this. So, with the general consumer and the folks on headfi who use a digital source, they all use a D/A converter and usually they can clock the incoming digital signal. But in every DAC you can find some kind of jitter reduction. In a consumer audio D/A, for example you receive incoming digital audio from a CD platter without conversion, directly from the pre out. So the consumer D/A will look at the incoming digital signal from the CD and try to clock off of that but you will a buffer that will remove the jitter from the digital stream and use that clock to ‘clock out’ the data through the converter. But in a professional rig (or high end hifi) where you have both the A/D and D/A in the chain, you have one clock that acts as the master clock that synchronizes both A/D and D/A processes. Keep in mind Apogee was one of the first companies that researched clocking and how in meaningfully affected audio. Many people in the audio realm today understand these concepts of clocking and oversampling, but back in the day these were esoteric issues. Apogee was one of the very first companies to not only discover them, but create products which addressed these issues. The crux of Apogee brand is ‘digital audio, practical issues’, and we try to bring the real world practice of digital audio closer to the theory.
Kishore: Can you touch on some of the significant misunderstandings in the world of digital audio?
Roger: Yes, of course. Earlier, I used graph paper as an analogy for oversampling, and you asked me if using smaller squares would yield a better result for music reproduction. Now, this statement leads to one of the biggest misunderstandings of digital audio. First we need to master the basics, we need to define sample. A sample is every digital sample represents a time and a voltage. So if you have a sine wave, each point on the curve represents a different time and voltage. You can then use those numbers, to reconstruct the analogue signal. Now, this would lead you to believe that the ‘finer’ the graph paper (a higher sample rate) the more accurately you will be able to produce the analogue wave. This is a reasonable expectation. This is also 100% wrong.
Kishore: What? That sounds counterintuitive.
Roger: Haha it does doesn’t it? For instance, if you use Illustrator or Photoshop there is an arc tool. This tool allows you to draw a specific arc between two points. Keep this visual in mind. Now, way digital audio works is when you apply filtering at the A/D stage that the filter ensures any frequency beyond a predetermined frequency point isn’t introduced into the system. The D/A stage filter, you have an identical filter, this is the reconstruction filter. Essentially a filter limits the speed at which a waveform can change. A filter acts like the arc tool in Illustrator. The arc tool forces a predefined curved, which determines how the points can connect, and this is how reconstruction filters work.
If your frequency is under half the sample rate, then it will perfectly connect those two points (samples) in only one way. Let’s say I have a 19kHz sine wave, but if my sample rate is 48kHz there isn’t going to be many samples to define the curve. The magic of digital audio is that if you have two discrete sample points, and you use a digital filter, then there is only one way those two points can ever be connected, and this will give us perfect reconstruction even with very few points. It’s the digital filter which gives us the predetermined path to perfect sound wave reproducibility.
This is one area where digital audio theory works perfectly. If you’re sampling at 48kHz, and nothing above 48kHz got into your digital audio source, and you feed it through a digital filter that’s exactly the same, then the theory directly translates into reality. It’s quite rare to have a direct translation of theory and practice.
In summary, contrary to popular belief, oversampling does not mean better quality. Oversampling means oversampling. There’s no need to put more samples. So what does a higher sampling rate give you? It gives you the ability to open up your filter. So now instead of having a filter at 24kHz, I can have a filter at 48 kHz. With a higher sampling rate, you can achieve higher bandwidth.
Kishore: Hold on. How do you determine a good sampling rate? Doesn’t human hearing top out at 20kHz?
Roger: That’s exactly the question. Does sampling above 20kHz provide anything useful? This is a topic you can spend days, weeks even years going around in circles. What we’ve found is there are some processes that do have benefits at 96 kHz, but this is an impossibly small amount. The vast majority of professionals work in 48kHz. Some, due to the debate of 96kHz vs 48 kHz, work in 96 kHz but I don’t think you’ll find many professionals at all recording at higher than 96 kHz. Anything higher than 96 kHz isn’t really useful for professionals.
After listening critically for many years in this industry, I’m somewhat thrilled I can hear up to 15kHz considering my age haha! The thing is, I’m sure you can debate on Headfi all day about the ‘usefulness’ of recording frequencies above 20kHz, but this is not high on our priority list at Apogee. We focus on what value can we bring to the professional. There are so many other avenues to work on than being concerned with higher kHz. That being said, most if not all of our products do 192 kHz because the market asked for it.
Kishore: I’ve noticed your DAC’s don’t support 384kHz, let alone DSD. But you’re saying even supporting 192kHz is overkill.
Roger: We just don’t think it’s adds enough value for our customers. We’ve had this exhausting conversation so many times with people who ask us ‘why don’t you support 784 kHz or DSD, my arglebargle from China does this for a fraction of the price’. And we say, ok cool. We have a hard time keeping our faces straight when we see folks talk about how songs with higher bitrates (eg 96 kHz) reveal new levels of detail. Also, objectively we’ve done the measurements between higher sample rates, and everything measures identical across the board. Believe me, I would’ve loved to see some differences, anything that can enable the artist to produce better music, and allow the listener to extract more value from the songs is the core of our values. However, all these years later we’ve learned to invest in proven technologies, and these new formats like SACD, DSD, MQA will come and go with time. Just look at how MQA is turning out. Let’s look at DSD, it’s an old technology that got repackaged. Do you remember when we discussed integrated chips? Sony came out with 1 bit oversampling for a generation of chips that was about 20 years ago. Then, they took that internal technology and made it into a public format called DSD. And for the longest time, there was no real acknowledgement that it was old technology. And now we see are multi-bit versions of DSD because there it is generally accepted that 1 bit DSD offers less performance than multi bit DSD, hence nearly all DSD converter chips actually convert 1 bit DSD to multi bit DSD (more info https://bit.ly/2zMqP2A).
Kishore: Can you expand upon what this ‘bit’ means, and its context in digital audio?
Roger: Of course. Let’s take a 48 kHz 24 bit file. 24 bit means you should have several hundred thousand different voltage levels. You have 24 ‘levels’ to define a voltage level between 0 dBFS and -144 dBFS if you’re in the digital domain. I think you have around 600,000 individual bit levels for a 24 bit file (recording format), and for a 16 bit file (CD audio) you have about 16,000 bit levels. The total amount of levels are your dynamic range. In the signal path, you are asking a computer or converter chip to look at the analogue signal 48,000 times per second (assuming a 48kHz file), look at the incoming voltage and finally associate that voltage with one of 600,000 different levels. For a 96kHz file, the converter has to do this 96,000 times per second and so on. The way DSD works is that the converter’s process examines the current sample and asks, is the level greater or lesser than the previous sample, then it adjusts as required. This is done by increments or decrements of one bit. DSD does this incredibly fast, and this is in the MHz range. Essentially, DSD performs an incredibly basic sampling operation, but because it’s simple, it can do this at an incredible speed. So DSD was seen as the next step to improve the A/D and D/A converters. Then they came to the realization that simply transitioning one level up or down wasn’t enough. Now they have more bits to increase the resolution to better characterize the next sample during the sampling process. So that’s what multi-bit DSD is. The tradeoff they are making is between the amazing number of digital levels a device has to detect for non DSD files, versus doing the sampling much simpler but far faster.
