Proper Definition of High-Resolution Music

[watchnerd]

Oh the scary part is with a good enough recording Bieber would do the trick.

Stimulated against my will?

#metoo?

But if past experience in an indicator, the answer is still no. I never got into other audiophile "standards" like Jennifer Warnes "Famous Blue Raincoat", no matter how much audio reviewers wet themselves over the quality of the recording.

 

[Midwest Blade]

A good clip of Dr. Mark Waldrep's discussion of what is Hi Res. I follow his blog and have a few of his Blu Ray discs, after listening in his space a couple of Axpona shows ago it was quite an audio experience. My views and systems have changed but I am a firm believer in well produced and engineered 16/44 vs much of the other gobeliegook.

 

[Blumlein 88]

A good clip of Dr. Mark Waldrep's discussion of what is Hi Res. I follow his blog and have a few of his Blu Ray discs, after listening in his space a couple of Axpona shows ago it was quite an audio experience. My views and systems have changed but I am a firm believer in well produced and engineered 16/44 vs much of the other gobeliegook.

I assume you have read about Dr. Waldrep's listening test of 44/16 vs his 96/24 recordings. Results don't show the difference is audible. The big difference, far larger than hires or not is how well the recording was done in all the other ways that go into a recording. I'd say there are no acoustic recordings that can exceed the dynamic limits of dithered 16 bit. A super tiny, tiny, tiny number that could even need 96 db.

 

[JJB70]

I could listen to Mark Waldrep and read his writing for hours, his blog is always worth reading. It displayed a lot of integrity to take the results of the listening test that indicated no audible advantage of 96/24 recordings on the chin and not try and invent excuses or reasons why listeners were wrong given his previous views on the matter.

 

[Midwest Blade]

I assume you have read about Dr. Waldrep's listening test of 44/16 vs his 96/24 recordings. Results don't show the difference is audible. The big difference, far larger than hires or not is how well the recording was done in all the other ways that go into a recording. I'd say there are no acoustic recordings that can exceed the dynamic limits of dithered 16 bit. A super tiny, tiny, tiny number that could even need 96 db.

Followed the test and have his book, biggest problem we all face is finding suitable source material.

 

[digicidal]

To me I think the problems with resolution as far as recording quality is concerned are the exact same as on the playback side (or more accurately an inverse). Which makes sense as it's the sound-in-air problem on both sides. It's likely that the acoustics of the venue/studio play a much more significant part than the technology responsible for recording it.

After all... if you have a 'perfect' recording chain, but the conditions are noisy - then you're likely just capturing the noise better. Same with playback, if it's perfectly transparent and has 120db of DR, but your listening room isn't completely silent - then it's quite possible that you now can't hear half the music in the recording. So either you don't listen at all, or you use loudness controls to eliminate most/all of the 'benefit' that you achieved.

And for what it's worth, Bieber singing live in front of me wouldn't do it... so I'm pretty sure it's not a problem of resolution. In fact, I think he actually sounds best at around 1db SNR.

 

[Blumlein 88]

To me I think the problems with resolution as far as recording quality is concerned are the exact same as on the playback side (or more accurately an inverse). Which makes sense as it's the sound-in-air problem on both sides. It's likely that the acoustics of the venue/studio play a much more significant part than the technology responsible for recording it.

After all... if you have a 'perfect' recording chain, but the conditions are noisy - then you're likely just capturing the noise better. Same with playback, if it's perfectly transparent and has 120db of DR, but your listening room isn't completely silent - then it's quite possible that you now can't hear half the music in the recording. So either you don't listen at all, or you use loudness controls to eliminate most/all of the 'benefit' that you achieved.

And for what it's worth, Bieber singing live in front of me wouldn't do it... so I'm pretty sure it's not a problem of resolution. In fact, I think he actually sounds best at around 1db SNR.

The biggest thing I think is the SNR. It appears there may be no recordings with more than something in the mid 70 db range. Even breaking it up into our most sensitive areas vs noise it appears you aren't going to get past the mid 90 db range. So dithered 16 bit is enough. Fewer than 1% of all recordings will get close to that.

