Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Saving music from destruction: upsampling (Read 29361 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Saving music from destruction: upsampling

This article starts off well , with a relatively sane (for audiophile-speak) statement of what good upsampling can and cannot do:

http://www.pstracks.com/av/ham-fisted-upsampling/

but then we get to David Chesky on the defense against accusations that his HDtracks download service is doing 'ham fisted' or pointless upsampling:

Quote
David responds. “Why do we sell hi res files at 192/24 when there is no musical energy up at 96khz? Lets look at it like this. In the current state of digital technology we do not have A/D chipsets that work at 44.1/16 that do not have pre and post ringing. This ringing destroys the music in the audio band. Audiophile labels do not record at 192/24 to try to capture what many think is air up over 50K. They do this as to keep the nasty anti-aliasing digital A/D filters as far away from the audio band as they can. This insures less corruption down where the music is. The sound becomes much more analog and relaxed. Think of  pre and post ringing like this. You are riding down a street in a Ferrari at 180 MPH over a cobblestone road and someone gives you a camera and ask you to take some photos out of the windows. All these little cobblestones will cause bumps and the camera will not be steady and you will loose the focus. This is what happens in current A/D conversion as well at 44.1/16. Doubling or quadrupling that rate helps smooth the analogous ride."


Emphasis mine.

See any listening test data in there anywhere?

Nah, me neither.  Nor any reason why 192kHz versus , say, 96 or 88.2*.  Just another inane analogy to sports cars and cameras.

It never ends, does it?


* (if I recall my Dan Lavry whitepapers correctly, the possible negative effects from pre- and post-ringing are taken care of at a SR of ~60kHz)

Saving music from destruction: upsampling

Reply #1
* (if I recall my Dan Lavry whitepapers correctly, the possible negative effects from pre- and post-ringing are taken care of at a SR of ~60kHz)


Not to mention that even going to 60kHz is irrelevant for distribution and playback. An oversampling DAC avoids the issue directly, even for 44.1kHz.

Saving music from destruction: upsampling

Reply #2
David's response = audiophile FUD and a weak attempt to distract from their mistakes
"I hear it when I see it."

Saving music from destruction: upsampling

Reply #3
This article starts off well , with a relatively sane (for audiophile-speak) statement of what good upsampling can and cannot do:

http://www.pstracks.com/av/ham-fisted-upsampling/

but then we get to David Chesky on the defense against accusations that his HDtracks download service is doing 'ham fisted' or pointless upsampling:

Quote
David responds. “Why do we sell hi res files at 192/24 when there is no musical energy up at 96khz? Lets look at it like this. In the current state of digital technology we do not have A/D chipsets that work at 44.1/16 that do not have pre and post ringing. This ringing destroys the music in the audio band. Audiophile labels do not record at 192/24 to try to capture what many think is air up over 50K. They do this as to keep the nasty anti-aliasing digital A/D filters as far away from the audio band as they can. This insures less corruption down where the music is. The sound becomes much more analog and relaxed. Think of  pre and post ringing like this. You are riding down a street in a Ferrari at 180 MPH over a cobblestone road and someone gives you a camera and ask you to take some photos out of the windows. All these little cobblestones will cause bumps and the camera will not be steady and you will loose the focus. This is what happens in current A/D conversion as well at 44.1/16. Doubling or quadrupling that rate helps smooth the analogous ride."


Emphasis mine.

See any listening test data in there anywhere?

Nah, me neither.  Nor any reason why 192kHz versus , say, 96 or 88.2*.  Just another inane analogy to sports cars and cameras.

It never ends, does it?


* (if I recall my Dan Lavry whitepapers correctly, the possible negative effects from pre- and post-ringing are taken care of at a SR of ~60kHz)


Consider the source.  For another audiophile recording source that has gone off completely into the weeds, please see Mapleshade records:

Audiophile recordings and Tweeks, all in one online store!

