48000Hz vs 192000Hz aka What The Heck Is Going On?

[dmills]

Next we will be seeing the 'R' word being applied to audio.... Video is sampled at way below the Nyquest limit, hence wagon wheels in old movies spinning backwards, plaid shirts looking awful on analogue telly and such, audio is not sampled this way and higher sample rates gain you nothing within the original passband.

As long as you limit the bandwidth before the sampler to strictly less then 1/2 the sample rate you capture everything that is there, no ifs, no buts.

Now there are issues with that statement, particularly with the required filtering and its resulting phase response, but that is more a small matter of engineering then a fundamental limitation.

And yes, having more bandwidth in a system then you need to obtain a flat group delay over the band of interest is a bad thing, it makes your electronics more susceptible to noise pickup, and potentially more prone to take off somewhere up above the audio band which will do your tweeters a world of good, it also wastes headroom amplifying stuff that is out of band (Usually more of a problem with excessive LF extension).
It also potentially passes energy above the useful band to transducers which are usually not particularly linear up there, this can give rise to IMD products within the audio band even if the original signals are above the audible range, DC to daylight is not a figure of merit in an audio reproduction chain.
This can actually be a problem with microphones, some that make a feature out of a HF response to many tens of kHz can actually produce difference frequency tones in the output when driven by a reasonable level of two tone ultrasonic power, what this does to a close miced recording of a violin I will leave as an exercise for the reader......

I couldn't find any proper documentation on those BAT amps (Think schematics), but I would regard a 250kHz bandwidth as excessive and as being an excuse to end up amplifying every smpsu in the area. I am also somewhat suspicious of 'zero NFB' as a claim, they might have no overall negative feedback, but there will be at least the intrinsic mosfet source resistance (And probably an external degeneration resistor) providing local feedback, then you have the Miller capacitance.....

Regards, Dan.

[Elephone]

I just wish I could see Hugh's face when he returns from his holidays, catches up with all the activity on the forums, and reads this thread.

Andy

Ha-ha! I was scanning for a response from Hugh when I read your comment.

I can see the purpose of much higher sample rates for capturing ultra high frequencies (e.g. tiny insects) with a 50kHz mic, but only for analysis or maybe to slow them down to human hearing. Otherwise, if there is anything audible with 192KHz, I can only think it's to do with the way digital processors are dealing with it...?

[dmills]

Yup, aliasing in the digital domain is a common issue in poorly written soft synths and dynamics plugins, also half band decimators in delta sigma converters can be an issue if not handled correctly.

When you do it wrong, it does not work properly, no surprise there!

Regards, Dan.

[TFHT]

It's my opinion that you over-thinking this and taking it all a bit too seriously.

Your Realtek soundcard seems, from what I know, a consumer thing for average audio listening (i.e. MP3 listening). It is by no means a 'professional' audio I/O. To think superlative results are possible is, I think, a little naive ... but I appreciate that you're new to all this and yes, it CAN be damned confusing sometimes.

The Acer you have is a fine little thing for general computing, gaming, music listening, video/Netflix/YouTube viewing but I'd be loathe to use it for any serious recording, especially without some quality external audio I/O interface.

I hate to bring the news to you but having a laptop does not necessarily give you a recording studio. And sorry, 192kHz recording is total arsewash - 44.1/16-bit is perfectly adequate. At a push, 44.1/24-bit is (arguably) better but this will have to be dithered down to 16-bit for CD release and if your target market is an MP3 release, 192kHz really is overkill ... or, to use a technical term, horsesh!t with more than a hint of "The Emperor's New Clothes" about it.

Sorry to be blunt

Well, I may not be able to afford a better computer, but I can afford an external soundcard / audio interface, given that even the cheap ones now seems loads better than my current soundcard.

But, I never expected superlative results from my gear, nor do I think I have a recording studio. I just want to see where my gear's limits peak, and to map out my capabilities so I know what to immediately upgrade and what not to. This thread is doing exactly that.

