Requesting metadata support for DSD1024 files

[Marco 007Yo]

The main reason for that would be marketing.
Be honest to yourself and ABX the originals against the upsampled files. Our brains play tricks with us.

I understand it differently. Oversampling simply allows us to transmit the original sound to our ears more accurately and in detail. In general, 15 years ago, they used to say that about 192/24khz and dsd 64, they say this is marketing and does not make sense. But now many have checked it out and are convinced that it sounds better. So from the technical side, pcm 1536 and DSD 2048 are not some kind of abstract figure, but a working improvement mechanism.

[garym]

I understand it differently. Oversampling simply allows us to transmit the original sound to our ears more accurately and in detail. In general, 15 years ago, they used to say that about 192/24khz and dsd 64, they say this is marketing and does not make sense. But now many have checked it out and are convinced that it sounds better. So from the technical side, pcm 1536 and DSD 2048 are not some kind of abstract figure, but a working improvement mechanism.

You are convinced it sounds better. So why not report the results of a 10 trial ABX test. That would confirm that you in fact can detect the difference in a one of these files and the same file (same mastering/mix) in 16/44.1. The problem with these discussions is that proponents of the benefits seem to never be willing to do a simple ABX test. foobar2000 can easily perform an ABX test for you with statistical analysis, etc. It won't tell you what sounds better. But it will definitively test whether you can even detect any difference in the versions. If you can't confirm that you can detect a difference, then the discussion about sound quality is moot.

(I know some will immediately say, listening to short snippits doesn't work, but keep in mind that one can do an ABX test with as long a listening session as you desire. It can be hours!).

[GBrown]

You are convinced it sounds better. So why not report the results of a 10 trial ABX test. That would confirm that you in fact can detect the difference in a one of these files and the same file (same mastering/mix) in 16/44.1. The problem with these discussions is that proponents of the benefits seem to never be willing to do a simple ABX test. foobar2000 can easily perform an ABX test for you with statistical analysis, etc. It won't tell you what sounds better. But it will definitively test whether you can even detect any difference in the versions. If you can't confirm that you can detect a difference, then the discussion about sound quality is moot.

(I know some will immediately say, listening to short snippits doesn't work, but keep in mind that one can do an ABX test with as long a listening session as you desire. It can be hours!).

I still stand firm on the fact that once you have converted any source to digital, that will be the upper limit of the resolution. Even if that is lossless FLAC, ALAC, OGG, whatever - you cannot make the signal any better than the weakest link in the chain. Colourize this any way you want, but the end result is nothing more than a larger file that perfectly preserves the quality of the previous compression. If the source is not based on a DSD2048 master, there is no point upconverting it IF your intent is not to use interpolation but to truly have bit-perfect playback.

[garym]

I still stand firm on the fact that once you have converted any source to digital, that will be the upper limit of the resolution. Even if that is lossless FLAC, ALAC, OGG, whatever - you cannot make the signal any better than the weakest link in the chain. Colourize this any way you want, but the end result is nothing more than a larger file that perfectly preserves the quality of the previous compression. If the source is not based on a DSD2048 master, there is no point upconverting it IF your intent is not to use interpolation but to truly have bit-perfect playback.

I agree. I would be shocked and amazed if someone could actually demonstrate anything other than random guessing with an actual ABX test. The existing evidence shows this is hard enough even with high quality, high bit rate mp3 file compared to a 16/44.1 wav (lossless) file.

[Marco 007Yo]

I agree. I would be shocked and amazed if someone could actually demonstrate anything other than random guessing with an actual ABX test. The existing evidence shows this is hard enough even with high quality, high bit rate mp3 file compared to a 16/44.1 wav (lossless) file.

You don't understand a bit what it's for. Increasing the frequency of discrediting without changing the original sound allows you to more accurately represent the original sound in detail. And yes, in general, the original studio master is not made in such resolutions. Usually they write in wav/dxd, then they process it and do it in DSD 256/512/1024 or more.

[Marco 007Yo]

Look at the DSD spectrograms (the official SACD and the one that is officially published on digital platforms). At the bottom is the audible part, then the area of noise filtering and noise removal into the inaudible range. This is all done to better convey the original 20kHz sound to us. The higher the DSD resolution, the better these 20kHz are fed into our ears.