DSD multi-bit is happening in every DAC on the market, it happens in the DAC chips we use for instance. At the end of our conversion process however, we use a decimation filter. This takes the output of all that and packages it into a 96 kHz 24 bit format in the digital domain. This is what we record and playback in.
When it’s happening in the converter and assigning the sample a digital value, this is the process that can get complicated. With all these novelty formats, they are taking what’s already occurring in the chip, and externalizing it. The chipmakers are pretty much a decade in front of these formats coming to market, and we at Apogee don’t think this is a coincidence. For example DSD 1 bit was already happening in the chips for about 10 years before Sony came out and made DSD a public format. This same thing is true with DSD multi bit, multi-bit DSD has been in chips for about a decade and in the last five years it has become a public format. By the way, the multi-bit DSD process occurs in our products and in nearly every DAC on the market.
Kishore: Is there any added value to the consumer, by making DSD a public format?
Roger: No. It simply doesn’t haha! If DSD had any value we would have included in our products immediately.
Kishore: Apogee is a unique electronics company, considering you have a full recording studio and have a relationship with the artists that use your products. Do you work with artists when developing new products, if so how does such a collaboration work?
Roger: Oh absolutely! So first all our products are made in the US. Most of our products are manufactured toward the San Diego area. The best way to deal with manufacturing/ calibration issues for instance, is to be able to drive over to the facility with our testing equipment and have it addressed same day. Our flagship product, the Symphony I/O is made in house, you’ve seen the board assembly downstairs, and we have nearly a dozen units undergoing burn in right now before we conduct few more tests and it’s sent to the shipping department.
Regarding artist collaboration, we work with them iteratively and throughout the prototype phase to ensure our product adds significant value to an artists workflow. Let’s look at one of our more popular products, the Duet. This was one of the very first professional interfaces that can be used in big sessions. You can connect microphones, speakers and headphones. Granted it doesn’t have the performance of our flagship series products, but Duet brings unmatched ease of use and portability. We provide prototypes to a number of artists, and they provide feedback on how not just sonic quality, but on ease of use as well. A big part of my job is to talk with customers and artists is to take their feedback and apply it to make our products better.
Kishore: Regarding artists, what's their perspective on these new high bitrate formats? Do they care about it?
Roger: I’ll put it this way, extremely few professionals deal with bitrates above 92kHz. Most professionals are working with 48 kHz.
Kishore: In the hifi world, it’s not uncommon to see $10k, $20k, and even $50k DAC’s. Yet Apogee’s top range studio DAC is relatively affordable, below $5k. Do you feel these other DAC’s are completely unnecessary, or is Apogee working on a premiere DAC to compete with these other super high priced flagships on the market?
Roger: Oh, that is a dangerous question my friend. I would say that professionals are less susceptible to hype. Everything we do in our products has an incredibly practical, but provable reason behind it. When we design and create products we need to see or measure the improvement objectively but we also need to subjectively experience the improvement. Keep in mind, it’s not one or the other, many people get lost in this. It’s a careful combination of the two. Specifications aren’t the only thing.
The trouble is, many products start with older assumptions. For example, it used to be that one had to literally build one’s own converter process out of discrete components. The other way of doing it is using converter chips, which is an integrated chip that has all the conversion performed on the chip itself (such as ESS Sabre or AKM chips). So you may have heard of discrete analogue versus integrated circuit (IC) chips in the analogue domain. When IC chips were first developed, there was some truth to that, in order to get better performance, you had to do everything with discrete parts and an integrated chip resulted in a lower quality solution. However, this was decades ago, and modern chips have astounding quality, the performance far exceeds what you will find in a discrete converter. Now, you will find people that will sell discrete converter which takes this process that is usually encapsulated in a DAC chip but then separates out the processes which involves a custom design for the converter.
Kishore: Are you referring to the FPGA architecture?
Roger: No, I’m referring to the converter chip itself. So, on a DAC you have a digital in, clock inputs, and analogue outputs. The quality of these converter chips is really astounding. Now, sure you can make a good sounding product with a good chip, but not many folks are interested in merely ‘good’ performance, everyone wants the best performance possible. So, to take performance further, it comes down to providing a perfect environment for those inputs and outputs. For example, a clock input that is ultra low jitter and taking those analogue outputs and providing the chip with an electrical environment that facilitates best performance. This is the concept of ‘DAC implementation’ we see being discussed on forums, and it’s this implementation that becomes an art. What consumers need to know is that it’s not all about the type of chip a manufacturer uses, yes that’s a part of it, but it really comes down to the implementation of the chip in the circuit design. I’m sure you’ve seen on some HeadFi threads where someone states ‘This DAC doesn’t use xyz-123 chip, therefore it’s out of spec and a waste of money’. This blanket statement is simply not true. Some manufactures have made this into a discrete process, and it’s based on an older assumption that is ‘discrete will always be better than integrated’. And this is a false assumption.
Kishore: You may have upset some folks reading this.
Roger: Haha…well, so be it. We find that professionals more concerned with getting mission critical work done, rather than experiment with various things for, at best, debatable gains in performance. At Apogee it’s always been about giving the professional the best performance possible at a reasonable cost.
Kishore: Moving to the creative aspect, did you develop your iOS oriented products in anticipation of the ‘bedroom artist’ movement?
Roger: Yes, as you noticed our iOS products and portable professional tools sort of overlap. Actually, the beginning of our iOS product line is an interesting story, and is a big part of our long-standing relationship with Apple Inc. Many years ago, artists who wanted to use their iOS device needed to use products that interfaced with Apple’s headphone jack but such products were terrible quality. Many artists were asking us for a quality product that would allow them to record with iPhones/iPads and Macs. So, while we developed the hardware product called Jam, we reached out to Apple and collaborated with them to create a digital audio protocol for iOS. This protocol is open, and we pushed the industry forward as people weren’t restricted to the headphone jack, and could use the 30 pin port instead. This served as the foundation for our push into the mobile front, and we are proud of it.
Bharat: What’s the nature of your relationship with Apple?