If I, or anyone waited until they had superlative gear to start making music, I think there would be far less music in general, and far less quality music. You gotta start somewhere. If Eric Clapton waited until he had a $10,000 guitar to learn how to play, instead of playing on a crap cigar-box guitar, we wouldn't know who Mr.Clapton was because he would have probably starting playing at 30 instead of 10. I'm no Clapton, and I don't expect to appease the Claptons of the world just yet. I'd rather push my crap gear to its limits and make good music than have the best gear in the world and not know how to use it.

Gotta start somewhere.

[ernzo]

In short, 44.1KHz is perfectly capable of completely reproducing all frequencies up to Nyquist (half the sample rate) - more samples does not equal "more detail".

Great explanation thanks!

I would like to see the same done with real complex audio tho..

Is this true regardles of source complexity?

I mean what if you have a super beautifully complex sounding guitar, violin and piano, playing complex harmonies alla Piazzolla?

Will all those harmonics, supra/sub harmonics and the harmonic relationships of and between every instrument, go thru completely clear regardless of samplerate? (up to nyquist limit, and considering ideal recording conditions/method..)

[dmills]

Yes, **ANY** signal that is band limited to a bandwidth strictly less then half the sample rate can be captured entirely by sampling at that rate.

Consider that any signal can be broken down into a (generally infinite) set of summed sine waves (Not going to demonstrate this without math rendering, but it is VERY well established fact), and that as any sine wave can be fully characterized by at least 3 points over the course of a cycle, then it follows that any sum of sine waves can be fully characterized by measuring the instantaneous value of the sum at a rate greater then twice the bandwidth (Hence ensuring the three points thing for the highest frequency). Addition is after all a linear operation.

44.1k is entirely sufficient for **ANY** signal band limited to DC-(just under) 22.05khz.
Practically of course the filters at both ends are not perfect and leaving a little space for a transition band does not hurt, but in principle....

Music is magic, audio is engineering, and should be treated as such.

Regards, Dan.

[Exalted Wombat]

In short, 44.1KHz is perfectly capable of completely reproducing all frequencies up to Nyquist (half the sample rate) - more samples does not equal "more detail".

Great explanation thanks!

I would like to see the same done with real complex audio tho..

Is this true regardles of source complexity?

I mean what if you have a super beautifully complex sounding guitar, violin and piano, playing complex harmonies alla Piazzolla?

Will all those harmonics, supra/sub harmonics and the harmonic relationships of and between every instrument, go thru completely clear regardless of samplerate? (up to nyquist limit, and considering ideal recording conditions/method..)

Yes, all music - by definition - adds up to a mixture of audible frequencies (do we need to argue over how you "hear" particularly low frequencies?). Sample at twice the highest your ear can make out (44.1KHz is plenty) and you're good.

This is not particularly intuitive. It's easy to convince yourself that complex music must contain "more frequencies than you can hear". But think about it!

The system relies on effecient filtering at half the sample rate, free of side-effects. We're pretty good at doing that now.

[Goddard]

Sampling rate discussions like this thread are so fun. Really takes me back... to about 15 years ago.

[Folderol]

Well that was a fun read, and I think just underlines that when someone says they can hear/see/smell/feel something, you are wasting your breath trying to explain why they can't.

[Exalted Wombat]

Very true however on those high end systems I'm referring to your not hearing higher frequencies as much as your now hearing things more linearly. What's going on in those stratosphere frequencies is pretty important. To be a little poetic... It's the sonic glue that holds the whole thing (on a really interesting level) together. After you've heard 192 that was recorded at 192 on a high end audiophile system and then you hear the same track at 44.1 it becomes obvious what your missing. The 44.1 track sounds ailiased (is that a word? My spell check doesn't think so) well regardless I'm not arguing with you because your right except for a few random lucky people and dogs most of us can't hear higher than around 19-20k but again it's not that your hearing higher frequencies but that due to the ultra high sample rate those 15k-20k we can hear are WAY better represented with the overtone structure that is there, albeit that you can't hear even at a live performance, but that you can feel.

Eh? Want to try that again, in English, with your brain turned on?

[dmills]

No argument about that!

However, when someone makes a claim that they can see/hear/feel something BECAUSE then taking the time to call it is a good thing as it keeps misinformation from spreading.

