Requesting metadata support for DSD1024 files

[vilsen]

We've already heard the hypothesis behind hi-res / oversampling many times. You could be the first one to actually prove that you can hear an improvement - or even a difference - by performing an ABX test. Please take the challenge.

[Marco 007Yo]

We've already heard the hypothesis behind hi-res / oversampling many times. You could be the first one to actually prove that you can hear an improvement - or even a difference - by performing an ABX test. Please take the challenge.

This test is not objective to the end, since the comparison takes place in the playback of music through the browser. On more technically advanced equipment and players, the difference between mp3 flac dsd is more clearly audible. I suggest you take another test, compare the sound on SACD vs CD vs HI RES vs MP3. On higher-quality and specialized equipment. Or take a player, turn on the usual FLAC on it with real-time oversampling options and evaluate the result.

[Marco 007Yo]

Everyone decides for himself what to listen to and how to listen. I don't think it's worth arguing about these topics. I just would like to get the DSD 2048/PCM 1536 option in Dbpoweramp like many other fans of this resolution. Anyway, there is a DSD 64-1024 option in it. There is just a proposal to increase it to 2048.

[Marco 007Yo]

It's just very convenient when there is high-quality software that is able to support any sound options. It doesn't matter if it's mp3 128 or DSD 1024+. That's why I prefer poweramp and love the diverse functionality.

[garym]

This test is not objective to the end, since the comparison takes place in the playback of music through the browser.

This is incorrect. One can easily do an ABX test using the exact same playback chain they use for regular music listening. But at this point we're talking past each other. Enjoy the music.

[GBrown]

Modern players and DACs support audio resolution up to DSD 2048. Therefore, you need to add this to the poweramp.

There is a big difference between files from a native source that may have been encoded at these extremely high rates versus an attempt to upsample a file created at a lower rate. What benefit would be realized from upconverting even a 24-bit/96kHz track to DSD2048?

[Marco 007Yo]

There is a big difference between files from a native source that may have been encoded at these extremely high rates versus an attempt to upsample a file created at a lower rate. What benefit would be realized from upconverting even a 24-bit/96kHz track to DSD2048?

There is an advantage of improving the sound by means of: oversampling, stronger noise shaping, as well as giving a different structure to the sound in the form of delta sigma modulation.

[GBrown]

There is an advantage of improving the sound by means of: oversampling, stronger noise shaping, as well as giving a different structure to the sound in the form of delta sigma modulation.

You have stated this previously. However you simply cannot create a higher resolution file with more information than existed from a previously converted file. You are just creating a bigger file that still represents exactly what the original file contained. No different than converting a highly compressed mp3 file to FLAC. The end file may be in FLAC format and show a bigger file size, but this will still be a duplicate of what the mp3 file contained, not better.

[Marco 007Yo]

You have stated this previously. However you simply cannot create a higher resolution file with more information than existed from a previously converted file. You are just creating a bigger file that still represents exactly what the original file contained. No different than converting a highly compressed mp3 file to FLAC. The end file may be in FLAC format and show a bigger file size, but this will still be a duplicate of what the mp3 file contained, not better.

Again, you don't understand how it works. Oversampling and other processes improve the sound without changing its natural frequency. The Nyquist Kotelnikov theorem tells us that the original sound needs to be resampled to the X2 frequency in order to be definitely hear it. But x2 is not enough for this, so the higher the oversampling, the more accurate and detailed the original sound is presented.

[Dat Ei]

Therefore it is necassary to sample the original, analog sound signal (sound waves) with a high samplerate and high bit depth. Once you have converted the analog signal to a digital signal, this data have a strictly limited information content. Umsampling this data just increases the digital signal (file size), but not the information content (entropy). It's like scanning only every second page of a book. Once done, you can't restore the missing pages by increasing the number of pages. You can guess, and some algorithms might have a good guess, but one can discuss, if this music file still represents the pure music. It's a bit like photoshop for music.


