[theaudiologist1]

Is DSD in general more quiet than PCM? I have a DAC that outputs 28mW into 300Ohms and my headphones are 470 Ohms at 98dB sensitivity, but surprisingly, it gets very loud even with classical on the PCM side. It still does loud enough on DSD on non-classical music, but with classical, it gets a bit more quiet and I sometimes had to put my DAC on max gain and volume for the quiet parts and it becomes harder to separate the instruments. This problem mostly got fixed by turning off replaygain in Audirvana but some classical songs are still a bit quiet in the quieter parts. But I'm just asking is it natural for DSD to sound more quiet and "flat" than PCM? Even the PCM files I upsample to DSD get more quiet. Is it also because I'm using -6dB gain in the PCM to DSD settings of Audirvana?

 

[bigshot]

First for the theory part... An SACD has a lower noise floor, but for the purposes of listening to recorded music in the home, it is no better than a CD. A CD has a noise floor at about -90dB The quietest listening room (think the quietest public library possible) has a room noise floor of about 35dB. So, to be able to hear the quietest sound on a CD, you would have to raise the volume above the ambient noise in the room- 90 + 35 is 125dB. That is beyond the threshold of pain. If your stereo is even capable of playing that loud, you will incur hearing damage pretty quickly. Bigger numbers are not always better in home audio. There is a point where it's good enough and more won't make it better. That is the limit of your own human ability to hear. A CD is all you need. There's an article in my sig on that if you'd like to read more.

Now, to address your practical problem... Where are you getting the PCM and DSD audio from? Are they both from the two layers of an SACD? Or are you comparing a CD to an SACD? The reason I'm asking is because different releases of a recording might be mastered quite differently. A CD is targeted at regular people, so the level of compression is a little higher. SACDs are targeted at audiophiles, so they may have a broader dynamic range. From what you are describing, the SACD has too much dynamic range. I've found that myself on some classical recordings, particularly those on the BIS label. You might want to try a variety of recordings and see if it varies. If so, it's probably mastering. I've found that some SACDs have different mastering on the PCM layer than they do on the SACD one. In that case, just choose which sounds better to you. There's no sin in listening to the PCM layer of an SACD if you like it better. It's not like there is a difference in sound fidelity between the two. Whichever works is fine.

 

[theaudiologist1]

First for the theory part... An SACD has a lower noise floor, but for the purposes of listening to recorded music in the home, it is no better than a CD. A CD has a noise floor at about -90dB The quietest listening room (think the quietest public library possible) has a room noise floor of about 35dB. So, to be able to hear the quietest sound on a CD, you would have to raise the volume above the ambient noise in the room- 90 + 35 is 125dB. That is beyond the threshold of pain. If your stereo is even capable of playing that loud, you will incur hearing damage pretty quickly. Bigger numbers are not always better in home audio. There is a point where it's good enough and more won't make it better. That is the limit of your own human ability to hear. A CD is all you need. There's an article in my sig on that if you'd like to read more.

Now, to address your practical problem... Where are you getting the PCM and DSD audio from? Are they both from the two layers of an SACD? Or are you comparing a CD to an SACD? The reason I'm asking is because different releases of a recording might be mastered quite differently. A CD is targeted at regular people, so the level of compression is a little higher. SACDs are targeted at audiophiles, so they may have a broader dynamic range. From what you are describing, the SACD has too much dynamic range. I've found that myself on some classical recordings, particularly those on the BIS label. You might want to try a variety of recordings and see if it varies. If so, it's probably mastering. I've found that some SACDs have different mastering on the PCM layer than they do on the SACD one. In that case, just choose which sounds better to you. There's no sin in listening to the PCM layer of an SACD if you like it better. It's not like there is a difference in sound fidelity between the two. Whichever works is fine.

My DSD files are mostly SACD rips to DSD64 but I also have some DSD128 and DSD256 stuff. My PCM are all from various resolutions from 16/44.1 FLAC to 32/192 Wavpack, but I upsample them all to DSD128. I don't think it has to do with the mastering as much as it does it being DSD, since even converting the DSD to PCM on the go makes those recordings louder. I never even listen to the PCM layers of SACD's (I don't even have an SACD player, I just convert my SACD's to ISO using my PS3 and then convert the ISO into DSD64 files).

Regarding my classical music, I feel like I'm not getting "absorbed" into the music as much as the other genres I have. I feel like I'm experiencing the opposite of clipping (volume not loud enough to hear all the details). People say the volume of an orchestra at full force is 120dB, and with my dac and headphones it only goes to around 110dB-115dB max volume, so I feel like I'm missing out on some of the details and resolution of listening to them on a higher volume (like with my lower resistance 32 Ohm headphones). Idk if it's just my paranoia just "feeling" it's not loud enough on the quieter parts, or if that's just meant to be like that.

 

[bigshot]

I'm not going to bother to answer all your misconceptions individually, because you didn't pay any attention to what I said in my previous comment. Maybe someone with more patience will answer you point by point.