The claim to see/hear/feel something is not anything that is realistically tested online without a lot of pain for all concerned. The statement that such is because of is best called out as nonsense in those cases when clearly cannot be the cause.

Regards, Dan.

[Exalted Wombat]

Well that was a fun read, and I think just underlines that when someone says they can hear/see/smell/feel something, you are wasting your breath trying to explain why they can't.

Oh come on! We know the OP was a troll. But then there were some vaguely sensible questions...

[TFHT]

Well that was a fun read, and I think just underlines that when someone says they can hear/see/smell/feel something, you are wasting your breath trying to explain why they can't.

Now, I know you're not talking to me, because the difference between my 48kHz and 192kHz renders is night and day. If there was a reason why I couldn't actually hear the difference, it would be because I have no ears. If you want, I could send you the two different renders, and you can hear for yourself. I am hearing a vast difference, but I do NOT claim that I can hear beyond human limits. There is probably something wrong with my setup, and I aim to find that problem. Once corrected, I fully expect to hear virtually no difference.

Screw it. Here: https://soundcloud.com/tinfoilhattrick

Even on Soundcloud you can hear the differences.

[Kwackman]

A group of scientists is perfoming a series of experiments on a spider.
They cut one leg off the spider and tell it to crawl. The spider crawls.
Next they cut a second leg off the spider. "Crawl!" they command. The spider crawls.
They cut off a third leg. "Crawl!" The spider crawls.
They continue cutting the legs off the spider one at a time and commanding it to crawl.
Each time, with one leg less, the poor spider crawls.
Finally, they cut the eighth leg off the spider and tell it to crawl. The spider does not respond.
The one scientist says to the other, "Make a note, when we cut off the eighth leg, the spider became deaf.".

[Tim Gillett]

Well that was a fun read, and I think just underlines that when someone says they can hear/see/smell/feel something, you are wasting your breath trying to explain why they can't.

Now, I know you're not talking to me, because the difference between my 48kHz and 192kHz renders is night and day. If there was a reason why I couldn't actually hear the difference, it would be because I have no ears. If you want, I could send you the two different renders, and you can hear for yourself. I am hearing a vast difference, but I do NOT claim that I can hear beyond human limits. There is probably something wrong with my setup, and I aim to find that problem. Once corrected, I fully expect to hear virtually no difference.

Screw it. Here: https://soundcloud.com/tinfoilhattrick

Even on Soundcloud you can hear the differences.

Thanks for the audio samples. To my ears there's not much difference but then for frequencies around 15 to 20khz my ears arent exactly stellar any more.

Very hard to tell much from your samples in absolute terms because this "electronic" type of music cant be directly compared with something like a human voice or instrument like a piano whose sounds we know well. To me BOTH samples sounds heavily distorted but then again maybe that's how you want the music to sound.

Is there an audible difference in the sound before and after it is recorded?

Tim

[largon203]

Just for the record, BAT amps were simply a random drop to explain how the audiophile industry is wide and very expansive. I could have name dropped "Blue Circle" or "Dan D Agoslino" or Even older McIntosh amps the list goes on and on. I'm not saying you don't have the right to your thoughts on the matter and I respect it, but do NOT tell me and others that I or we don't hear what we hear. Maybe it's time you had another listen seeing you admittedly have been out of the audiophile world for some time. Seeing that the Audiophile world has a tendency to Change as quick as the PC industry; one could argue you suffer from an overdose of contempt prior to investigation. You seem like a smart guy. Give it another go aye?