Saving music from destruction: upsampling

Reply #4
Consider the source.  For another audiophile recording source that has gone off completely into the weeds, please see Mapleshade records:

Audiophile recordings and Tweeks, all in one online store!


Wow...

Quote
The Mapleshade Records website features free MP3 samples and downloads. You'll be very pleasantly surprised by the sound quality of these MP3s, probably the best sounding 128k MP3s you've heard any where on the net. They are encoded with a Mapleshade proprietary signal path and process that we call Maplestreaming, a process that yields startling dynamics, stereo detail and a wonderfully resolved sound stage with ambience and depth you'd never expect from an MP3.

Saving music from destruction: upsampling

Reply #5
Wow...

Quote
The Mapleshade Records website features free MP3 samples and downloads. You'll be very pleasantly surprised by the sound quality of these MP3s, probably the best sounding 128k MP3s you've heard any where on the net. They are encoded with a Mapleshade proprietary signal path and process that we call Maplestreaming, a process that yields startling dynamics, stereo detail and a wonderfully resolved sound stage with ambience and depth you'd never expect from an MP3.


Just another commercial trick. They have encoded mono-like easy-compressible source into 128 kbps. Scam.


Saving music from destruction: upsampling

Reply #6
Upsampling does not remove the need for a sharp filter. The upsampling itself must use that sharp filter. So what are they talking about here? Allowing aliasing, and by doing that making the interchannel time resolution worse?
-----
J. D. (jj) Johnston

Saving music from destruction: upsampling

Reply #7
Quote
David responds. “... Think of  pre and post ringing like this. You are riding down a street in a Ferrari at 180 MPH over a cobblestone road and someone gives you a camera and ask you to take some photos out of the windows. All these little cobblestones will cause bumps and the camera will not be steady and you will loose the focus. This is what happens in current A/D conversion as well at 44.1/16. Doubling or quadrupling that rate helps smooth the analogous ride."


So what would the doubling or quadrupling be in the Ferrari analogy? Driving 720 mph instead of 180? What a stupid comparison. What he actually describes is a visualization of dither*... which can be a good thing...

Quote
Just another commercial trick. They have encoded mono-like easy-compressible source into 128 kbps. Scam.

The tao1.mp3 demo on that page is quite stereo, but more than 95% of the time the bandwidth is limited to 13 kHz. Probably audibly low-pass to quite a few people when compared to the lossless source.

Chris

* or recording noise in general
If I don't reply to your reply, it means I agree with you.

Saving music from destruction: upsampling

Reply #8
Upsampling does not remove the need for a sharp filter. The upsampling itself must use that sharp filter. So what are they talking about here? Allowing aliasing, and by doing that making the interchannel time resolution worse?


Wow, I never thought about that little twist!  ;-)

I guess the answer to your question is that upsampling is a way of making the playback of a file less dependent on the quality of the brick wall filter in the playback device. The brick wall filter in the upsampler tends to supercede the brick wall filter in the playback device since the latter generally is at a far higher frequency.

Saving music from destruction: upsampling

Reply #9
Wow, I never thought about that little twist!  ;-)

I guess the answer to your question is that upsampling is a way of making the playback of a file less dependent on the quality of the brick wall filter in the playback device. The brick wall filter in the upsampler tends to supercede the brick wall filter in the playback device since the latter generally is at a far higher frequency.


I doubt the upsampling algorythm is using a lowpass with far higher frequency as the build in the DAC one. I even wonder if many of these upsamplers use lower corner frequencies as a default DAC when decoding 44.kHz to prevent any pre-ringing.
You shouldn´t also forget when you upsampled lets say from 44.1 to 96kHz that besides the ~20kHz in the upsampling routine another lowpass is applied in the DAC at ~40kHz at playback.
So with upsampling you trade the single 20kHz LP in the DAC for 44.1kHz material against a 20kHz LP in the upsampler + a second LP in the DAC at higher frequency.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Saving music from destruction: upsampling

Reply #10
Wow, I never thought about that little twist!  ;-)

I guess the answer to your question is that upsampling is a way of making the playback of a file less dependent on the quality of the brick wall filter in the playback device. The brick wall filter in the upsampler tends to supercede the brick wall filter in the playback device since the latter generally is at a far higher frequency.


