Hi Res music

[Miguelón]

Would you please confirm the validity of the following argument. Thank you.

Sampling rates higher than 44.1 kHz are used to preserve audio in the band above 22.05 kHz. From the format of the audio file, it is then easy to infer what band the audio will be stored in a Hi-Res Audio recording, because accurate reconstruction of a continuous, frequency-limited signal from its samples is possible if the sample rate was higher than twice the highest harmonic component of the sampled signal. Then there is the bit depth, which defines the maximum usable dynamic range.

More or less the frequency part is the Nyquist theorm applied to acoustics, you should use sample rates twice the highest frequency you want to reconstruct. 96 kHz allows to encode 48 kHz.

More or less because when you approach Nyquist limit, that’s it if you try to encode 99 Hz with 200 Hz sample rate you will find modulation sounds.

One can define a “passband”, in the case of 44.1 kHz is the frequencies between 20 kHz and 22 kHz that are not well encoded. It was suggested is a little bit narrow for filtering and avoid aliasing or imaging, and can be improved with 48 kHz files.

With some filters one can listen slight alterations on very high frequencies, especially young people I guess. I don’t really know so much about the relevance of 48 kHz on most people, the passband in this case is 20 to 24 kHz. I cannot hear the difference with filters neither.

As Anticollinet told, 24 bits also low quantization noise, but you’re right: in some action films for example it can get better peak SPL as explosions or other effects, increasing the disposable intensity range

 

[antcollinet]

Sampling rates higher than 44.1 kHz are used to preserve audio in the band above 22.05 kHz

Above 20kHz really. The space between 20 and 22.05kHz is needed for the reconstruction filter roll off. We don't typically have perfect brick wall filtering.

From the format of the audio file, it is then easy to infer what band the audio will be stored in a Hi-Res Audio recording, because accurate reconstruction of a continuous, frequency-limited signal from its samples is possible if the sample rate was higher than twice the highest harmonic component of the sampled signal.

Sort of. We can infer the highest audio frequency a file is capable of recording - as you say, up to half the sample rate. Thing is though - as well as us being unable to hear above typically 20Khz, there is little to no musical content above that point either. There is really very little point of sample rates higher than 48kHz (The increase from 44.1 to 48 creates more space for a typical filter to roll off to good levels of attenuation - though whether this really gives an audible benefit is debatable)

EDIT - or what @Miguelón said.

Then there is the bit depth, which defines the maximum usable dynamic range.

Correct.

 

[Casper]

I really enjoy Qobuz and it integrates well with Volumio which runs well on a Raspberry Pi. They have a great selection of jazz and classical as well as just about everything else. I don't remember ever not being able to find what I was looking for.

 

[Miguelón]

I really enjoy Qobuz and it integrates well with Volumio which runs well on a Raspberry Pi. They have a great selection of jazz and classical as well as just about everything else. I don't remember ever not being able to find what I was looking for.

Love Qobuz for classical and jazz too!

But when I try pop/rock playlists one discover they contain a strange unexpected distortion: it is the so called “french pop”. I tried to apply AI filters to jump automatically to the next track but was in vain

(sarcasm, they have interesting singer-songwriters as Jacques Brel and Georges Brassens with social content and ironic humor, also rap is well developed in France)

 

[dualazmak]

I assume it would be worthwhile sharing here on this thread about my personal "Summary of rationales for ""on-the-fly (real-time)"" conversion of all music tracks (including 1 bit DSD tracks) into 88.2 kHz or 96 kHz PCM format for DSP (XO/EQ) processing"; if you would be interested, and for your possible reference, please visit my post #532 on my audio project thread. I pasted the contents of that post under the below spoiler cover.

Spoiler: Summary of my rationales for "on-the-fly (real-time)" conversion of all music tracks (including 1 bit DSD tracks) into 88.2 kHz or 96 kHz PCM

My present answer for you is "It is quite feasible enough and even ""needed"" to feed all the audio digital signals in 88.2 kHz or 96 kHz PCM (or 192 kHz, if you like) by JRiver's on-the-fly format conversion to be sent into DSP (XO/EQ) software EKIO."

Various background and justifications for this answer are as follows;

Before starting this project, I had been enjoying music with ordinary PC audio setup with one DAC (OPPO Sonica DAC)) and one HiFi integrated amplifier (ACCUPHASE E-460) driving all the SPs through passive LC (inductors capacitors resistors) network. And I had been sticking to "native format feed" into OPPO Sonica DAC up to 1-bit/DSD256(4x), as you kindly pointed.

When I started considering possible multichannel multi-driver multi-way multi-amplifier project with software DSP (XO/EQ), I did intensive search and desk evaluations on various DSP software solutions, and I found the maximum PCM processing format is 192 kHz 24 bit in these DSP software solutions. (Even with the extraordinary expensive TRINNOV ALTITUDE 32 DSP processor, actually having PC in it, the internal DSP processing is up to 192 kHz).

