John
Goddard;
I agree that he has a style that is not paper quality. I also have a problem with people talking about science in the way he does. I know he is addressing the public. However, I have been saying much the same thing for many years now. I did as you suggest and read the bit depth article and again I am in total agreement.
What he says is the way I understand both sample rate and bit depth. I shall put it to you to clearly state where and what he is saying is wrong. Please give citations and where your objections are coming from.
John, first off, as you might be aware (and Noel certainly should be) the question of recording at sample rates >44.1/48kHz (and at bit depths >16) is hardly new and has been discussed (and debated) for many years already, at least ever since it became feasible and the DVD Video spec supported higher rate/depth LPCM audio. See e.g. this article on (the then newfangled) 24/96 recording:
http://web.archive.org/we...8E706886256688000FBE08 and (of course) the ensuing online debate:
https://groups.google.com/d/msg/cakewalk.audio/sKdF8ZHvlJc/jLbRbsNd75QJ As hardware and software and computers advanced to the point where even higher (i.e. "quad) sampling rate recording became feasible (and SACD and DVD-A fizzled onto the scene), the discussions and debates were recast around the higher rates/bit-depths.
For people concerned with latency, one attraction of recording/playing at higher sampling rates was that some audio interfaces exhibited significantly reduced latency (i.e., shorter filter "group delay") when operated at "double" rate. But to be sure, there were certainly valid "downside" concerns wrt recording at higher sampling rates, such as higher system and disk i/o loading (now no longer of as much concern given today's vastly more powerful systems) or more quickly bumping up against Windows' 2GB WAV file size limit (now no longer a concern thanks to WAVE-64).
On top of that, everyone still heard differently, and still had a different opinion about what they (and everyone else) could/couldn't hear. Plenty of marketing hype always flying around too, just confusing and obscuring things even more. Oh the questions and indecision...
Along the way Dan Lavry put out a paper in which he attempted to dispel certain "myths" regarding higher-than-Nyqist frequency sampling rates, subsequently cited as definitive gospel by many (Lavry has since come out with a follow-up addressing more "myths"). In his papers, Lavry, whose company's converters go no higher than 96k, asserted that higher sampling rates were deleterious to audio quality, although his arguments on that point lacked any actual proof (imo) and the limitations he pointed to were really not (at least, no longer) valid for many commercially available audio converter chips (except perhaps less expensive and older ones, or at the possible loss of a slower filtering option in some converter ICs when operating at quad rate), while at the same time downplaying or omitting some advantages (such as simpler/more effective filtering) of higher rates.
Anyway, point being, the topic of sampling rates has already been and continues to be discussed and opined upon in various contexts in various on/offline nooks and crannies, even apart from that "facetious scientist" blogging a so-called "science of..." rehashed article of other's work on it, e.g.
http://www.stereophile.com/reference/104law/index.html http://mixonline.com/recording/mixing/audio_emperors_new_sampling/ http://www.gearslutz.com/board/music-computers/659500-digital-audio-sampling-rates.html http://www.gearslutz.com/board/remote-possibilities-acoustic-music-location-recording/834281-recording-16-44-1-more-resolution.html#post9001744 http://www.computeraudiophile.com/f6-dac-digital-analog-conversion/oversampling-who-does-it-best-6088/ Now turning to the subject "Science of Sampling Rates", let's consider this portion:
But be careful of designers who go for super-sonic sampling rates and set their filters too high. If you include too much super-sonic information in the signal it becomes likely that you will introduce super-high frequency “intermodulation distortion” on playback.
It turns out that in many cases, we can hear the sound of higher sample rates not because they are more transparent, but because they are less so. They can actually introduce unintended distortion in the audible spectrum, and this is something that can be heard in listening tests. More on that later.
...
In any case, by increasing rates from 44.1kHz to anywhere up to about 96kHz, you might not get incredible increases in sound quality — but with all other things being equal, it shouldn’t hurt.
Once you go past that however, you introduce the possibility of not just using too much power, but of introducing unintended distortion.
Yes, there is a point where you can have too high a data rate. Some would argue that point is closer to 96kHz, but almost any computer scientist or circuit designer today will tell you that you’ve definitely reached that point by 192 kHz.
...
Since analog circuits are almost never linear at super-high frequencies, they can and will introduce a special type of distortion called intermodulation distortion.
which I call bogus, mis-info and pseudo-science.
Yes, Lavry in his articles intimated about a possible risk of ultrasonic IM distortion and a loss of accuracy, but offered no actual evidence of such (nor afaik has anyone else definitively, including the xiph article which the blogger cited). And the spiel about non-linearity of analog circuits (as if quantization were itself ever "linear"?) introducing IM distortion just plain misses the that the sampling frequencies only exist in the digital domain (as do the Sigma-Delta/Delta-Sigma converters' MHz modulator sampling frequencies), and that with effective filtering at the ADC and DAC (which is simpler to accomplish digitally and can be more effective at higher rates) a properly operating audio DAC converting "double" or "quad" rate PCM audio still only outputs audio passband frequencies at its analog output just like it does when converting good old 44.1k "single" rate PCM audio.
Ok, maybe if one's studio has an ultrasonic intrusion alarm and a microphone that picks up ultrasound, then perhaps one should avoid recording at higher sampling rates, but otherwise... Or have I somehow missed all the accounts of howling dogs, nasty audio artifacts and fried tweeters in studios using high rate audio? Bah, hogwash.
With all respect to Dan Lavry, his papers have raised the occasional question, as well as su****ion of an underlying commercial agenda. Questions like, if 44.1/48kHz are sufficient, then why even bother to offer double-rate-capable converters? Or. if 88.2/96kHz are ok because they're close enough to your suggested as perhaps more optimal 50~60~70kHz rate while actually being widely supported and adopted rates, then how exactly are 176.4/192kHz not ok in any respect differently from (as in, any "cons" thereof not applicable to) 88.2/96kHz (other than that your company's converter offerings don't happen to support quad-rate)? Maybe I missed the overwhelming evidence to which the facetious scientist referred?
But sorry, that blogger deserves no respect. Now, if the blogger had mentioned how a sampling rate of 44.1kHz had originally been adopted... (hint: think videotape, not CD)