Dat Ei

[Spoon]

Oversampling is done on hardware (DACs) to simply the circuit design by moving the upper frequency away from the upper frequency, this allows negative effects from DACs to be reduced, such as aliasing, etc:

https://hydrogenaud.io/index.php/topic,101971.0.html

[Marco 007Yo]

Oversampling is done on hardware (DACs) to simply the circuit design by moving the upper frequency away from the upper frequency, this allows negative effects from DACs to be reduced, such as aliasing, etc:

https://hydrogenaud.io/index.php/topic,101971.0.html

A slightly incorrect analogy with Photoshop. Oversampling adds a computational boost in the form of a discrediting frequency. Due to this, the original frequency is reproduced very accurately, in detail and in volume. The original natural frequency of any music is mostly up to 20kHz. In hi-res recordings above 20kHz, only harmonics and overtones of sound are used. Oversampling has been used for a very long time and not only in DACs, but in players and audio processing programs.

[Marco 007Yo]

Therefore it is necassary to sample the original, analog sound signal (sound waves) with a high samplerate and high bit depth. Once you have converted the analog signal to a digital signal, this data have a strictly limited information content. Umsampling this data just increases the digital signal (file size), but not the information content (entropy). It's like scanning only every second page of a book. Once done, you can't restore the missing pages by increasing the number of pages. You can guess, and some algorithms might have a good guess, but one can discuss, if this music file still represents the pure music. It's a bit like photoshop for music.


Dat Ei

The frequency information increases this process to better reproduce the original sound.

[GBrown]

Oversampling adds a computational boost in the form of a discrediting frequency. Due to this, the original frequency is reproduced very accurately, in detail and in volume. The original natural frequency of any music is mostly up to 20kHz. In hi-res recordings above 20kHz, only harmonics and overtones of sound are used. Oversampling has been used for a very long time and not only in DACs, but in players and audio processing programs.

The frequency information increases this process to better reproduce the original sound.

Just so you have a real point a reference, Spoon is the developer of the software, and has contributed significant white paper level detail regarding digital music handling here. So any attempt to contradict his responses will not likely be taken seriously by others.

You may want to look a little further into music recording techniques. Analog has long been exceeding the 20kHz range, this is only a theoretical limit that was used for the Redbook standard written for CD. The decision was based on the best human hearing range of 20Hz to 20kHz, although most cannot hear this full range ever. Many newer recordings even in the digital realm extend far beyond 20kHz.

There are plenty of solid technologies on the hardware side that can maximize handling of digital media. Some are better than others. But no matter how you refine it, the original media source plays the only role in defining the maximum resolution. If you slice a true analog recording into digital words using any king of transfer, big lossy chunks or tiny "lossless" resolution, that is the best it can be moving forward. Even a 24bit/192kHz FLAC file becomes a stepped digital representation of an analog signal. Of course the higher the resolution, the smaller the steps. But no matter what process or conversion you try apply after each stage, the end result will never be better than the lowest resolution that was previously applied.

[Marco 007Yo]

Just so you have a real point a reference, Spoon is the developer of the software, and has contributed significant white paper level detail regarding digital music handling here. So any attempt to contradict his responses will not likely be taken seriously by others.

You may want to look a little further into music recording techniques. Analog has long been exceeding the 20kHz range, this is only a theoretical limit that was used for the Redbook standard written for CD. The decision was based on the best human hearing range of 20Hz to 20kHz, although most cannot hear this full range ever. Many newer recordings even in the digital realm extend far beyond 20kHz.

There are plenty of solid technologies on the hardware side that can maximize handling of digital media. Some are better than others. But no matter how you refine it, the original media source plays the only role in defining the maximum resolution. If you slice a true analog recording into digital words using any king of transfer, big lossy chunks or tiny "lossless" resolution, that is the best it can be moving forward. Even a 24bit/192kHz FLAC file becomes a stepped digital representation of an analog signal. Of course the higher the resolution, the smaller the steps. But no matter what process or conversion you try apply after each stage, the end result will never be better than the lowest resolution that was previously applied.

I did not argue and contradict Spoon, this should be perceived as an adequate discussion. Regarding recordings where the sound is above 20khz, as I wrote earlier, these are only harmonics and overtones. This can be seen on the spectra. Regarding oversampling, it is noticeable to me by ear, as well as to many other people. You can check for yourself. Take the same source file, encode it in DSD 64, 128, 256, 512, 1024 and further, listen ideally on the DAC and hear the changes. Or the same with WAV 44.1, 88,2, 176,4, 352,8, 705,6, 1411,2. Or by a 48x2 x32 conversion series.
The transformation can be done with any technique, no matter what it is. The DAC, the Player, the Converter will all work.