I carefully considered the pros and cons of "DSP processing all tracks in 192 kHz or 96kHz" instead of "native format feed", and concluded that multichannel multi-amplifier approach would surpass the cons, at least in my system setup with still amazingly wonderful Yamaha SP drivers and cabinet.

Consequently, I decided to go into "multichannel multi-amplifier" world of "max. 192 kHz 24 bit processing", as you kindly have read through this project, including the "all in max. 192 kHz ASIO I/O within PC".

Then, rather recently, I (we) fully discussed and evaluated the UHF (ultra-high frequency) noise issue in poorly QC-ed HiRes music tracks including DSD formats, as you clearly noticed;
- "Near ultrasound - ultrasound" ultra-high frequency (UHF) noises in improperly engineered/processed HiRes music tracks, and EKIO's XO-EQ configuration to cut-off such noises: #362-#386, #518
I wrote that such a high amount of UHF noises would be "possibly" harmful (and useless, meaningless) for our tweeters and super tweeters. I also pointed they would be highly possibly harmful for our beloved pets including dogs, cats, birds.

Having my intensive objective measurements of these "poorly QC-ed" HiRes tracks, and having so many intensive discussions on "enough PCM sampling rate in HiFi audio", now I conclude that 88.2 kHz or 96 kHz processing (i.e. up to 44.1 kHz or 48 kHz in L and R channels) would be just enough and feasible in my setup (and I believe so also in your setup) since I decided always having high-cut (low-pass) -48 dB/Oct filters at 25 kHz in my EKIO configuration to cut-off any of the possible UHF noises very frequently existing in HiRes tracks.

This means that I have finally landed on agreement with @mikessi's "enlightenment and belief" of "There is really no audible benefit to playback beyond 24/96 sampling, especially with any recordings other that those done with the most advanced high res gear and high fidelity values."　

Another important aspect of this issue would be relating to our hearing ability in high frequency zones. Recently, I participated in the interesting thread entitled "Audio Listening With Age Diminished Hearing". You would please read my posts #70, #72 and #74 on that thread.

BTW, as I wrote here, here and here, my digital music library of about 25,000 files consists of mixture of various formats;

16-bit/44.1kHz CD ripped non-compressed aif (majority!),
24-bit/192kHz down-sampled or up-sampled aif,
24-bit/96kHz flac,
24-bit/192kHz flac,
1-bit/DSD64(1x) 2.8MHz dsf,
1-bit/DSD128(2x) 5.6 MHz dsf,
1-bit/DSD256(4x) 11.2 MHz dsf,

and now JRiver MC feeds all of the tracks usually (mainly) in 88.2 kHz 24 bit (i.e. max. 44.1 kHz Fq window in 2-ch stereo) by on-the-fly conversion into EKIO for crossover/EQ processing. As I have high-cut (low-pass) -48 dB/Oct LR filters at 25 kHz, max. 44.1 kHz in L & R channels are more than enough.

 

[Miguelón]

I assume it would be worthwhile sharing here on this thread about my personal "Summary of rationales for ""on-the-fly (real-time)"" conversion of all music tracks (including 1 bit DSD tracks) into 88.2 kHz or 96 kHz PCM format for DSP (XO/EQ) processing"; if you would be interested, and for your possible reference, please visit my post #532 on my audio project thread. I pasted the contents of that post under the below spoiler cover.

Spoiler: Summary of my rationales for "on-the-fly (real-time)" conversion of all music tracks (including 1 bit DSD tracks) into 88.2 kHz or 96 kHz PCM

My present answer for you is "It is quite feasible enough and even ""needed"" to feed all the audio digital signals in 88.2 kHz or 96 kHz PCM (or 192 kHz, if you like) by JRiver's on-the-fly format conversion to be sent into DSP (XO/EQ) software EKIO."

Various background and justifications for this answer are as follows;

Before starting this project, I had been enjoying music with ordinary PC audio setup with one DAC (OPPO Sonica DAC)) and one HiFi integrated amplifier (ACCUPHASE E-460) driving all the SPs through passive LC (inductors capacitors resistors) network. And I had been sticking to "native format feed" into OPPO Sonica DAC up to 1-bit/DSD256(4x), as you kindly pointed.

When I started considering possible multichannel multi-driver multi-way multi-amplifier project with software DSP (XO/EQ), I did intensive search and desk evaluations on various DSP software solutions, and I found the maximum PCM processing format is 192 kHz 24 bit in these DSP software solutions. (Even with the extraordinary expensive TRINNOV ALTITUDE 32 DSP processor, actually having PC in it, the internal DSP processing is up to 192 kHz).

I carefully considered the pros and cons of "DSP processing all tracks in 192 kHz or 96kHz" instead of "native format feed", and concluded that multichannel multi-amplifier approach would surpass the cons, at least in my system setup with still amazingly wonderful Yamaha SP drivers and cabinet.

Consequently, I decided to go into "multichannel multi-amplifier" world of "max. 192 kHz 24 bit processing", as you kindly have read through this project, including the "all in max. 192 kHz ASIO I/O within PC".