Roger: Haha, everyone asks us that. Folks ask if Apple owns us, or if there are foot high piles of contracts involved. The truth is, it’s an informal relationship. We’ve known their audio engineers for decades, we provide their core audio team with insider knowledge of audio development and recording, and of course they provide us with their massive customer base as they have our products in stock at the Apple Store.
Bharat: Can you tell us about the Ambeo Headset, what did it involve and what was it like working with Sennheiser?
Roger: For us, it was incredible to be able to partner with this legendary audio company, guided by German engineering. I found it interesting that we were able to provide something to Sennheiser, and it’s the same thing with Apple. These companies, Apple especially, are massive and what can our humble outfit offer to them? It turns out there are a couple of things. When working with Sennheiser, it became quickly apparent we can offer them our iPhone expertise as Apogee has a great working relationship with Apple, but also many of our products are designed to work with Apple hardware. Also being a small company, we can move incredibly quickly and turn around a product fairly quickly. And it turns out we can ensure high quality conversion of the headset audio, all while within the confines of the iPhone. Working within Apple’s limitations can be challenging, any developer will tell you this, and when designing a DAC for this application, we always need to operate within tolerances such as the low current load and so on. We were able to provide Sennheiser connectivity with the iPhone, the app, and the conversion technology. We’ve also become adept at creating excellent preamps, and in Ambeo, you have two microphones, and we developed the pre amps for those as well. Throughout this product, there were many areas where we could bring our expertise. Sennheiser were able to bring their earbud and microphone technologies to the table. The reason why we embarked on this project was because we were enchanted by this possibility to create absolutely convincing audio. The binaural audio experience is captivating, and a consumer can experience memories via recorded video with binaural in a whole different way. The Ambeo is dead simple, you just plug it in and begin using the Camera app to record video. We wanted to take binaural technology and put it on the most popular device on the planet. Being able to stun the world with a whole new audio experience, that doesn’t require high spec equipment to experience. All you need is some headphones, and you can have the binaural experience.
Bharat: Is there a way to take a stereo source and make it binaural?
Roger: Facebook has explored this area and developed host of ProTools plugins, one of which is a binaural plug in which spatializes even mono sources. It’s not bad per se, but it’s nothing like the real thing.
Kishore: You mentioned consumers are the target for Ambeo, but how have music professionals reacted to this? Are they using it?
Roger: Some artists have expressed interest in this technology, but there isn’t any specific products or artist creations that utilize Ambeo yet. This is a new tool for them, and figuring out how this can work with their creative vision can take time. Sennheiser is dabbling in this area, and are pushing into the virtual reality space, so we may see something come from that in the near future.
Bharat: Going back to the Sennheiser collaboration, did you have Sennheiser engineers flying in from Germany, or did some folks from Apogee go to Sennheiser’s HQ?
Roger: Actually we have a Sennheiser project manager who’s almost an Apogee employee. Here is here quite a lot, and he is much beloved by the staff here. Betty actually got to know the Sennheiser brothers who currently run the company. We had some high level conversations with these two young brothers who run Sennheiser, and it’s incredible to relate each company’s vision for the future of audio as Apogee and Sennheiser bring complementary skillsets and perspectives to the table. To be honest, Apogee acquired the parts below the earbuds and developed the core technology behind the Ambeo. We manufactured the enclosure, circuit board, the iOS certification but we relied on Sennheiser for the earpieces in the Ambeo.
Kishore: Looking closer at your product line, I am surprised to learn Apogee doesn’t offer DJ products as this is an area Apogee can leverage their expertise and knowledge in. Does Apogee have plans to offer products in the DJ segment?
Roger: Hmm that is true. I would say Duet is a toe in the water for that area, we do have many customers that use Duet for sound mixing in clubs and other such venues. This area is definitely interesting to us, but we haven’t found the right product yet but we definitely have some things on the horizon which can apply more directly.
Kishore: What does the future look like for Apogee?
Roger: We will continue to refine our flagship professional series products such as Symphony I/O, as you saw we have a rack of those being burned in downstairs that are being shipped out at the end of the week. However we are always on the look out for new product categories, for instance we transitioned from being a professional only solutions company to offering consumer, prosumer and professional portable solutions which work with iOS and Mac. We actually have an audio recording program called MetaRecorder (https://bit.ly/2Kts6AL). MetaRecorder is a professional grade field recording app for iOS. At AES, we just announced MetaRecorder works with a UltraSyncBlue system to record and synchronize timecode onto your iPhone (https:// bit.ly/2DTLEOe). This allows people to use iPhones to make movies, which Apple is very eager to push. For instance, if you are a film student this can be very useful, as many students have iPhones and you can shoot multiple angles with multi camera synchronized timecode without exceeding your budget. These kinds of features used to be pro only, and having this available on iPhone is unheard of.
Kishore and Bharat: Thank you so much for your time, we greatly appreciate it.
Roger: No worries, I didn’t think we’d talk about all these topics and in such depth, but I actually enjoyed it. This was quite fun.
*All images are taken from Wikipedia, Apogee or Sennheiser unless otherwise stated.
Hi folks, it’s good to be back! Last time you saw us was when we toured the Audeze factory, and sat down with Sankar Thiagasamudram to discuss the development of high performance planar magnetic headphones. If you haven’t already seen it, please check it out (https://bit.ly/2yLcHHb). We were thrilled with the response we got from the Headfi community, and we hope you enjoy our next factory tour!
Introduction
After covering the development of cutting edge headphones at Audeze, we brainstormed what to cover next. All of us on the forums participate in the final leg of the music’s journey, its transmission from the source to our ears. The sounds we enjoy from our favorite pair of headphones or IEMs is the result of meticulous craftsmanship and purpose so we, as listeners, can achieve the most pleasure from it. After enjoying countless albums, I find myself asking this question, what is involved during the process of music creation? If audiophiles are so exacting in DAC/amp requirements for our gear, what is it like for the creators and professionals? How can technology be used to bridge the gap between the recording studio and the artist? A small company in Santa Monica may hold the answer to these questions.
Apogee Electronics
Apogee Electronics has a history that spans back to the 1980’s, and found its success in the Apogee 924 and 944 digital filters that were designed for top flight studio recorders such as the Sony PCM 3324.
Apogee’s filters were so popular in fact, it became almost impossible to find a studio without them and equipment distributors would bundle Apogee filters with nearly every recorder sold. Since then, Apogee has innovated its way through the past 35 years with the debut of the first standalone A/D D/A converter in 1991, the first multi channel 24 bit audio interface in 1997, and an extensive array of professional grade music production equipment. Apogee’s product line has evolved over the past few decades to reflect a shift from onsite recording equipment to the increasing market of portable mixing solutions with products such as Duet, Quartet and Apogee’s portable DAC Groove. While hifi consumers may not use Apogee’s professional grade equipment, many of us most likely listen to albums that utilized Apogee’s equipment in some way during the production process. The opportunity to understand how professionals use such equipment can give us a perspective outside world of consumer hifi.