I doubt the upsampling algorythm is using a lowpass with far higher frequency as the build in the DAC one.


I wouldn't expect a higher corner frequency, but I might expect a better filter as regards to in-band ripple, sharpness of cutoff, phase and therefore impulse response, etc.

Quote
I even wonder if many of these upsamplers use lower corner frequencies as a default DAC when decoding 44.kHz to prevent any pre-ringing.


You can read all about that, as illustrated by real world examples, here:

Sample rate conversion technical tests

Quote

You shouldn´t also forget when you upsampled lets say from 44.1 to 96kHz that besides the ~20kHz in the upsampling routine another lowpass is applied in the DAC at ~40kHz at playback.
So with upsampling you trade the single 20kHz LP in the DAC for 44.1kHz material against a 20kHz LP in the upsampler + a second LP in the DAC at higher frequency.
[/quote

The effects of the 48 KHz brick wall on operations at 22 KHz should be pretty small.

Saving music from destruction: upsampling

Reply #11
I think there's two different issues.

The first is taking a 44.1kHz or 48kHz digital recording, and upsampling it, and then charging people for the privilege. Well, if you tell them what you've done, why not? Let them decide whether they want to pay you for that service, or buy the CD and upsample it themselves, or just play it as normal (which will oversample it anyway!). Whereas if you claim your result is somehow natively 96kHz or 192kHz or whatever, it's a scam. Plain and simple.

The second is taking an analogue recording, and digitising it at whatever sample rate. The analogue recording itself may have little content above 20kHz. Again, if the provenance of the recording is made clear, that's fine. Whereas if you pop a brick wall LPF at 20kHz, this again is a scam IMO.

I've put aside whether or not it sounds better - the point is, if higher sample rates could sound better, to you or a bat or whatever, then when people sell "hi res audio", it should have the potential to offer this advantage.


Unless of course the whole thing is a placebo induced con - in which case you can take the analogue output of a $99 CD player, re-digitise it at 192kHz/24-bits, sell it as HD, and it'll sound better to the purchaser.

Cheers,
David.

Saving music from destruction: upsampling

Reply #12
You can read all about that, as illustrated by real world examples, here:

Sample rate conversion technical tests

Not conclusive and lets open what filter you prefer, Aliasing over Pre-ringing or other evils that are non-issues, like always. And imagine the 44.1KHz version was downsampled already. What kind of filter did they use? If i apply the same filter again, may it hurt or clip?

I better trust the DAC designer that he chooses a LP in the DAC itself that doesn´t cause to much side-effects.

The first is taking a 44.1kHz or 48kHz digital recording, and upsampling it, and then charging people for the privilege. Well, if you tell them what you've done, why not? Let them decide whether they want to pay you for that service, or buy the CD and upsample it themselves, or just play it as normal (which will oversample it anyway!). Whereas if you claim your result is somehow natively 96kHz or 192kHz or whatever, it's a scam. Plain and simple.


Like they tried to sell High bitrate/kHz stuff here until a customer noticed the content ended sharp at ~22kHz. After complaining the studio admitted to use that for sonically reasons for preventing to "distort" the electronics behind with to high content. LOL!
http://www.computeraudiophile.com/Kent-Poo...ogue-III-Review



Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Saving music from destruction: upsampling

Reply #13
It's hardly any surprise that there's almost no-one left who gives a damn about this (high resolution audio, I mean).

People say "oh, the kids have got used to mp3 - they don't know what high quality sound is, or care".

B*ll*cks!!!!

People have been watching SD video for years, and when they see HD, they mostly go "wow" and save up to buy a nice big TV. They never go "oh, I've got so used to low quality low bitrate SDTV that I really don't need HD".