-Dave

P.S. = Lavry Whites' paper wasn't all bad (again in my opinion) he raised some good points, and I learned a fair bit, but before I learned anything about sampling rates and the such I learn the physics of sound and I'm afraid based on the laws of Audible physics he is fundamentally flawed. You can try to record something at 192k and if no signal is present above 20khz Aka 44.1 sample rates (primarily overtones and other ultra high noises relating to both the eligible instruments, environment and the like) the "Quality" 192k Analog to Digital converter records important supersonic info that supports and provides special inner detail that is kind of unexplainable. You have to hear it to know what I'm saying. Unless your over amplifying the signal to produce undue noise which is of course ridiculous. He mentions how you wouldn't want to record @ 100mhz sampling rate because of high distortions... Duh... I think the following statement gets me the most... "Again, ultrasonic energy will at best cause no harm to the sound you need, but it certainly cannot help;" he then goes on to say how it's not worth the hard disk space. My problem with this statement is if you have a pair of the best Neumann mics you can get and you record a simple stereo recording of a symphony and / or piano and you want the most natural real to life sound you can get: 192 will Blow you away. I have had the pleasure of assisting in such a production and have had the opportunity to hear the honest difference and Mr White, while incredibly intelligent and articulate is just wrong when he says " ultrasonic energy will at best cause no harm, but it certainly cannot help" With an Audiophile system that is setup properly playing back a signal recorded at 24 bit, 192k sampling rate not only can I/we hear the difference between it and lower sample rates, It seems to capture the essence of the instruments and their environments in a non-subtle majorly positive way, well beyond the lower sample formats. Call me a troll, call me what you will but don't tell me what I and so many around me hear.

Until next time...
-Dave

[TFHT]

Well that was a fun read, and I think just underlines that when someone says they can hear/see/smell/feel something, you are wasting your breath trying to explain why they can't.

Now, I know you're not talking to me, because the difference between my 48kHz and 192kHz renders is night and day. If there was a reason why I couldn't actually hear the difference, it would be because I have no ears. If you want, I could send you the two different renders, and you can hear for yourself. I am hearing a vast difference, but I do NOT claim that I can hear beyond human limits. There is probably something wrong with my setup, and I aim to find that problem. Once corrected, I fully expect to hear virtually no difference.

Screw it. Here: https://soundcloud.com/tinfoilhattrick

Even on Soundcloud you can hear the differences.

Thanks for the audio samples. To my ears there's not much difference but then for frequencies around 15 to 20khz my ears arent exactly stellar any more.

Very hard to tell much from your samples in absolute terms because this "electronic" type of music cant be directly compared with something like a human voice or instrument like a piano whose sounds we know well. To me BOTH samples sounds heavily distorted but then again maybe that's how you want the music to sound.

Is there an audible difference in the sound before and after it is recorded?

Tim

Well, the sounds are distorted on purpose. As I said in the beginning of the thread, the song is two parts. #1 Being the first chord played by the Zebra2 soft synth (which sounds better at 192kHz) and the distorted drums and distorted guitar being #2, which sound worse at 192kHz. #1 has no effects, and #2 has amp distortion and bit crushing effects.

My computer can't perform with FL Studio at above 48kHz without the songs sounding like they are having seizures inside an AM radio station, so I use 48kHz inside FL Studio, and the 48kHz renders sound no different.

For all I know, the synth (Zebra2) is supposed to sound at least a little better, and the distortion effects are supposed to sound even more so distorted at higher sample rates. This could be why, when I rendered a different song without those effects, the synths still sounded better, but the non-synth VSTs and samples sounded the SAME.

[ernzo]

Is there any serious 192khz vs Xkhz Test that you know?

-What about 64 bit, or n-bit audio?

If 32bit is such an apparent improvement over 16bit, maybe it's bit depth what gives digital audio part of its precision.. wouldnt 64bit be a theoretical improvement? ever been tried?

-Btw, I remember reading some years ago about some odd A/D converter that guitarist Eric Johnson was using.. i cant recall the details, or find it, but it was something about it being based on pure bit streams, rather than having a sample rate...

[Tim Gillett]