 

[digicidal]

The biggest thing I think is the SNR. It appears there may be no recordings with more than something in the mid 70 db range. Even breaking it up into our most sensitive areas vs noise it appears you aren't going to get past the mid 90 db range. So dithered 16 bit is enough. Fewer than 1% of all recordings will get close to that.

And few output devices will take advantage of it as well. That's one of my biggest dilemmas... as someone with a more modest budget for audio gear and room treatments... DSP is the simplest way to alleviate necessary compromises in listening environment. However, with all but the most expensive gear there is a limitation of processing capabilities which is surpassed as the sampling rate increases.

Maybe in a world with a direct neural interface to the music, this would be a much bigger deal. However, even in that fantasy world it's not likely - since although I might be able to 'hear' something in the 20kHz-40kHz range in my brain - we don't have instruments that play notes that high. So at most I'm getting a little better sense of 'air' or whatever (mostly harmonic noise).

In the real world however, I have these amazing but ultimately crappy (when compared to 'hi res' audio) ears... which are my brain's only interface with the music itself.

 

[Krunok]

we don't have instruments that play notes that high.

Actually we do. Electric keyboard organ can easilly be adjusted to play in 20-40kHz range. The only reason it is limited is we wouldn't hear it and not because of it's inability to play high notes.

 

[Peter Leyenaar]

This is what I am talking about. Then, how can we evaluate the quality of recording / mastering?

Use your ears, you will only hear 18KHz max. anyway

 

[digicidal]

Actually we do. Electric keyboard organ can easilly be adjusted to play in 20-40kHz range. The only reason it is limited is we wouldn't hear it and not because of it's inability to play high notes.

Although pointlessly argumentative IMO, you are completely correct. And depending on the definition of 'instrument' (and 'music' for that matter) entirely possible without adjustment:

You succeeded in casting light on my oversimplification! I guess if you're modifying that keyboard to play ultra-sonics... you'd also better add another 88 keys to it while you're at it or no one will know you're even playing the damn thing! Although could be some fun times with the family dog...

Out of curiosity, what's the notation for a note 7 octaves above middle C?

 

[Pluto]

how can we evaluate the quality of recording / mastering?

It is no more possible to evaluate, precisely, a music recording than it is to quantify the pleasure of viewing a painting by an Old Master. In any case, are the same criteria valid for evaluating a record of baroque chamber music as punk rap? I doubt it. We have those genres where the “duty” of the recording is, arguably, to be as transparent as possible and convey the listener to the best seat in the concert venue, and we have those in which the recording itself is a significant component of the art.

There is absolutely no indication, of which I am aware, that any of the “number wars” criteria (such as bit depth, sample rate etc.) make any significant difference to the end user and I would be happy to prepare material to demonstrate this if a suitable methodology can be agreed upon. I guess the role of the record-reviewing journalist hasn't had its day just yet!

 

[hetzer]

Use your ears, you will only hear 18KHz max. anyway

Actually, its 17khz
But I can still figure out whether this album is heavily compressed or not.

 

[digicidal]

It is no more possible to evaluate, precisely, a music recording than it is to quantify the pleasure of viewing a painting by an Old Master. In any case, are the same criteria valid for evaluating a record of baroque chamber music as punk rap? I doubt it. We have those genres where the “duty” of the recording is, arguably, to be as transparent as possible and convey the listener to the best seat in the concert venue, and we have those in which the recording itself is a significant component of the art.

There is absolutely no indication, of which I am aware, that any of the “number wars” criteria (such as bit depth, sample rate etc.) make any significant difference to the end user and I would be happy to prepare material to demonstrate this if a suitable methodology can be agreed upon. I guess the role of the record-reviewing journalist hasn't had its day just yet!

I might disagree with the minutiae of that first sentence - there are objective ways to quantify values of both (and many like dynamic range apply whether visual or aural). However, to your second point, being more the subjective evaluation... I agree 100%. I have good hearing (for my age at least) and can still hear 17kHz tones fine. I can, on certain selections, reliably differentiate between 320kbps MP3 and lossless... On anything I actually enjoy listening to however - it drops to a statistical insignificance and becomes guessing.