When my father was interviewing industry insiders at CanJam SoCal (https://bit.ly/2KwWmdG), he was introduced to Ambeo through a Sennheiser representative. The Ambeo headset is an interesting piece of technology, as it allows consumers to recording binaural tracks without needing to purchase professional grade equipment which is not only expensive, but is quite bulky. The Ambeo headset works seamlessly with iPhone, and allows the user to capture a convincing 3D recording of any space they reside in. This product is a unique collaboration between Sennheiser and Apogee, with Apogee’s logo adorning the control module of the headset. Curiosity got the better of us, and we made inquiries to Apogee if we could discuss this amongst other topics with a representative. We were lucky enough that Apogee’s Product Evangelist Roger Robindore was able to speak with us. Roger really knows his stuff, as he’s been with Apogee for 20 years and we had a meaningful discussion across a breadth of topics in the audio industry.
Tour & Interview
Apogee’s HQ resides in a nondescript brick face building, with the Apogee logo emblazoned on the front. They offer two doors at the entrance, one for artists (right) and one for visitors/employees (left). Why do artists have a separate entrance? Stay tuned, we will get to this later! We were greeted Roger by reception, and asked him what was the deal with separate entrances. Roger smiled and said “There’s no better place to begin, let me show you Apogee’s crown jewel”. Walking through a separate entrance, we arrive at a nondescript door, which leads into an expansive studio. At the center of the studio is massive mixing console, which I am told is a vintage Rupert Neve Model 8068 console, there are only a few of this specific model in the world. Roger began to explain to use the purposes behind the studio, and the discussion quickly transitioned into a technical one, exploring different aspects of the audio field.
Roger: This studio represents a key aspect of Apogee, our relationship with artists and how their feedback informs our design decisions. One of the most interesting things about Apogee is we produce incredibly high end professional systems that reside in top studios around the world.
We have three divisions of products. First, is our professional products, this is our bread and butter and how we got started in the first place. Our second product category is the semi professional, such as the home studio. Lastly, is our consumer portable, this is reflected in our iOS focused product line, such as Duet, Quartet and others. You mentioned the Ambeo headset, and it is in part result from the movement of consumers and professionals using iOS products as part of their workflow. This whole idea of recording audio on your iPhone has permeated our industry, and we think introducing quality 3D audio recorders for iPhone users can give our customers a new creative canvas.
Kishore: Can you tell us about this mixing console?
Roger: The beating heart of Apogee is this studio Rupert Neve Model 8068. Though it’s Bob’s [Bob Clearmountain] studio, employees have access to it. We all get to share this wonderful gear during jam sessions, for instance we took delivery of some brand new recently, but we also have a selection of vintage mics that impart that give us that unique quality to recordings. While we are working on developing gear, we also come in this studio and put it to the test. I can’t tell you how much fun I’ve had in this studio jamming with the people here. We also have a unique partnership with KCRW who make use of this studio as well.
Kishore: When you develop new gear, how often do you go back to older gear for inspiration? There is some tremendous vintage stuff here, or do you start with a clean sheet design?
Roger: Hahaha that’s an interesting question. The legacy of Apogee is our A/D D/A technology. When Apogee started, the language, technology and science of conversion was an esoteric idea at the time. We had specialized terminologies such as Nyquist rates, convolution, filtering, and others. Whereas now, these ideas are more widespread. You folks understand them and I imagine many Headfi members read it for fun in their spare time.
Let’s go back to the days when the CD first came out. Do you remember the tagline for CD? Perfect sound forever? So technically, CD was supposed to be absolutely perfect. But there’s an old saying ‘In theory, there’s no difference between theory and practice’. In reality however, there’s a Grand Canyon of difference between theory and practice. So the digital audio theory is very sound (pun not intended) and everyone thought they had perfect sound forever with CD, but of course in the ensuing years, people were thinking “there’s something wrong with CD audio”. It took a lot of effort to determine what is wrong with the theoretically perfect process that wasn’t working. In the early days, it turned out to be the filtering. The fact you need to limit your bandwidth before you go into the digital process.
Bharat: What do you mean by bandwidth?
Roger: Let’s start with a quick visual representation of the wagon wheel optical illusion (https://bit.ly/2Ryf1bC). When you spin a wagon wheel past a certain speed, the spokes appear to move backward. This visual ‘trick’ occurs due to a camera’s sampling rate. Film, for instance 24 fps, is 24 individual pictures moving fast enough for our eye to perceive it as a movie. The film rate and the wagon wheel rate are interacting so you get this illusion of the wagon wheel spokes moving backward (illustrated below).
You can also see this effect if you film a helicopter where the rpm of the rotor blades are synchronized perfectly with the cameras capture rate [https://bit.ly/2n2v8Pv]. What does this mean in the for digital audio? Well, you need to limit your frequency range for your audio to half of the sample rate. If you don’t do this, then you introduce aliasing to your audio sample. Aliasing is where you get these false frequencies that get fed back into the audio range.
Let’s take a simple example, let’s take 48kHz audio. If you have a sine wave that has an amplitude higher than 24kHz, any audio at 24kHz will get sampled at different places so when it gets reconstructed it appears as 8kHz. The basic functionality of a digital system is, before you start sampling, you need to limit the frequency response. Now, for bizarre historical reasons the very first sample rate was 44.1 kHz, that means half of that is 22kHz which means you can have no audio beyond 22kHz, but up to 22kHz you want to have full frequency response right? Now, that filtering is crucial, nothing works without that filtering in place. If the filtering isn’t good enough, then the whole theory of digital audio goes out the window.
The way how Apogee got started was providing the audio industry with audio filtering technologies that corrected many long-standing issues. Before Apogee, the way the filters were being implemented, while technically fulfilling the requirements for digital audio, rendered audio so sharply that there were resonances being introduced into the audible high end bands. Apogee was one of the first companies to carefully examine the science of how filters affect audio quality. After determining this as a problem that needed to be solved, Apogee introduced a set of filters that were designed for the professional multi track tape machines of the 80’s. So we were approaching the problem from the production side as opposed from the consumer side. At the time, Sony pushed out most of the studio tape machines along with Mitsubishi and a few others. We made drop in filter modules which addressed those resonance issues among others. These filters became immensely popular, for instance in the United States there were 100 multi track machines, these went for $250,000 at least. Well, 99 of them had Apogee filters, and at a certain point Apogee converters became a stock configuration option when a studio bought these multi track machines. And that was Apogee’s starting point.
Kishore: Speaking of studio equipment, can you tell us more about the Apogee Studio?