Then, rather recently, I (we) fully discussed and evaluated the UHF (ultra-high frequency) noise issue in poorly QC-ed HiRes music tracks including DSD formats, as you clearly noticed;
- "Near ultrasound - ultrasound" ultra-high frequency (UHF) noises in improperly engineered/processed HiRes music tracks, and EKIO's XO-EQ configuration to cut-off such noises: #362-#386, #518
I wrote that such a high amount of UHF noises would be "possibly" harmful (and useless, meaningless) for our tweeters and super tweeters. I also pointed they would be highly possibly harmful for our beloved pets including dogs, cats, birds.

Having my intensive objective measurements of these "poorly QC-ed" HiRes tracks, and having so many intensive discussions on "enough PCM sampling rate in HiFi audio", now I conclude that 88.2 kHz or 96 kHz processing (i.e. up to 44.1 kHz or 48 kHz in L and R channels) would be just enough and feasible in my setup (and I believe so also in your setup) since I decided always having high-cut (low-pass) -48 dB/Oct filters at 25 kHz in my EKIO configuration to cut-off any of the possible UHF noises very frequently existing in HiRes tracks.

This means that I have finally landed on agreement with @mikessi's "enlightenment and belief" of "There is really no audible benefit to playback beyond 24/96 sampling, especially with any recordings other that those done with the most advanced high res gear and high fidelity values."　

Another important aspect of this issue would be relating to our hearing ability in high frequency zones. Recently, I participated in the interesting thread entitled "Audio Listening With Age Diminished Hearing". You would please read my posts #70, #72 and #74 on that thread.

BTW, as I wrote here, here and here, my digital music library of about 25,000 files consists of mixture of various formats;

16-bit/44.1kHz CD ripped non-compressed aif (majority!),
24-bit/192kHz down-sampled or up-sampled aif,
24-bit/96kHz flac,
24-bit/192kHz flac,
1-bit/DSD64(1x) 2.8MHz dsf,
1-bit/DSD128(2x) 5.6 MHz dsf,
1-bit/DSD256(4x) 11.2 MHz dsf,

and now JRiver MC feeds all of the tracks usually (mainly) in 88.2 kHz 24 bit (i.e. max. 44.1 kHz Fq window in 2-ch stereo) by on-the-fly conversion into EKIO for crossover/EQ processing. As I have high-cut (low-pass) -48 dB/Oct LR filters at 25 kHz, max. 44.1 kHz in L & R channels are more than enough.

I didn’t understand half of arguments but for simplicity, the improvement of upsampling is just for the mentioned software or concerns all DSPs precessing?

 

[dualazmak]

I didn’t understand half of arguments but for simplicity, the improvement of upsampling is just for the mentioned software or concerns all DSPs precessing?

Of course, I have been never expecting sound quality improvement for upsampling of 44.1 kHz digital music tracks into 88.2 kHz or 96 kHz.

As I shared in my post under the spoiler cover, about 93 % of my around 28,000-track digital music library in SSD storage are CD-ripped intact 44.1 kHz 16 bit non-compressed PCM format (I prefer AIFF for flexible/reliable tag info and cover-art embedding; ref. here for my digital music library organization).

The rest of the library, i.e. 7 %, consists of higher Hi-Res tracks including 96 or 192 kHz AIFF (including digitized LPs), 96 kHz or 192 kHz FLAC, and 1-bit DSD of 2.8 or 5.6 or 11.2 MHz.

I really would like to feed all of my digital tracks, as well as other streaming audio including YouTube listening with web browser, into DSP EKIO through system-wide one-stop ASIO/VASIO/VAIO routing center "VB-Audio MATRIX" (ref. here for my setup) in common sampling rate without changing the sampling rate parameters in music player (I use JRiver MC), VB-Audio MATRIX and DSP EKIO. All the sound signals from the PC is, therefore, fed into DSP EKIO for DSP processing, and then 8-CH multichannel DAC process is done by OKTO DAC8PRO into multi-amplifier multi-SP-driver setup; DSP-EKIO, therefore, serves as system-wide one-stop DSP center. (If you would be interested, please refer here for the details of my latest system setup.)

Consequently, nowadays, I usually prefer feeding all the library tracks and other streaming audio channel in 88.2 kHz (rather than 96 kHz) 24 bit by JRiver's (or VB-MATRIX's) on-the-fly sampling-rate conversion. As you may agree, 88.2 kHz is well-high-rate-enough for me as described!

 

[Booker]

Suppose the sampling rate is twice that of CD, say 96kHz. Should I expect more stored information in the audible frequency range up to 20kHz? I should not, correct? Thanks.

 

[staticV3]

Suppose the sampling rate is twice that of CD, say 96kHz. Should I expect more stored information in the audible frequency range up to 20kHz? I should not, correct? Thanks.

Correct.

 

[Miguelón]

Anybody with some knowledge about MQA files from Tidal? Not sure what they do with this masters, sound a little bit different than CD in EQ. They state that the artist verifies the master after compression to ensure music quality but I don’t see really any advantage of doing so, from (for example) a regular CD album.

Are different masters?