The reason they don't go "wow" with HD audio has nothing to do with mp3 or not caring - it's because there's nothing to say "wow" to. High resolution audio fixes a problem that either doesn't exist, or (at best) is very small.


When you think of all the wonderful things the audio industry could be doing (better transducers, "you are there" binaural, holographic multi-channel etc), and then see people selling upsampled 2-channel audio as "the next big thing", well... it makes me so angry! (you'd never guess, would you?  )

Cheers,
David.

Saving music from destruction: upsampling

Reply #14
B*ll*cks!!!!

I second that
Really depressing when for example 2 titan´s of self-proclaimed diamond ears start to argue what sounds better. More kHz or more bits?
It may be fun to read all the superlative word-creations when reading such threads but when reading this at many places more and more often makes me really fear of what comes next!!
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Saving music from destruction: upsampling

Reply #15
The reason they don't go "wow" with HD audio has nothing to do with mp3 or not caring - it's because there's nothing to say "wow" to. High resolution audio fixes a problem that either doesn't exist, or (at best) is very small.

Hmm, makes me think. CD audio is 30 years old. Which other media technology for consumers from the early 1980s is still being used today? VHS? Compare how far we've come in 2D video (VHS -> Full-HD BluRay) and stereo audio (CD -> lossly compressed CD  )

Isn't CD stereo audio actually the equivalent/counterpart of 2D Full-HD video? The CD specification was ahead of its time (or video far behind). I think the recording industry tried for years to come up with much-better-than-CD stereo sound in order to keep the market alive. Even though they failed again and again, some still try, like Chesky records.

I fully agree with David, what's still missing is "you are there" binaural sound over head- or earphones. Convincing surround sound for-here or to-go, so to speak  The keys to that are not 24 bit or 96 kHz. It's listener motion tracking and coding more than two channels, or even a channel-number agnostic approach to surround sound coding.

I hope the music industry will have realized this before I'm old and deaf...

$0.02 from Chris
If I don't reply to your reply, it means I agree with you.

Saving music from destruction: upsampling

Reply #16
...
I fully agree with David, what's still missing is "you are there" binaural sound over head- or earphones. Convincing surround sound for-here or to-go, so to speak  The keys to that are not 24 bit or 96 kHz. It's listener motion tracking and coding more than two channels, or even a channel-number agnostic approach to surround sound coding.

I hope the music industry will have realized this before I'm old and deaf...

$0.02 from Chris

Why do you think that some people are willing to pay for "3D" tvs and Bluray players, and wear silly glasses? Why wont they do the same for audio?


-k

Saving music from destruction: upsampling

Reply #17
I'm not against surround sound or any other approach to realism, but I have no enthusiastic longing either. I have no place for a proper multi-speaker setup and I suggest that only a very small, mostly wealthier part of the population does either. I don't see any way for that to change for the better.

I do have a two channel setup that produces a half way decent stereo field. There is frequently a spread of performers across the room, and occasionally some obvious depth. The image of a few (very few), recordings seem to be wider, and occasionally deeper, than the room. Results are very dependent on the material being played.

I have some evidence that more is possible with the hardware I have, but there is no practical way to realize it. I cannot dedicate a suitable room, or any room, to the configuration and specialization necessary to achieve it.

My limited stereo setup is definitely more enjoyable (to me, anyway) than the haphazard arraignment used by about 98% of the population. I've occasionally pointed out the deficiency but gave up doing so years ago. Response were either disinterested or "I know, but ...", indicating that the equipment being convenient out of the way (with the speakers close together, or even pointing away from the listener) had an overwhelming priority. This also applies to people who have spent significant money on "theater" or "surround sound" systems.

I think the vast majority of people don't even know what stereo really is. I've had a couple or so, well educated to the convention standards and far from indigent, freeze in open mouthed surprise upon hearing the instruments of a recording coming from different, distinct positions in my modest front room. It was their first such experience.