But I will say this... You are approaching this completely backwards. You are taking your subjective impressions, and looking for some sort of technical minutia to justify them, while simultaneously ignoring the most obvious reasons for your subjectivity- bias and perceptual error. Your sole goal here seems to be to validate your bias. I offered you the opportunity to eliminate bias and perceptual error from the equation so you could find out the truth. But you aren't interested in that. You're just interested in cherry picking irrelevant factoids to back up your predetermined conclusions. That's what salesmen trying to convince audiophools do.

I'm not too keen on providing you any more information, because your intentions are all wrong. You'll just take whatever I tell you out of context and twist it to fit your agenda. You can count me out of that game. Have a nice day!

 

[theaudiologist1]

I'm not going to bother to answer all your misconceptions individually, because you didn't pay any attention to what I said in my previous comment. Maybe someone with more patience will answer you point by point.

But I will say this... You are approaching this completely backwards. You are taking your subjective impressions, and looking for some sort of technical minutia to justify them, while simultaneously ignoring the most obvious reasons for your subjectivity- bias and perceptual error. Your sole goal here seems to be to validate your bias. I offered you the opportunity to eliminate bias and perceptual error from the equation so you could find out the truth. But you aren't interested in that. You're just interested in cherry picking irrelevant factoids to back up your predetermined conclusions. That's what salesmen trying to convince audiophools do.

I'm not too keen on providing you any more information, because your intentions are all wrong. You'll just take whatever I tell you out of context and twist it to fit your agenda. You can count me out of that game. Have a nice day!

I was NEVER biased towards DSD. I simply said it was has a sound I prefer. I simply told you what I heard from my DSD and PCM files that DSD is less "punchy" and more flat, and also, less loud, than my PCM (whether it actually is like or if it's my DAC idk). I ALREADY KNOW that 99% of DSD was recorder or mixed or mastered in PCM or Analog. No need to punch me in the face for that. All I simply said was, WHEN it IS fully native DSD (from places like NativeDSD.com), the audio PERSONALLY sounds more lively than a PCM recording and that's if it is recorded right. I just wanted to know WHY, even when I upsample my ALREADY PCM files to DSD on the fly using Audirvana, that the upsampled DSD audio ends up being more quiet. Is it just the nature of DSD?

 

[bigshot]

Sorry. You should read a little more. In this forum, people don’t care how you feel about sound. Your subjective impressions are off topic here. We only are interested in facts. You’re not doing well there, because you haven’t taken the time to figure out how audio works. You could learn about that if you listened more and talked less. The other option is to go back to the forums that will welcome your blather.

 

[gregorio]

[1] I was NEVER biased towards DSD. I simply said it was has a sound I prefer.
[2] I ALREADY KNOW that 99% of DSD was recorder or mixed or mastered in PCM or Analog. No need to punch me in the face for that. All I simply said was, WHEN it IS fully native DSD (from places like NativeDSD.com), the audio PERSONALLY sounds more lively than a PCM recording and that's if it is recorded right.
[3] I just wanted to know WHY, even when I upsample my ALREADY PCM files to DSD on the fly using Audirvana, that the upsampled DSD audio ends up being more quiet. Is it just the nature of DSD?

1. Isn't preferring the sound of DSD the definition of being "biased towards DSD"? Especially as there's no audible sound difference.

2. But apparently there is a need to "punch you in the face for that" because there is no "WHEN it IS fully native DSD (from places like NativeDSD.com)"! NativeDSD writes on it's website:

"All DSD acoustic recordings start as musicians in an acoustic space being picked up with multiple microphones. The analog signals from these mics are first amplified by mic preamps, and fed to DSD Analog to Digital converters. Some labels will perform a session analog mix and balance before A/D conversion, but most labels, perform that mix and balance digitally in post production. In any case, post production always includes editing to correct note or tempo phrasing, and playing errors" - In order to edit and "mix and balance digitally" the 1bit DSD must be converted and therefore is NOT "fully native DSD"!

3. It's probably just some setting in Audirvana. In theory, upsampling to DSD would often require some level reduction, in order to avoid intersample peaks, but most probably it wouldn't need more than a dB or two and wouldn't noticeably sound much quieter. Check that you're not using replaygain or some other setting that could make it much quieter.

G

 

[Brahmsian]

different releases of a recording might be mastered quite differently. A CD is targeted at regular people, so the level of compression is a little higher. SACDs are targeted at audiophiles, so they may have a broader dynamic range. From what you are describing, the SACD has too much dynamic range. I've found that myself on some classical recordings, particularly those on the BIS label. You might want to try a variety of recordings and see if it varies. If so, it's probably mastering. I've found that some SACDs have different mastering on the PCM layer than they do on the SACD one. In that case, just choose which sounds better to you. There's no sin in listening to the PCM layer of an SACD if you like it better. It's not like there is a difference in sound fidelity between the two. Whichever works is fine.

Sometimes low-level detail will be easier to hear in a recording than in a concert hall. I attribute that not only to mic placement but to compression. If that’s true, I would say that, to a certain extent, compression in classical music recordings is good. Too much will obviously ruin the recording.