P.S. = Lavry Whites' paper wasn't all bad (again in my opinion) he raised some good points, and I learned a fair bit, but before I learned anything about sampling rates and the such I learn the physics of sound and I'm afraid based on the laws of Audible physics he is fundamentally flawed. You can try to record something at 192k and if no signal is present above 20khz Aka 44.1 sample rates (primarily overtones and other ultra high noises relating to both the eligible instruments, environment and the like) the "Quality" 192k Analog to Digital converter records important supersonic info that supports and provides special inner detail that is kind of unexplainable. You have to hear it to know what I'm saying. Unless your over amplifying the signal to produce undue noise which is of course ridiculous. He mentions how you wouldn't want to record @ 100mhz sampling rate because of high distortions... Duh... I think the following statement gets me the most... "Again, ultrasonic energy will at best cause no harm to the sound you need, but it certainly cannot help;" he then goes on to say how it's not worth the hard disk space. My problem with this statement is if you have a pair of the best Neumann mics you can get and you record a simple stereo recording of a symphony and / or piano and you want the most natural real to life sound you can get: 192 will Blow you away. I have had the pleasure of assisting in such a production and have had the opportunity to hear the honest difference and Mr White, while incredibly intelligent and articulate is just wrong when he says " ultrasonic energy will at best cause no harm, but it certainly cannot help" With an Audiophile system that is setup properly playing back a signal recorded at 24 bit, 192k sampling rate not only can I/we hear the difference between it and lower sample rates, It seems to capture the essence of the instruments and their environments in a non-subtle majorly positive way, well beyond the lower sample formats. Call me a troll, call me what you will but don't tell me what I and so many around me hear.

Until next time...
-Dave

For the record, the man's name is not Lavry White but Dan Lavry. Dan Lavry wrote a White Paper which apparently you read.

I believe Dan Lavry has also designed and sold a range of well regarded converters. Perhaps as a matter of principle, I believe he refused to design a converter capable of 192khz sample rate, even though he could have done so.

I own a soundcard capable of 192khz sample rate and capable of an essentially flat response to over 80khz. I have never used it for real time audio although I could. I purchased it to digitally capture signals from a professional high speed audio tape duplicator which played back tapes at 8 times normal speed, resulting in this case in frequencies up to 80khz. The soundcard does a good job in that application.

As I suspected, this is not really about Shannon Nyquist or Dan Lavry's discussion of it, but the place of ultrasonics in reproducing audio. You seem to believe that preserving ultrasonics plays a big part in the the naturalness of the sounds we hear.

But if the difference between 44.1 and 192 "blows you away" we have a problem. Many instruments dont produce ultrasonics. Many if not all human voices also. Do you exempt those cases from your generalisation?

More critically how can we be "blown away" by sounds that under any controlled tests we cannot even hear? Cannot recognize in any way through any of our five senses? Surely the "essence" of any sounds we humans hear is in what our ears CAN hear, rather than what they CANT hear. If we filtered out everything below 20khz, leaving only ultrasonics, what would we hear? The essence? Far from hearing some distilled essence I suggest we'd hear nothing.

There is a story that goes around about Geoff Emerick, the Beatles engineer. A mixing desk which he was accustomed to working with had had some repairs done to it and after the repair he wasnt happy with the sound. Others couldnt hear what he was talking about. It turned out that part of the repair had been made in correctly and there was now an ultrasonic tone present. That tone was way above what was humanly audible, but it apparently was interacting with other audible signals, producing a distortion product which WAS audible to Emerick.

It all depends on the interpretation. Some people wrongly concluded from this that Emerick had the ability to hear 50 to 60khz signals. Others with more of an audio engineering background rightly deduced the far more likely explanation was that he was hearing a distortion product, related to a beat frequency.

This is also an example of what Lavry was saying, where if an audio signal chain is designed not to reproduce ultrasonics in the first place it is less likely that these sorts of distortions will occur.

Tim

[molecular]

Is there any serious 192khz vs Xkhz Test that you know?

-What about 64 bit, or n-bit audio?

If 32bit is such an apparent improvement over 16bit, maybe it's bit depth what gives digital audio part of its precision.. wouldnt 64bit be a theoretical improvement? ever been tried?

-Btw, I remember reading some years ago about some odd A/D converter that guitarist Eric Johnson was using.. i cant recall the details, or find it, but it was something about it being based on pure bit streams, rather than having a sample rate...

ernzo, I spent a long time grappling with this stuff and have started one or two of these bonkers threads myself in the last couple of years. Having set aside issues of cheap convertors, badly manufactured CDs and all kinds of other stuff that comes up my simplest understanding of it is this:

1. A higher sample rate simply allows you to capture sounds of a higher frequency.
2. A higher bit depth simply lowers the noise floor / gives you more headroom.