But I've also heard some amazing recordings on headphones. Perhaps this is my general ignorance coming through, but a few made-for-headphone-listening binaural recordings I've purchased have not been particularly impressive. On the other hand, I've heard a few others that have.

One in particular I found on line some years ago was made on a minidisk player/recorder by someone walking down an alley. I don't recall if there was more information on technique provided, but there was nothing used that was not readily available years ago in the hardware. This was true surround, with sounds coming from ahead, behind, to the sides, and from above, all very clearly placed.

Saving music from destruction: upsampling

Reply #18
Like they tried to sell High bitrate/kHz stuff here until a customer noticed the content ended sharp at ~22kHz. After complaining the studio admitted to use that for sonically reasons for preventing to "distort" the electronics behind with to high content. LOL!
http://www.computeraudiophile.com/Kent-Poo...ogue-III-Review



Ugh. I made it as far as the references to James Boyk's and Milind Kuncher's work as evidence for the necessity of >22kHz content in recordings, then gave up.

Saving music from destruction: upsampling

Reply #19
I'm not against surround sound or any other approach to realism, but I have no enthusiastic longing either. I have no place for a proper multi-speaker setup and I suggest that only a very small, mostly wealthier part of the population does either. I don't see any way for that to change for the better.

Technology progression? Flexible capture/distribution formats that can be presented at any playback venue using speakers etc as best possible. Cheap distributed loudspeaker panels that can be hidden in the decor of the room, or directive "soundbars" that use room-reflections to create virtual sources.
Quote
One in particular I found on line some years ago was made on a minidisk player/recorder by someone walking down an alley. I don't recall if there was more information on technique provided, but there was nothing used that was not readily available years ago in the hardware. This was true surround, with sounds coming from ahead, behind, to the sides, and from above, all very clearly placed.

It is sad that we see so little commercial binaural recordings of good content. I know that I am willing to pay, and there have to be some iPod/Phone/Pad users out there who are willing?

-k

Saving music from destruction: upsampling

Reply #20
I'm not against surround sound or any other approach to realism, but I have no enthusiastic longing either. I have no place for a proper multi-speaker setup and I suggest that only a very small, mostly wealthier part of the population does either. I don't see any way for that to change for the better.
...but that's the problem we should be looking to solve. There are some very convincing ways of making sounds come from locations where there are no speakers. The commercialised versions of this technology are generally poor, but there's great stuff in various universities and research labs (well, there was a decade or more back).

Quote
I do have a two channel setup that produces a half way decent stereo field. There is frequently a spread of performers across the room, and occasionally some obvious depth. The image of a few (very few), recordings seem to be wider, and occasionally deeper, than the room. Results are very dependent on the material being played.
But this audiophile beloved trick with stereo is just that: a trick. That's why it falls apart so easily, and is so recording, equipment, and room specific. It is possible to create a wonderful, almost 3D effect with simple mic'd 2 channel stereo - it's very impressive - except when compared with doing it "properly" (i.e. more channels, proper breadth/depth clues, etc) which just blow it out of the water.

Quote
My limited stereo setup is definitely more enjoyable (to me, anyway) than the haphazard arraignment used by about 98% of the population. I've occasionally pointed out the deficiency but gave up doing so years ago. Response were either disinterested or "I know, but ...", indicating that the equipment being convenient out of the way (with the speakers close together, or even pointing away from the listener) had an overwhelming priority. This also applies to people who have spent significant money on "theater" or "surround sound" systems.

I think the vast majority of people don't even know what stereo really is. I've had a couple or so, well educated to the convention standards and far from indigent, freeze in open mouthed surprise upon hearing the instruments of a recording coming from different, distinct positions in my modest front room. It was their first such experience.
Yes, most people put their speakers where they will look nice. That said, in 2011, we really ought to be able to make sonically near-transparent speakers that blend into the decor, for a reasonable price.

You've made a point though: good audio is still a jaw-dropping experience. How has the industry failed to commercialise it as effectively as good video?