 

[bigshot]

Tools aren't usually inherently good or bad. It all depends on how the tool is used. (I guess the exception are things like nooses and iron maidens.) An old time sound engineer told me about a broadband noise reduction unit that had been introduced- a lot of people were up in arms because the dial could go up so high it could melt the sound. They felt that the unit should have been only given a range that could do no harm to the recording. But the question is, how do you know what too much sounds like if the dial doesn't go up that high? People like answers that are absolute- flip this switch and all your problems will be solved- but the world doesn't work like that. Everything is relative, like Goldilocks and the Three Bears. This porridge is too hot. This porridge is too cold. This porridge is JUST RIGHT. It takes discernment to judge things like that. It isn't an absolute black and white thing.

 

[danadam]

I just wanted to know WHY, even when I upsample my ALREADY PCM files to DSD on the fly using Audirvana, that the upsampled DSD audio ends up being more quiet. Is it just the nature of DSD?

I don't know how accurate this is, but here you go. From Roon Knowledge Base:

One of the design tradeoffs in Sigma-Delta modulator design is the management of mathematical instability. Sigma-Delta modulators utilize feedback (or feed-forward) loops that can become unstable if the input signal is too loud.

One of the most common practical measures used to prevent instability is limiting the amplitude of the input signal. In fact, the Scarlett Book (SACD) specifications require that the input signal be limited to -6 dBFS (meaning, 6dB less dynamic range than the original signal).

From mansr comment:

DSD is usually at -6dB or so relative to the same PCM recording.

That's a bit of a misconception. All it means is that if you digitally low-pass filter the DSD signal, the peak value of the result is 0.5. Nothing says a DAC can't, or shouldn't, produce the same peak voltage level for PCM and DSD. Then again, nothing says they should, so in practice they are all over the place.

As an example, here's an iFi DAC playing a 0 dbFS 1 kHz tone as PCM:
IMG
Same DAC playing a full-scale tone as DSD256:
IMG
That's a reduction by 4.5 dB.

 

[Brahmsian]

No possibility that DSD colors the sound? If not, why not? I don't say that my DSD recordings sound better than PCM, just different. Warmer. Now I don't know whether that is inherent to DSD, mastering, my own subjective bias, or the house sound of my favorite DSD labels. Note that DSD sounds the same to me via PCM, so it's not that PCM can't replicate what I'm hearing.

 

[bigshot]

Take your choice- mastering or expectation bias. To eliminate the first option, take a DSD native recording and bump it down to 16/44.1 and compare it. (I've done this.) To eliminate the second, do a blind comparison. (I did this too.) Sounded exactly the same. If you are interested in setting up a test yourself to try what I did and see if you come out the same, let me know and I'll help you.

 

[Brahmsian]

Take your choice- mastering or expectation bias. To eliminate the first option, take a DSD native recording and bump it down to 16/44.1 and compare it. (I've done this.) To eliminate the second, do a blind comparison. (I did this too.) Sounded exactly the same. If you are interested in setting up a test yourself to try what I did and see if you come out the same, let me know and I'll help you.

I sold my SACD player a few years ago. I'm currently listening to native DSD recordings as 16/44.1 and 24/96 ALAC files downloaded from Qobuz. DSD recordings sound the same to me when converted to 16/44.1. For me, it's not a matter of PCM not being able to faithfully reproduce native DSD. It's that material recorded as DSD sounds a little darker or warmer even when I play it back as 16/44.1. Not sure how I can A/B that. I would have to simultaneously record the same performance in DSD and PCM then compare the recordings. From what you understand, there is no reason why music recorded as DSD should sound different than if it were recorded as PCM? They should sound identical?

 

[bigshot]

I compared a native DSD recording to the same recording 16/44.1. I compared using headphones, speakers and a friend's rig who is a professional sound mixer. I could never hear any difference at all. In my own system, I run the output of my Oppo player as PCM, because it just doesn't make any audible difference. It isn't surprising to me. The line for audible transparency lies below even 16/44.1. You can't do better than perfect transparency.

If you are hearing a difference, it's either expectation bias or some funky transcoding going on somewhere. The former is much more likely than the latter. However if you're playing DSD on your computer, the comp may have some sort of problems playing DSD, making it sound different. I suppose that is possible.

 

[sander99]

If you are hearing a difference, it's either expectation bias or some funky transcoding going on somewhere.

I think you misunderstood @Brahmsian, he is not hearing a difference between a DSD recording and that same DSD recording converted to PCM 44.1/16, but he thinks maybe a DSD recording sounds darker or warmer compared to a PCM 44.1/16 recording were it made at the same event. In other words: it seems he is not doubting the transparency of PCM 44.1/16, but he is doubting the tranparency of DSD! Certainly he draws the wrong conclusion from the observations he made, because how would he know how a certain piece would have sounded if it were recorded different? If I understand him correctly he just thinks the DSD recordings that he heard on average sound darker or warmer than the PCM recordings he heard (comparing completly different recordings/performances/masters).
To test the transparency of DSD he would have to start with a good PCM recording, convert that to DSD and then do a blind comparison of those two.