Does DSD sound better than PCM?

[pkane]

As you should. You can't meaningfully compare things without matching the reference levels. Beware though, some DSD content exceeds this level for brief periods.

It's not strictly speaking necessary in DeltaWave, as it's computing level difference between two files and doesn't rely on them being matched.

 

[mansr]

It's not strictly speaking necessary in DeltaWave, as it's computing level difference between two files and doesn't rely on them being matched.

You misunderstand. Two PCM files have, by definition, the same reference level. If the RMS value of one file is higher, it will sound louder when played on the same equipment. DSD is a different system with a different reference level corresponding to maximum loudness. Consider the analogue signal from a vinyl pickup compared to a line level signal. The same loudness is represented by very different voltage levels. To compare the actual levels of vinyl and, say, tape playback, the appropriate gain must first be applied.

 

[Ron Texas]

if you want it to sound better it will.

 

[pkane]

You misunderstand. Two PCM files have, by definition, the same reference level. If the RMS value of one file is higher, it will sound louder when played on the same equipment. DSD is a different system with a different reference level corresponding to maximum loudness. Consider the analogue signal from a vinyl pickup compared to a line level signal. The same loudness is represented by very different voltage levels. To compare the actual levels of vinyl and, say, tape playback, the appropriate gain must first be applied.

I get that. But that’s what Deltawave is designed to do: adjust two files to match gain, sampling frequency and phase, regardless of how these files were generated. No prior corrections are required.

 

[mansr]

I get that. But that’s what Deltawave is designed to do: adjust two files to match gain, sampling frequency and phase, regardless of how these files were generated. No prior corrections are required.

I must not be making myself clear. Let's try again. Consider two PCM files, one 16-bit and one 24-bit. The first encodes full scale as 2^15, the other as 2^23. That doesn't mean the 24-bit file is 256 times louder. It's the same reason you can't compare a distance measured in metres to one measured in feet without a conversion. Of course, if you're only interested in comparing aspects other than level differences, automatic gain setting will do fine, although you still need to make sure you're matching the right thing.

 

[pkane]

I must not be making myself clear. Let's try again. Consider two PCM files, one 16-bit and one 24-bit. The first encodes full scale as 2^15, the other as 2^23. That doesn't mean the 24-bit file is 256 times louder. It's the same reason you can't compare a distance measured in metres to one measured in feet without a conversion. Of course, if you're only interested in comparing aspects other than level differences, automatic gain setting will do fine, although you still need to make sure you're matching the right thing.

Maybe it's me.. A sample value of 2^15 in 16-bit PCM will be treated as the same level when compared to a sample at 2^23 in a 24-bit recording. Deltawave is smart like that, and will scale correctly to an internal double floating point format, regardless of whether the input is 8-bits, 16-bits, 24, or 32 bits PCM or 1bit DSD. SACD reference level of 0dBFS is 6dB down from PCM. Hence the 2x level adjustment.

 

[mansr]

A sample value of 2^15 in 16-bit PCM will be treated as the same level when compared to a sample at 2^23 in a 24-bit recording. Deltawave is smart like that, and will scale correctly to an internal double floating point format, regardless of whether the input is 8-bits, 16-bits, 24, or 32 bits PCM

So far, so good. This is so obvious we tend to not think about it.

or 1bit DSD. SACD reference level of 0dBFS is 6dB down from PCM. Hence the 2x level adjustment.

Right. The whole "DSD is 6 dB quieter than PCM" thing is nonsense. It's like saying things in America are 3x bigger because they are measured in feet.

 

[Sal1950]

None of these have had any significant impact on music recordings. Almost all multichannel music recordings released on disc are SACD. Multichannel PCM (not Dolby/dts) has recently become a factor in file downloads.
Unfortunate but true. So?.

The first statement is true, but has a limit. SACD came along way before the latest BluRay formats and has that "audiophile magic dust" aura of DSD attached to it. It is "special" to it's devotees even requiring special players to be able to access it. But DSD as a superior sounding format is a myth as is being highlighted in this thread.
Truth is SACD as a format has a foot in the grave, and DSD-A is there already, and rightly so. Despite the audiophile magic dust aura of DSD, there is nothing special about it. For the multichannel world, BD offers about as many channels in lossless high data rate PCM as we could ever ask for. It's way past time to move to it for the production of the multich music standard format. Hanging on to SACD or DVD-A makes as much sense as hanging on to vinyl as the 2ch standard for high fidelity playback.
Look at what Mark Waldrep's AIX records is offering on one BD, 5.1 Stage mix, 5.1 Audience mix, 2 channel stereo 24/96 mix, with HD video included.
SACD and DSD-A were great formats, but progress and the SOTA has passed them by.