I can appreciate (technically speaking) a recording with significant dynamic range and spatial separation of instruments, but what I spend 90% of my time listening to is compressed recordings of 4-5 instruments - mostly amplified with effects/distortion - with hardly any range/separation at all! Add to that a problematic listening environment and gear and it's much worse. Am I missing something measurable? Definitely. Could I hear those missing elements in an ABX? Possibly, though less likely. Does that significantly impede my ability to enjoy listening? Not at all.

Measurements and specifications (once validated) are fantastic for choosing equipment that doesn't add something undesirable - they're utterly pointless however, in quantifying the enjoyment of a particular performance/genre. Considering how long it took for recordings to even fully utilize 16bit resolution... I'd say we've got another 30-40 years minimum before an argument could be made for a requirement for 24bit and higher. And even that won't settle the debate over whether or not any of the 'improvements' are significantly audible even under the best conditions.

 

[Krunok]

I guess if you're modifying that keyboard to play ultra-sonics... you'd also better add another 88 keys to it while you're at it or no one will know you're even playing the damn thing!

Well, you are not the first one here with the idea to interface audio nerve - it has been discusssed several times before.
You don't neeed 88 more keys, some foot pedals to shift several octave bands would do just fine, just like pipe organ has.

Out of curiosity, what's the notation for a note 7 octaves above middle C?

That would be C11.

 

[Soniclife]

Considering how long it took for recordings to even fully utilize 16bit resolution... I'd say we've got another 30-40 years minimum before an argument could be made for a requirement for 24bit and higher.

I'm not at all convinced we have music that needs 16 bit dynamic range now, outside of the odd test or very niche recording, and I don't see it changing, it's great to have the technology capable of doing it, and it makes sense to record in it, but playback we don't need it.

 

[digicidal]

I'm not at all convinced we have music that needs 16 bit dynamic range now, outside of the odd test or very niche recording, and I don't see it changing, it's great to have the technology capable of doing it, and it makes sense to record in it, but playback we don't need it.

Totally agree with this. It's similar with the video side of things... it makes sense from an archival accuracy perspective to capture video in 8K at 120Hz (or even higher)... but not on the playback side, where it's an imperceptible difference that creates other problems (transmission bandwidth, processing requirements, etc.). A little different naturally, since you also have the option of zooming or slowing down video at higher than perceptible speed/resolution. In either case (audio/video) there are some advantages on the studio/production side of things but very few when it comes down to consuming the final product.

Unless we evolve better hearing, in anything targeting human beings, it seems a pursuit with significantly diminishing returns at the very least. Then again we're developing content for dogs and cats now... so maybe they'll benefit.

 

[Sergei]

You are probably aware of a rather long thread discussing, among other things, the technical aspects of hi-res: https://www.audiosciencereview.com/...qa-creator-bob-stuart-answers-questions.7623/.

I narrowed down that thread to another one, and then this newer thread to a single post, which I believe captured the core issue:
https://www.audiosciencereview.com/...higher-sampling-rates.7939/page-2#post-194272

IMHO, the core issue is that, formally speaking, a limited-duration non-perfectly-periodic piece of music can't be Fourier-transformed to a finite spectrum representation, and thus the Nyquist–Shannon Sampling Theorem, which relies upon Fourier transform, strictly speaking is not applicable to music.

Thus, we must assume that a sampled digital representation of real-life music contains distortions, as compared to the original analog sound. Mathematically, increasing the sample rate and number of bits per sample makes this distortion lower.

Most of the time listeners can't perceive the difference of distortions between 44/16 and, say, 192/24. Sometimes they can: A Meta-Analysis of High Resolution Audio Perceptual Evaluation.

The 5% figure keeps coming up. Like: the difference can only be perceived by 5% of listeners, on 5% of music. A naive approach is to multiply these probabilities: the thinking goes that on average ~0.25% of listening sessions would be affected by the regular vs high fidelity differences.