Roger: Sure! First and foremost this space is our universal jamming space. Sometimes we have performances here with top artists, and it’s really fun. I mentioned earlier nearly all of our employees play instruments, and this space gives them a chance to jam and have sessions. This really helps bring the workforce together and bridges our passion for music to what we do at Apogee. However, I am most proud of how this studio plays a role in Apogee’s philanthropy. This is a private studio, and sometimes artists will rent it out for private use but when they do, all the proceeds this studio generates go to an orphanage in Mexico, Corazon de Vida.
When we moved into this building, all the other spaces were developed, and this area was one big empty space. So Bob has always had a vision of a live performance space in which great recordings could be made. So a lot of hard work later, we got it done, and on top of creating an excellent recording space we are able to donate proceeds to charity. The success of our studio has exceeded our expectations in every way. The caliber of artists that have recorded here, it’s just incredible.
Kishore: Can I ask who?
Roger: Sure, we had a show with John Legend in here with a piano and a string quartet, just incredible. Also, Queens of the Stone Age, Beck and a lot of folks with KCRW who we partner with. John Mayer rented our studio for a month, it was a blast to work with him. We’ve had such a wide variety of people come in and record here, this studio has just exceeded our wildest dreams.
Realizing we still had much to see on the tour, Roger asked us to walk through the studio and into the main offices. The aroma of toasted bagels and coffee drifted towards as we whizzed by an array of cubicles. We made way towards the back of the building and entered what looked like a workshop. Roger led us to the corner of the workshop area where the scent of warm electronics wafted from a rack with numerous wires probing the innards of Symphony I/O’s. The top panels of Symphony’s were off and we could see inside the intricate workings of Apogee’s flagship systems.
Roger: So this is the Symphony I/O assembly room. We receive the parts, circuit boards, panels, and everything is assembled by hand onsite. Our team goes through a multi stage testing process with many hours of burn in required. The Symphony I/O range (https://bit.ly/2r9osTC) represents a mission critical product and needs to empower artists and their audio professional teams to create the best music possible. Even after all the audiophile testing, we perform a final set of mission critical quality checks because not only is our flagship equipment going to audiophiles, but it also finds itself in top studios around the world and we need to not only meet, but exceed those standards. An audiophile can tinker all day with R2R DAC’s, mono blocks and spend five thousand dollars in cables, but if the recording studio doesn’t have high caliber recording equipment then it’s pointless. We want to deliver the best audio experience, and our equipment allows this for both the artist and the listener.
Kishore: How can studio equipment serve both the needs for studio engineer and audiophile?
Roger: I’m glad you asked! The Symphony I/O is modular by design, you can configure it via different card modules we offer. These I/O cards can configure a Symphony for a large number of I/O, you can configure a Symphony for up to 16 inputs and 16 outputs, giving you 32 total I/O. We have recently released the new Symphony 2X6 SE which began as a passion project of one of our analog engineers. The goal of the 2X6 SE is to offer the absolute best conversion possible, with all the latest components which are carefully implemented to our custom design. What’s interesting is that our Symphony is selling quite well to audiophile customers in China, many of these units you see being burned in here are going to China. We didn’t anticipate this sort of demand, but there it is.
Kishore: Which product would you recommend to an audiophile who’s looking to get studio grade performance?
Roger: I was actually thinking about this haha! Apogee Groove is our portable DAC/headphone amp which is trusted by musicians around the world. If anything, this product is a bit too modest considering the technology in it. It looks unassuming, yet it packs a high power amplifier which drives industry standard reference monitors from companies such as Sennheiser and Beyerdynamic with authority. These include T1’s, HD800’s and HD650’s. Of course, you can use nearly any headphone you like, but we’ve found most music studios utilize these headphones.
What makes the Groove special is the fundamentally different amplifier topology. Nearly every amplifier operates with constant voltage. We have an adapted constant current output driver in Groove. There’s only one other company, Korean I think, that does something like this. Why is this adaptive constant current needed in amplifiers? Well, today there are a litany of headphones for a consumer and professional to choose from, each are designed with different specifications, materials, technologies, and amplifiers that are expected to work perfectly together. However, if your headphone impedance is too close to your headphone output impedance, then you’ll get variations in frequency response. Clearly, this one size fits all approach isn’t good enough for us, so we designed an amplifier topology which can deliver performance that is headphone specific. The end goal is to deliver absolute consistent performance with whatever headphone regardless of the headphone impedance differences. For an audio enthusiast, this translates to a consistent flat frequency response across a variety of headphones. Many music professionals use different headphones, some swear by Sennheiser, others by Sony, and others use Beyerdynamic. We wanted to provide a consistent experience for both the music professional and listener and let people hear music as the artist intended.
After taking a few pictures, we went upstairs and headed into a conference room to continue our discussion.
Kishore: In our haste we forgot to ask about your background, how did you come to work at Apogee?
Roger: Sure! I was a recording engineer since my early twenties. I recorded classical music onto tape, then edited it using a razor blade on to a score I got from the conductor. Then I got into the pop world, I was on the recording team of the last Pink Floyd Live Album, and I also was able to work with Prince when he went to France and recorded there. The technology side drew my interest especially, and I started working for the Paris Apogee distributor. Soon after I met the team here in Santa Monica, and I joined Apogee full time here in 1999. I’ve had different positions within Apogee since then, and it’s been a blast so far.
Kishore: What is it like working here?
Roger: It’s pretty incredible, because nearly all of us are practicing creatives and due to that we are able to have amazing collaboration and create value based products.
Kishore: How do you create a great product at Apogee? What is the criteria?
Roger: When we start with a new product, we need to distill the idea down to its core principle. If one can summarize the product into two or three sentences which communicates the need, purpose and why clearly, then we will begin development. Generally this has served us quite well over the past few decades. In every instance of product development, we need to deliver a product which is far and beyond the current gold standard. The Duet is a great example, the market for this product was quite crowded and competition was fierce. But Duet offered such a compelling value for artists, the market has accepted the Duet as the de-facto gold standard. Duet has such a simple interface, very durable, and offers quality at such a high standard that large studios and artists use this for final recordings.
For instance, we make a microphone for iPhone and many of our customers use it for recording on the go. Now, one of our customers is a voice for Fox broadcasting. When he’s onsite at Fox studios they can offer him multi million dollar recording suites and so on. However, when he’s vacationing in another country and needs to do call backs, he uses one of our Apogee portable microphones and uses it with his iPhone. He can email these recordings to the studio, and they use them the next day for broadcasting. We loved how our customer used our portable product and used it for endpoint solutions such as final voice overs for broadcasting.