[garym]

I do understand your hypothesis. But that is what it is: a testable hypothesis. You can easily prove your assertions with an ABX test. Unless you are willing to provide such a simple test, all of this is just your opinion. Again, in all these conversations, a simple ABX test would answer the question, but again, rarely does anyone arguing your position provide such basic evidence. Anyhow, it's your files, your storage and your life, so I have nothing personally against doing whatever you want to do. But don't claim it as fact with no rigorous testing (which in this case is trivial to actually do).

[Marco 007Yo]

I do understand your hypothesis. But that is what it is: a testable hypothesis. You can easily prove your assertions with an ABX test. Unless you are willing to provide such a simple test, all of this is just your opinion. Again, in all these conversations, a simple ABX test would answer the question, but again, rarely does anyone arguing your position provide such basic evidence. Anyhow, it's your files, your storage and your life, so I have nothing personally against doing whatever you want to do. But don't claim it as fact with no rigorous testing (which in this case is trivial to actually do).

It's more simple how I perceive it. DSD itself gives the sound a different structure, due to a different modulation and noise suppression mechanism. The sound becomes very rich and concentrated. It works the same way in wav pcm, but a little worse. We are not talking about lossy vs lossless difference tests.

[Marco 007Yo]

Regarding sound testing with oversampling, I personally hear the difference between DSD 64/128/256/512/1024. Many people do too. In my subjective opinion, with an increase in the frequency of oversampling, the sound becomes more voluminous, crystal clear, and the detail increases.

[garym]

To be clear, you are unwilling to do an ABX test. Correct?

[GBrown]

Regarding sound testing with oversampling, I personally hear the difference between DSD 64/128/256/512/1024. Many people do too. In my subjective opinion, with an increase in the frequency of oversampling, the sound becomes more voluminous, crystal clear, and the detail increases.

Are you refering to a song mastered natively in these DSD 64/128/256/512/1024 rates? If yes then I can agree there is potential for each to have progressively better handling of noise, pushing it further and further into the inaudible spectrum. Assuming you have decoding hardware that can handle the higher resolution formats. And it is definitely subjective at these rates as to whether or not you can actually perceive any difference, real or otherwise.

But you simply cannot convert from one file format to another and magically recover data that is not there. Even from "lossless" compression formats, you are still limited by the bit depth and sample rates that were originally used. You are just using larger containers to store the same amount of remaining information. This is not subjective, it is 100% quantitative.

[Marco 007Yo]

DSD he gives a different representation of the sound. And the source of this sound is not important. By means of the format algorithm, the sound changes its structure to a different one than PCM.

[Marco 007Yo]

The first picture shows a DSD spectrogram. Take a good look at what's below. This is just the audible part. This is necessary for Noise Shaping. Increasing the frequency of discrediting through oversampling gives us a more accurate idea of the original sound. The second picture shows the difference in sound structure in DSD and PCM.

[Marco 007Yo]

Are you refering to a song mastered natively in these DSD 64/128/256/512/1024 rates? If yes then I can agree there is potential for each to have progressively better handling of noise, pushing it further and further into the inaudible spectrum. Assuming you have decoding hardware that can handle the higher resolution formats. And it is definitely subjective at these rates as to whether or not you can actually perceive any difference, real or otherwise.

But you simply cannot convert from one file format to another and magically recover data that is not there. Even from "lossless" compression formats, you are still limited by the bit depth and sample rates that were originally used. You are just using larger containers to store the same amount of remaining information. This is not subjective, it is 100% quantitative.

It's not about restoring data, but about giving the sound a different shape. Learn more about how the DSD format works.

[Marco 007Yo]

Also, learn more about what oversampling is and what it is for. This is not necessary in order to hear ultrasound, but to give the accuracy of the calculation to the sound. The technology is very old. Even in the early and mid-90s there were CD players with this feature. Only a lower level of oversampling was used than at present. For example, 8x CD quality/20 bit depth. Further, the technical capabilities increased as well as the amount of storage media. Now there are DACs, players, converters that allow you to process audio in resolutions up to 1536khz pcm and DSD 2048. As well as studios that release music in DSD 1024 format. And yes, I personally hear the differences between CD quality, HI RES quality, DSD SACD quality, sound is reproduced using oversampling by various equipment such as players and DACs.