Cheers,
David.

P.S.

Quote
One in particular I found on line some years ago was made on a minidisk player/recorder by someone walking down an alley. I don't recall if there was more information on technique provided, but there was nothing used that was not readily available years ago in the hardware. This was true surround, with sounds coming from ahead, behind, to the sides, and from above, all very clearly placed.
It's quite easy to do with half-decent microphone mounted at the sides of your head. There are two problems:
1) try it with music, and often it's not quite as satisfying or convincing.
2) such recordings sound pretty awful over speakers.
...given how popular headphone listening is, it still surprises me that there's not a "binaural" (or whatever you choose to name it) version in iTunes for various releases.

Saving music from destruction: upsampling

Reply #21
Why do you think that some people are willing to pay for "3D" tvs and Bluray players, and wear silly glasses? Why wont they do the same for audio?

Good question. I think because

1) you don't have to wear silly additional gadgets to experience 3D audio, just your headphones, maybe with motion sensors built into them. You can do the rest in software.* But it hasn't been done, at least not in an affordable way.
2) the binaural sound is most of the time not convincing for everybody. See also David's post above. Each person has a different head and ear shape, so there's no "standard" binauralization. In principle you'd need a custom binaural profile for each listener (or at least a preset which comes close). See the fine differences between individuals' HRTFs.

Chris

* The same also goes for 3D TV, by the way. Wearing silly glasses is indeed silly. You could do without, as some prototypes have demonstrated.
If I don't reply to your reply, it means I agree with you.

Saving music from destruction: upsampling

Reply #22
2) the binaural sound is most of the time not convincing for everybody. See also David's post above. Each person has a different head and ear shape, so there's no "standard" binauralization. In principle you'd need a custom binaural profile for each listener (or at least a preset which comes close). See the fine differences between individuals' HRTFs.

Chris

But my stereo is not customized for my ears, either.

The question is if generic binaural can bring us a significant step towards the ideal compared to regular stereo. Having adaptation to individual head shape and dynamic head tracking would be a welcome refinement.

I guess that our eyes have slightly different inter distance as well, and current "3D" displays show a static viewpoint (no headtracking), still people find it useful. The analogy is not a very good one, but anyways...

-k

Saving music from destruction: upsampling

Reply #23
But my stereo is not customized for my ears, either.
Doesn't need to be though - it doesn't bypass your head and outer ear.

Quote
The question is if generic binaural can bring us a significant step towards the ideal compared to regular stereo.
It needs decent headphones to work well and sound natural, with a known frequency response.

Mismatched HRTFs make it sound very unnatural.

Without head tracking, and without reverb, the image collapses into the head.

Most significantly, lots of music just doesn't sound as fun presented in half-decent binaural compared with a normal mix. The normal mix sounds more immediate and exciting, even over headphones.

Quote
I guess that our eyes have slightly different inter distance as well, and current "3D" displays show a static viewpoint (no headtracking), still people find it useful. The analogy is not a very good one, but anyways...
I can't disagree - it's an interesting and reasonable analogy. Then again, we've had binaural and 3D for years (both with varying degrees of technical success!). Now it seems 3D's time has come.

Cheers,
David.

Saving music from destruction: upsampling

Reply #24
You've made a point though: good audio is still a jaw-dropping experience. How has the industry failed to commercialise it as effectively as good video?


My guess: golden cables make more money so the incentive isn't focused on real quality audio, and given that we're eye-creatures, pretty pictures are a much easier sale than pretty audio.

The same bollocks trend is now emerging for HD televisions. Going from DVD to full-HD is a big step, but beyond that there's little to be gained. I know of a TV that has special yellow subpixels because they supposedly make the image seem more alive over "common" HD. The benefits of a dedicated yellow subpixel are dubious. I expect it's not the last trick to be invented, and before long, we'll have true pure-blood videophiles. When that happens, it's time for hydrogenvideo.org to be registered.