Bharat: That’s quite impressive! We can see Apogee has made significant contributions to iOS/Mac. But let’s turn back the clock. Earlier you mentioned Apogee created filters for studio equipment in the 1980’s. What did Apogee do after creating filters?
Roger: Ah yes, thank you. After we became successful in the area of filters, we started to address the idea of clocking.
Bharat: Can you clarify the concept of clocking?
Roger: Sure. The analogy I like to use for clocking is, imagine you have a wave form. And you overlay a piece of graph paper on the wave form. Using the graph paper, I can then assign coordinates to several points. If you then have a blank piece of graph paper, in theory you should be able to reconstruct the waveform perfectly using just the coordinates I mapped earlier. This is a good analogy for digital audio and represents reconstruction process I mentioned earlier.
Kishore: This is true only if the squares on the graph paper are small enough and allow enough resolution correct? Does that mean smaller graph paper squares mean better reconstruction?
Roger: Ah ha! I’ll get to that in a minute. But first, I need to cover the concept of jitter. Imagine if you have faulty graph paper that was unevenly spaced. You can deduce if you plotted the coordinates of the sine wave I measured earlier on to the uneven graph paper, the wave form wouldn’t be accurately produced. This is exactly what jitter is. You see, digital audio makes assumption, chiefly that the sample is incredibly regular, that is the interval between each sample is precisely the same. As soon as the spacing is not the same, you get these distortions of your wave form, and this is jitter.
Now we have this understanding of jitter, we can cover the concept of clock in digital audio. Moving to the analogue side, we have this clock that generates the coordinates on the graph paper. On the digital side, we will have a clock that reassembles the coordinates and the tiniest variations happen to be audible, which can cause music to sound unnatural or ‘artificial’. By the way, this happens in both the A/D and D/A side.
Bharat: So you have two clocks that work together?
Roger: Correct, but we’ll have to get real deep into this. So, with the general consumer and the folks on headfi who use a digital source, they all use a D/A converter and usually they can clock the incoming digital signal. But in every DAC you can find some kind of jitter reduction. In a consumer audio D/A, for example you receive incoming digital audio from a CD platter without conversion, directly from the pre out. So the consumer D/A will look at the incoming digital signal from the CD and try to clock off of that but you will a buffer that will remove the jitter from the digital stream and use that clock to ‘clock out’ the data through the converter. But in a professional rig (or high end hifi) where you have both the A/D and D/A in the chain, you have one clock that acts as the master clock that synchronizes both A/D and D/A processes. Keep in mind Apogee was one of the first companies that researched clocking and how in meaningfully affected audio. Many people in the audio realm today understand these concepts of clocking and oversampling, but back in the day these were esoteric issues. Apogee was one of the very first companies to not only discover them, but create products which addressed these issues. The crux of Apogee brand is ‘digital audio, practical issues’, and we try to bring the real world practice of digital audio closer to the theory.
Kishore: Can you touch on some of the significant misunderstandings in the world of digital audio?
Roger: Yes, of course. Earlier, I used graph paper as an analogy for oversampling, and you asked me if using smaller squares would yield a better result for music reproduction. Now, this statement leads to one of the biggest misunderstandings of digital audio. First we need to master the basics, we need to define sample. A sample is every digital sample represents a time and a voltage. So if you have a sine wave, each point on the curve represents a different time and voltage. You can then use those numbers, to reconstruct the analogue signal. Now, this would lead you to believe that the ‘finer’ the graph paper (a higher sample rate) the more accurately you will be able to produce the analogue wave. This is a reasonable expectation. This is also 100% wrong.
Kishore: What? That sounds counterintuitive.
Roger: Haha it does doesn’t it? For instance, if you use Illustrator or Photoshop there is an arc tool. This tool allows you to draw a specific arc between two points. Keep this visual in mind. Now, way digital audio works is when you apply filtering at the A/D stage that the filter ensures any frequency beyond a predetermined frequency point isn’t introduced into the system. The D/A stage filter, you have an identical filter, this is the reconstruction filter. Essentially a filter limits the speed at which a waveform can change. A filter acts like the arc tool in Illustrator. The arc tool forces a predefined curved, which determines how the points can connect, and this is how reconstruction filters work.
If your frequency is under half the sample rate, then it will perfectly connect those two points (samples) in only one way. Let’s say I have a 19kHz sine wave, but if my sample rate is 48kHz there isn’t going to be many samples to define the curve. The magic of digital audio is that if you have two discrete sample points, and you use a digital filter, then there is only one way those two points can ever be connected, and this will give us perfect reconstruction even with very few points. It’s the digital filter which gives us the predetermined path to perfect sound wave reproducibility.
This is one area where digital audio theory works perfectly. If you’re sampling at 48kHz, and nothing above 48kHz got into your digital audio source, and you feed it through a digital filter that’s exactly the same, then the theory directly translates into reality. It’s quite rare to have a direct translation of theory and practice.
In summary, contrary to popular belief, oversampling does not mean better quality. Oversampling means oversampling. There’s no need to put more samples. So what does a higher sampling rate give you? It gives you the ability to open up your filter. So now instead of having a filter at 24kHz, I can have a filter at 48 kHz. With a higher sampling rate, you can achieve higher bandwidth.
Kishore: Hold on. How do you determine a good sampling rate? Doesn’t human hearing top out at 20kHz?
Roger: That’s exactly the question. Does sampling above 20kHz provide anything useful? This is a topic you can spend days, weeks even years going around in circles. What we’ve found is there are some processes that do have benefits at 96 kHz, but this is an impossibly small amount. The vast majority of professionals work in 48kHz. Some, due to the debate of 96kHz vs 48 kHz, work in 96 kHz but I don’t think you’ll find many professionals at all recording at higher than 96 kHz. Anything higher than 96 kHz isn’t really useful for professionals.
After listening critically for many years in this industry, I’m somewhat thrilled I can hear up to 15kHz considering my age haha! The thing is, I’m sure you can debate on Headfi all day about the ‘usefulness’ of recording frequencies above 20kHz, but this is not high on our priority list at Apogee. We focus on what value can we bring to the professional. There are so many other avenues to work on than being concerned with higher kHz. That being said, most if not all of our products do 192 kHz because the market asked for it.
Kishore: I’ve noticed your DAC’s don’t support 384kHz, let alone DSD. But you’re saying even supporting 192kHz is overkill.