The only occasions I can think of when these two pieces of information aren't sufficient to allow me to do what I want to do is when I'm reading an SoS thread about sample rates and bit depth.

there is a totally separate argument being laid out by the audiophile camp on this thread which I think is derailing the facts, which is one about whether we can hear/feel/sense audio which is either/both above 20Khz in frequency or below the normal threshold of hearing (in terms of volume). Your man a couple of posts back obviously thinks that this is possible, and that it explains why 192khz sample rates are better (in his view). That's got nothing to do with the question of whether 192khz or 32bit does a much better job of reproducing the "detail" or "resolution" of audible sound in the 0-20khz range, though. As has been mentioned, the 'search forum' button will provide you with up to two years worth of Hugh setting people straight!

[molecular]

As I suspected, this is not really about Shannon Nyquist or Dan Lavry's discussion of it, but the place of ultrasonics in reproducing audio. You seem to believe that preserving ultrasonics plays a big part in the the naturalness of the sounds we hear.

beaten to it.

what he said ^^^

[muzines]

If 32bit is such an apparent improvement over 16bit, maybe it's bit depth what gives digital audio part of its precision.. wouldnt 64bit be a theoretical improvement? ever been tried?

Terms like "precision" and "resolution" are kind of nasty if the digital audio world, because if you aren't clear on how things work, then can easily be interpreted wrongly to mean something different.

*All* a higher bit depth does is provide a higher potential signal-to-noise ratio, it lowers the noise floor. As 24-bit audio already handles a noise far generally below that which is achievable by typical analog front ends, it's pretty sufficient, 32-bit is not that useful, and 64-bit audio recording is pointless.

There is a reason to have higher bit depths in DSP processing, where DSP processes can produce numbers which require that precision for calculating (for various technical reasons dependent on the algorithms being used) but this more about the maths than about how accurate the sampled representation of audio is.

44/24 is really capable of producing perfectly great quality audio.

[muzines]

I'm not saying you don't have the right to your thoughts on the matter and I respect it, but do NOT tell me and others that I or we don't hear what we hear.

Obviously I cannot comment on this/your particular experience but a big part of the problems in the audiophile world is that people swear they can hear something when there are other things going on that are affecting perception.

These tests need to be very carefully done, and under completely double blind conditions and a number of people need to have clear and reproducible independent agreeing results, otherwise the results are meaningless. Those listening tests should also be backed up by solid experimental data (for example, there is no point in having ten people that say they can hear above 20K if the source equipment is shown to be not capable of generating it...)

Unfortunately, there is too much of the former and too little of the latter... Listening tests only show the results of individuals interpretations of what their senses are telling them, biased by their own opinions and experience. Not to say listening tests aren't useful, of course, as ultimately all music is designed for that purpose.

[johnny h]

Now, I know you're not talking to me, because the difference between my 48kHz and 192kHz renders is night and day. If there was a reason why I couldn't actually hear the difference, it would be because I have no ears. If you want, I could send you the two different renders, and you can hear for yourself. I am hearing a vast difference, but I do NOT claim that I can hear beyond human limits. There is probably something wrong with my setup, and I aim to find that problem. Once corrected, I fully expect to hear virtually no difference.

Screw it. Here: https://soundcloud.com/tinfoilhattrick

Even on Soundcloud you can hear the differences.

Well yes there is a huge difference I can hear even on my laptop .. its just a lot more harsh and distorted. Its probably due to one of the distortion plugins not working at 192kHz or something. You will not be able to hear a difference between 48kHz and 192kHz on a compressed soundcloud link on laptop speakers. Some other process is involved.

I think its really funny how people get so involved in these kind of debates, especially when you often see guys running fruity loops and some free plugins playing to 5000 people every weekend while these nerds post 384kHz-recorded buchla noodles to 80 disinterested soundcloud followers.

[narcoman]

Hey Tim

I'm afraid that what I heard and experienced wasn't really up for debate. When you get a chance to listen to a proper demo on an Audiophile system that is tuned properly costing upwards around 1/4+ million USD you let me know what you experienced.
Kudos mate!

-Dave

I have such a system.

I also know Dan Lavry. Knows his stuff. A brave fool would dare to argue with him on digital audio.

Just saying. So i call troll