 

[jsrtheta]

No recording medium is without intrinsic issues that limit fidelity ...
PCM and DSD should - and can co-exist, in the recording industry, IMHO, each has certain advantages over the other, technically-wise, primarily based on what they do best.

What does DSD do best? Serious question.

 

[jsrtheta]

The first statement is true, but has a limit. SACD came along way before the latest BluRay formats and has that "audiophile magic dust" aura of DSD attached to it. It is "special" to it's devotees even requiring special players to be able to access it. But DSD as a superior sounding format is a myth as is being highlighted in this thread.
Truth is SACD as a format has a foot in the grave, and DSD-A is there already, and rightly so. Despite the audiophile magic dust aura of DSD, there is nothing special about it. For the multichannel world, BD offers about as many channels in lossless high data rate PCM as we could ever ask for. It's way past time to move to it for the production of the multich music standard format. Hanging on to SACD or DVD-A makes as much sense as hanging on to vinyl as the 2ch standard for high fidelity playback.
Look at what Mark Waldrep's AIX records is offering on one BD, 5.1 Stage mix, 5.1 Audience mix, 2 channel stereo 24/96 mix, with HD video included.
SACD and DSD-A were great formats, but progress and the SOTA has passed them by.

I never heard DVD-A, but I've heard SACD, and I found nothing great about it. It was nice that some material came out that was made from different masters than CD, but the solution there would be to release the different masters on CD. I am not aware of any DBT that proves SACD sounded a damned bit different. (There may have been one, but I never saw it.)

 

[Wombat]

What does DSD do best? Serious question.

FWIW. Interesting read or start at the conclusions: DSD vs PCM

 

[jsrtheta]

FWIW. Interesting read or start at the conclusions: DSD vs PCM

I will endeavor to try to understand this - my science/engineering aptitude is miniscule. One question: Doesn't DSD require a lot of PCM to process? And isn't virtually uneditable? Again serious, if ignorant, questions.

 

[Sal1950]

FWIW. Interesting read or start at the conclusions: DSD vs PCM

The layout and graphics of that guys site gives me a eye-ache after a few minutes. It tells me to just go somewhere else before I freak out.

 

[jsrtheta]

The layout and graphics of that guys site gives me a eye-ache after a few minutes. It tells me to just go somewhere else before I freak out.

You too?

 

[Wombat]

You guys need to read more textbooks rather than color-in books.

 

[bennetng]

I must not be making myself clear. Let's try again. Consider two PCM files, one 16-bit and one 24-bit. The first encodes full scale as 2^15, the other as 2^23. That doesn't mean the 24-bit file is 256 times louder. It's the same reason you can't compare a distance measured in metres to one measured in feet without a conversion. Of course, if you're only interested in comparing aspects other than level differences, automatic gain setting will do fine, although you still need to make sure you're matching the right thing.

Not long ago someone misinterpreted my posts in the same way:
https://www.audiosciencereview.com/...nd-dacs-with-volume-controls.5432/post-121603

BTW, since you are a SoX maintainer, could you advise how to solve the clipping problem below? Thanks!
https://hydrogenaud.io/index.php/topic,101850.0.html

 

[Thomas savage]

This thread is going round in circles, unless you have some data to contribute please satisfy yourself by reading though whats be written.

 

[Wombat]

To the OP. No.

Yes there are technical differences in processing and file format conversions but these should be of no concern to the consumer if competently done. Software is available that does this well.

Overthinking minutiae again, I feel. There is more involved for those engaged in recording but they should learn what is required from reputable sources(dare I say it, get a qualification?) rather than internet audio forums.

Asking simple questions from an uniformed stance and expecting simple understandable answers is often too big an expectation. One has to ask, though because one may get a satisfactory response.

 

[FrantzM]

This thread is going round in circles, unless you have some data to contribute please satisfy yourself by reading though whats be written.

That DSD is Bivine manure? That conclusion was reached about 10 posts into the thread.

 

[Wombat]

Bivine? Is that related to Unicorns?

The OP asked about sound quality. Not the content capacity of disc types or other extraneous matters.