But, and there is a big but! For a particular listener in the 5%, whose favorite music genres happen to fall in the 5% too, the number of affected listening sessions can be much higher, closer to 100% actually.

And vice versa, for a listener in the 95%, or for a listener whose favorite music genres happen to be in the 95%, the advantages of hi-res are immaterial. For them, the promise of hi-res is 100% snake oil.

 

[Peter Leyenaar]

You are probably aware of a rather long thread discussing, among other things, the technical aspects of hi-res: https://www.audiosciencereview.com/...qa-creator-bob-stuart-answers-questions.7623/.

I narrowed down that thread to another one, and then this newer thread to a single post, which I believe captured the core issue:
https://www.audiosciencereview.com/...higher-sampling-rates.7939/page-2#post-194272

IMHO, the core issue is that, formally speaking, a limited-duration non-perfectly-periodic piece of music can't be Fourier-transformed to a finite spectrum representation, and thus the Nyquist–Shannon Sampling Theorem, which relies upon Fourier transform, strictly speaking is not applicable to music.

Thus, we must assume that a sampled digital representation of real-life music contains distortions, as compared to the original analog sound. Mathematically, increasing the sample rate and number of bits per sample makes this distortion lower.

Most of the time listeners can't perceive the difference of distortions between 44/16 and, say, 192/24. Sometimes they can: A Meta-Analysis of High Resolution Audio Perceptual Evaluation.

The 5% figure keeps coming up. Like: the difference can only be perceived by 5% of listeners, on 5% of music. A naive approach is to multiply these probabilities: the thinking goes that on average ~0.25% of listening sessions would be affected by the regular vs high fidelity differences.

But, and there is a big but! For a particular listener in the 5%, whose favorite music genres happen to fall in the 5% too, the number of affected listening sessions can be much higher, closer to 100% actually.

And vice versa, for a listener in the 95%, or for a listener whose favorite music genres happen to be in the 95%, the advantages of hi-res are immaterial. For them, the promise of hi-res is 100% snake oil.

Okidoki

 

[Blumlein 88]

You are probably aware of a rather long thread discussing, among other things, the technical aspects of hi-res: https://www.audiosciencereview.com/...qa-creator-bob-stuart-answers-questions.7623/.

I narrowed down that thread to another one, and then this newer thread to a single post, which I believe captured the core issue:
https://www.audiosciencereview.com/...higher-sampling-rates.7939/page-2#post-194272

IMHO, the core issue is that, formally speaking, a limited-duration non-perfectly-periodic piece of music can't be Fourier-transformed to a finite spectrum representation, and thus the Nyquist–Shannon Sampling Theorem, which relies upon Fourier transform, strictly speaking is not applicable to music.

Thus, we must assume that a sampled digital representation of real-life music contains distortions, as compared to the original analog sound. Mathematically, increasing the sample rate and number of bits per sample makes this distortion lower.

Most of the time listeners can't perceive the difference of distortions between 44/16 and, say, 192/24. Sometimes they can: A Meta-Analysis of High Resolution Audio Perceptual Evaluation.

The 5% figure keeps coming up. Like: the difference can only be perceived by 5% of listeners, on 5% of music. A naive approach is to multiply these probabilities: the thinking goes that on average ~0.25% of listening sessions would be affected by the regular vs high fidelity differences.

But, and there is a big but! For a particular listener in the 5%, whose favorite music genres happen to fall in the 5% too, the number of affected listening sessions can be much higher, closer to 100% actually.

And vice versa, for a listener in the 95%, or for a listener whose favorite music genres happen to be in the 95%, the advantages of hi-res are immaterial. For them, the promise of hi-res is 100% snake oil.

For the 5% with the 5% recordings what is the perceptual quality difference? If rating on a scale of 0-100% with higher being max fidelity how much is lost without that last 5%? Does it drop you to 95% of what is possible? Does it have an overly large effect and cause your perception of the recording quality to drop to 80%?