Roger: We just don’t think it’s adds enough value for our customers. We’ve had this exhausting conversation so many times with people who ask us ‘why don’t you support 784 kHz or DSD, my arglebargle from China does this for a fraction of the price’. And we say, ok cool. We have a hard time keeping our faces straight when we see folks talk about how songs with higher bitrates (eg 96 kHz) reveal new levels of detail. Also, objectively we’ve done the measurements between higher sample rates, and everything measures identical across the board. Believe me, I would’ve loved to see some differences, anything that can enable the artist to produce better music, and allow the listener to extract more value from the songs is the core of our values. However, all these years later we’ve learned to invest in proven technologies, and these new formats like SACD, DSD, MQA will come and go with time. Just look at how MQA is turning out. Let’s look at DSD, it’s an old technology that got repackaged. Do you remember when we discussed integrated chips? Sony came out with 1 bit oversampling for a generation of chips that was about 20 years ago. Then, they took that internal technology and made it into a public format called DSD. And for the longest time, there was no real acknowledgement that it was old technology. And now we see are multi-bit versions of DSD because there it is generally accepted that 1 bit DSD offers less performance than multi bit DSD, hence nearly all DSD converter chips actually convert 1 bit DSD to multi bit DSD (more info https://bit.ly/2zMqP2A).
Kishore: Can you expand upon what this ‘bit’ means, and its context in digital audio?
Roger: Of course. Let’s take a 48 kHz 24 bit file. 24 bit means you should have several hundred thousand different voltage levels. You have 24 ‘levels’ to define a voltage level between 0 dBFS and -144 dBFS if you’re in the digital domain. I think you have around 600,000 individual bit levels for a 24 bit file (recording format), and for a 16 bit file (CD audio) you have about 16,000 bit levels. The total amount of levels are your dynamic range. In the signal path, you are asking a computer or converter chip to look at the analogue signal 48,000 times per second (assuming a 48kHz file), look at the incoming voltage and finally associate that voltage with one of 600,000 different levels. For a 96kHz file, the converter has to do this 96,000 times per second and so on. The way DSD works is that the converter’s process examines the current sample and asks, is the level greater or lesser than the previous sample, then it adjusts as required. This is done by increments or decrements of one bit. DSD does this incredibly fast, and this is in the MHz range. Essentially, DSD performs an incredibly basic sampling operation, but because it’s simple, it can do this at an incredible speed. So DSD was seen as the next step to improve the A/D and D/A converters. Then they came to the realization that simply transitioning one level up or down wasn’t enough. Now they have more bits to increase the resolution to better characterize the next sample during the sampling process. So that’s what multi-bit DSD is. The tradeoff they are making is between the amazing number of digital levels a device has to detect for non DSD files, versus doing the sampling much simpler but far faster.
DSD multi-bit is happening in every DAC on the market, it happens in the DAC chips we use for instance. At the end of our conversion process however, we use a decimation filter. This takes the output of all that and packages it into a 96 kHz 24 bit format in the digital domain. This is what we record and playback in.
When it’s happening in the converter and assigning the sample a digital value, this is the process that can get complicated. With all these novelty formats, they are taking what’s already occurring in the chip, and externalizing it. The chipmakers are pretty much a decade in front of these formats coming to market, and we at Apogee don’t think this is a coincidence. For example DSD 1 bit was already happening in the chips for about 10 years before Sony came out and made DSD a public format. This same thing is true with DSD multi bit, multi-bit DSD has been in chips for about a decade and in the last five years it has become a public format. By the way, the multi-bit DSD process occurs in our products and in nearly every DAC on the market.
Kishore: Is there any added value to the consumer, by making DSD a public format?
Roger: No. It simply doesn’t haha! If DSD had any value we would have included in our products immediately.
Kishore: Apogee is a unique electronics company, considering you have a full recording studio and have a relationship with the artists that use your products. Do you work with artists when developing new products, if so how does such a collaboration work?
Roger: Oh absolutely! So first all our products are made in the US. Most of our products are manufactured toward the San Diego area. The best way to deal with manufacturing/ calibration issues for instance, is to be able to drive over to the facility with our testing equipment and have it addressed same day. Our flagship product, the Symphony I/O is made in house, you’ve seen the board assembly downstairs, and we have nearly a dozen units undergoing burn in right now before we conduct few more tests and it’s sent to the shipping department.
Regarding artist collaboration, we work with them iteratively and throughout the prototype phase to ensure our product adds significant value to an artists workflow. Let’s look at one of our more popular products, the Duet. This was one of the very first professional interfaces that can be used in big sessions. You can connect microphones, speakers and headphones. Granted it doesn’t have the performance of our flagship series products, but Duet brings unmatched ease of use and portability. We provide prototypes to a number of artists, and they provide feedback on how not just sonic quality, but on ease of use as well. A big part of my job is to talk with customers and artists is to take their feedback and apply it to make our products better.
Kishore: Regarding artists, what's their perspective on these new high bitrate formats? Do they care about it?
Roger: I’ll put it this way, extremely few professionals deal with bitrates above 92kHz. Most professionals are working with 48 kHz.
Kishore: In the hifi world, it’s not uncommon to see $10k, $20k, and even $50k DAC’s. Yet Apogee’s top range studio DAC is relatively affordable, below $5k. Do you feel these other DAC’s are completely unnecessary, or is Apogee working on a premiere DAC to compete with these other super high priced flagships on the market?
Roger: Oh, that is a dangerous question my friend. I would say that professionals are less susceptible to hype. Everything we do in our products has an incredibly practical, but provable reason behind it. When we design and create products we need to see or measure the improvement objectively but we also need to subjectively experience the improvement. Keep in mind, it’s not one or the other, many people get lost in this. It’s a careful combination of the two. Specifications aren’t the only thing.
The trouble is, many products start with older assumptions. For example, it used to be that one had to literally build one’s own converter process out of discrete components. The other way of doing it is using converter chips, which is an integrated chip that has all the conversion performed on the chip itself (such as ESS Sabre or AKM chips). So you may have heard of discrete analogue versus integrated circuit (IC) chips in the analogue domain. When IC chips were first developed, there was some truth to that, in order to get better performance, you had to do everything with discrete parts and an integrated chip resulted in a lower quality solution. However, this was decades ago, and modern chips have astounding quality, the performance far exceeds what you will find in a discrete converter. Now, you will find people that will sell discrete converter which takes this process that is usually encapsulated in a DAC chip but then separates out the processes which involves a custom design for the converter.
Kishore: Are you referring to the FPGA architecture?
Roger: No, I’m referring to the converter chip itself. So, on a DAC you have a digital in, clock inputs, and analogue outputs. The quality of these converter chips is really astounding. Now, sure you can make a good sounding product with a good chip, but not many folks are interested in merely ‘good’ performance, everyone wants the best performance possible. So, to take performance further, it comes down to providing a perfect environment for those inputs and outputs. For example, a clock input that is ultra low jitter and taking those analogue outputs and providing the chip with an electrical environment that facilitates best performance. This is the concept of ‘DAC implementation’ we see being discussed on forums, and it’s this implementation that becomes an art. What consumers need to know is that it’s not all about the type of chip a manufacturer uses, yes that’s a part of it, but it really comes down to the implementation of the chip in the circuit design. I’m sure you’ve seen on some HeadFi threads where someone states ‘This DAC doesn’t use xyz-123 chip, therefore it’s out of spec and a waste of money’. This blanket statement is simply not true. Some manufactures have made this into a discrete process, and it’s based on an older assumption that is ‘discrete will always be better than integrated’. And this is a false assumption.
Kishore: You may have upset some folks reading this.
Roger: Haha…well, so be it. We find that professionals more concerned with getting mission critical work done, rather than experiment with various things for, at best, debatable gains in performance. At Apogee it’s always been about giving the professional the best performance possible at a reasonable cost.
Kishore: Moving to the creative aspect, did you develop your iOS oriented products in anticipation of the ‘bedroom artist’ movement?
Roger: Yes, as you noticed our iOS products and portable professional tools sort of overlap. Actually, the beginning of our iOS product line is an interesting story, and is a big part of our long-standing relationship with Apple Inc. Many years ago, artists who wanted to use their iOS device needed to use products that interfaced with Apple’s headphone jack but such products were terrible quality. Many artists were asking us for a quality product that would allow them to record with iPhones/iPads and Macs. So, while we developed the hardware product called Jam, we reached out to Apple and collaborated with them to create a digital audio protocol for iOS. This protocol is open, and we pushed the industry forward as people weren’t restricted to the headphone jack, and could use the 30 pin port instead. This served as the foundation for our push into the mobile front, and we are proud of it.
Bharat: What’s the nature of your relationship with Apple?
Roger: Haha, everyone asks us that. Folks ask if Apple owns us, or if there are foot high piles of contracts involved. The truth is, it’s an informal relationship. We’ve known their audio engineers for decades, we provide their core audio team with insider knowledge of audio development and recording, and of course they provide us with their massive customer base as they have our products in stock at the Apple Store.
Bharat: Can you tell us about the Ambeo Headset, what did it involve and what was it like working with Sennheiser?
Roger: For us, it was incredible to be able to partner with this legendary audio company, guided by German engineering. I found it interesting that we were able to provide something to Sennheiser, and it’s the same thing with Apple. These companies, Apple especially, are massive and what can our humble outfit offer to them? It turns out there are a couple of things. When working with Sennheiser, it became quickly apparent we can offer them our iPhone expertise as Apogee has a great working relationship with Apple, but also many of our products are designed to work with Apple hardware. Also being a small company, we can move incredibly quickly and turn around a product fairly quickly. And it turns out we can ensure high quality conversion of the headset audio, all while within the confines of the iPhone. Working within Apple’s limitations can be challenging, any developer will tell you this, and when designing a DAC for this application, we always need to operate within tolerances such as the low current load and so on. We were able to provide Sennheiser connectivity with the iPhone, the app, and the conversion technology. We’ve also become adept at creating excellent preamps, and in Ambeo, you have two microphones, and we developed the pre amps for those as well. Throughout this product, there were many areas where we could bring our expertise. Sennheiser were able to bring their earbud and microphone technologies to the table. The reason why we embarked on this project was because we were enchanted by this possibility to create absolutely convincing audio. The binaural audio experience is captivating, and a consumer can experience memories via recorded video with binaural in a whole different way. The Ambeo is dead simple, you just plug it in and begin using the Camera app to record video. We wanted to take binaural technology and put it on the most popular device on the planet. Being able to stun the world with a whole new audio experience, that doesn’t require high spec equipment to experience. All you need is some headphones, and you can have the binaural experience.
Bharat: Is there a way to take a stereo source and make it binaural?
Roger: Facebook has explored this area and developed host of ProTools plugins, one of which is a binaural plug in which spatializes even mono sources. It’s not bad per se, but it’s nothing like the real thing.
Kishore: You mentioned consumers are the target for Ambeo, but how have music professionals reacted to this? Are they using it?
Roger: Some artists have expressed interest in this technology, but there isn’t any specific products or artist creations that utilize Ambeo yet. This is a new tool for them, and figuring out how this can work with their creative vision can take time. Sennheiser is dabbling in this area, and are pushing into the virtual reality space, so we may see something come from that in the near future.
Bharat: Going back to the Sennheiser collaboration, did you have Sennheiser engineers flying in from Germany, or did some folks from Apogee go to Sennheiser’s HQ?
Roger: Actually we have a Sennheiser project manager who’s almost an Apogee employee. Here is here quite a lot, and he is much beloved by the staff here. Betty actually got to know the Sennheiser brothers who currently run the company. We had some high level conversations with these two young brothers who run Sennheiser, and it’s incredible to relate each company’s vision for the future of audio as Apogee and Sennheiser bring complementary skillsets and perspectives to the table. To be honest, Apogee acquired the parts below the earbuds and developed the core technology behind the Ambeo. We manufactured the enclosure, circuit board, the iOS certification but we relied on Sennheiser for the earpieces in the Ambeo.
Kishore: Looking closer at your product line, I am surprised to learn Apogee doesn’t offer DJ products as this is an area Apogee can leverage their expertise and knowledge in. Does Apogee have plans to offer products in the DJ segment?
Roger: Hmm that is true. I would say Duet is a toe in the water for that area, we do have many customers that use Duet for sound mixing in clubs and other such venues. This area is definitely interesting to us, but we haven’t found the right product yet but we definitely have some things on the horizon which can apply more directly.
Kishore: What does the future look like for Apogee?
Roger: We will continue to refine our flagship professional series products such as Symphony I/O, as you saw we have a rack of those being burned in downstairs that are being shipped out at the end of the week. However we are always on the look out for new product categories, for instance we transitioned from being a professional only solutions company to offering consumer, prosumer and professional portable solutions which work with iOS and Mac. We actually have an audio recording program called MetaRecorder (https://bit.ly/2Kts6AL). MetaRecorder is a professional grade field recording app for iOS. At AES, we just announced MetaRecorder works with a UltraSyncBlue system to record and synchronize timecode onto your iPhone (https:// bit.ly/2DTLEOe). This allows people to use iPhones to make movies, which Apple is very eager to push. For instance, if you are a film student this can be very useful, as many students have iPhones and you can shoot multiple angles with multi camera synchronized timecode without exceeding your budget. These kinds of features used to be pro only, and having this available on iPhone is unheard of.
Kishore and Bharat: Thank you so much for your time, we greatly appreciate it.
Roger: No worries, I didn’t think we’d talk about all these topics and in such depth, but I actually enjoyed it. This was quite fun.
*All images are taken from Wikipedia, Apogee or Sennheiser unless otherwise stated.
