48K vs 44.1 - That big a deal?

[Loudnoize Ent.]

Well I certainly got way more than I bargained for, especially Nika for participating in this topic and sharing his in depth theory. Thank you Nika and to you all...

As a more experienced producer/writer than I am an engineer, I tend to hear things from an artistic point, rather than a technical one. A few months ago I started a session only to realize later my settings were at 16/44. That won't happen again but I didn't complain and no one else did either. I come from the school where if it's a "feel good" situation, well then carry on. It amazes me at how many of us tend to hear something that is recorded at higher sampling rates (192k or even 96) which I personally believe isn't really there. Maybe I am wrong and I won't discredit the ones who think there is but all in all, this thread has given me a broader outlook and at how I will continue working.

Although it has been discussed at length but never the less fascinating, and for some this may be old news, I've decided to share with you a few paragraphs from Sampling Theory For Digital Audio...

.........................

The notion that more is better may appeal to one's common sense. Presented with analogies such as more pixels for better video, or faster clock to speed computers, one may be misled to believe that faster sampling will yield better resolution and detail. The analogies are wrong. The great value offered by Nyquist's theorem is the realization that we have ALL the information with 100% of the detail, and no distortions, without the burden of "extra fast" sampling.

Nyquist pointed out that the sampling rate needs only to exceed twice the signal bandwidth. What is the audio bandwidth? Research shows that musical instruments may produce energy above 20 KHz, but there is little sound energy at above 40KHz. Most microphones do not pick up sound at much over 20KHz. Human hearing rarely exceeds 20KHz, and certainly does not reach 40KHz. The above suggests that 88.2 or 96KHz would be overkill. In fact all the objections regarding audio sampling at 44.1KHz, (including the arguments relating to pre ringing of an FIR filter) are long gone by increasing sampling to about 60KHz.

Sampling at 192KHz produces larger files requiring more storage space and slowing down the transmission. Sampling at 192KHz produces a huge burden on the computational processing speed requirements. There is also a tradeoff between speed and accuracy. Conversion at 100MHz yield around 8 bits, conversion at 1MHz may yield near 16 bits and as we approach 50-60Hz we get near 24 bits. Speed related inaccuracies are due to real circuit considerations, such as charging capacitors, amplifier settling and more. Slowing down improves accuracy.

So if going as fast as say 88.2 or 96KHz is already faster than the optimal rate, how can we explain the need for 192KHz sampling? Some tried to present it as a benefit due to narrower impulse response: implying either "better ability to locate a sonic impulse in space" or "a more analog like behavior". Such claims show a complete lack of understanding of signal theory fundamentals. We talk about bandwidth when addressing frequency content. We talk about impulse response when dealing with the time domain. Yet they are one of the same. An argument in favor of microsecond impulse is an argument for a Mega Hertz audio system. There is no need for such a system. The most exceptional human ear is far from being able torespond to frequencies above 40K. That is the reason musical instruments, microphones and speakers are design to accommodate realistic audio bandwidth, not Mega Hertz bandwidth.


Record at 192KHz than process down to 44.1KHz?

There are reports of better sound with higher sampling rates. No doubt, the folks that like the "sound of a 192KHz" converter hear something. Clearly it has nothing to do with more bandwidth: the instruments make next to no 96KHz sound, the microphones don't respond to it, the speakers don't produce it, and the ear can not hear it.

Moreover, we hear some reports about "some of that special quality captured by that 192KHz is retained when down sampling to 44.1KHz. Such reports neglect the fact that a 44.1KHz sampled material can not contain above 22.05KHz of audio.

Some claim that that 192K is closer to the audio tape. That same tape that typically contains "only" 20KHz of audio gets converted to digital by a 192K AD, than stripped out of all possible content above 22KHz (down sample to CD).

“If you hear it, there is something there” is an artistic statement. If you like it and want to use it, go ahead. But whatever you hear is not due to energy above audio. All is contained within the "lower band". It could be certain type of distortions that sound good to you. Can it be that someone made a real good 192KHz device, and even after down sampling it has fewer distortions? Not likely. The same converter architecture can be optimized for slower rates and with more time to process it should be more accurate (less distortions).

The danger here is that people who hear something they like may associate better sound with faster sampling, wider bandwidth, and higher accuracy. This indirectly implies that lower rates are inferior. Whatever one hears on a 192KHz system can be introduced into a 96KHz system, and much of it into lower sampling rates. That includes any distortions associated with 192KHz gear, much of which is due to insufficient time to achieve the level of accuracy of slower sampling.

.........................

Let's do keep this thread alive and thank you all again.

[love666]

The benefit is a bandwidth issue.

And that is? What can you deliver in 24 bits that can't audibly be done in 16?

Nika

even using [IMO] the best-sounding dither (pow-r type 2) in [IMO] the best-sounding software (sonic studio hd) -
24-bit program material sounds better without the dither instantiated

as far as i know the tools available are not up to the mathematically possible task

[Nika]

What are the playback levels? And have you tried POW-r type III?

Nika

[Mount Royal]

even using [IMO] the best-sounding dither (pow-r type 2) in [IMO] the best-sounding software (sonic studio hd) -
24-bit program material sounds better without the dither instantiated

I can only agree. My mixes sound better undithered, to be sure. But you'll probably also agree, Love, that it's not the benefit those extra 8 bits are affording. Rather it's the process of lopping those 8 bits off (to yield 16) that's destructive. Certainly we can't actually use the dynamic range 24 bits can convey in a final mix, given listening environ noise floors and loudspeaker system limits. At the rate we're compressing things these days, I wonder if we truly need even 16 bits in the end product.

John-

[Mount Royal]

We have to delineate the physiology of hearing from the psychology/perception of hearing. ... What we really need to do is determine the bounds of the ear, for it is the large filter on the system. If we can ascertain the limits of the ear wrt frequency, phase, amplitude, and dynamic range - all at a physiological level - then we can design audio systems that exceed that.

... We don't actually need a model of the ear - just a measurement of its boundaries and enough of a model to understand/interpret why in order to validate/invalidate our measurements.

Nika

This is attractive to be sure Nika: the idea that the ear's behavior can be characterized as you suggest - writing a transfer function of sorts for the ear and designing your recording system to represent all data that falls within the bounds of that function. A few items still trouble me, though.

Our goal was to characterise how an audio device must perform to be measurably as good as, or worse than, some standard related to sound quality. As you are aware, the middle ear is the transducer, your A/D converter, which represents only one of the inputs to the acoustic nerve. The acoustic nerve which receives conductive input from the body as a whole, and the skull most of all. Further, the acoustic nerve represents only one in the brains inputs which contribute to sound perception. Our other tissues resonate sympathetically with sound, especially since the era of amplification, and events like chest resonance at rumble frequencies contribute to sound perception.

So the thought is that delineation of the physiology of hearing needs to entertain these pathways which operate in parallel with the predominant path of [outer ear > middle ear mechanical components > middle ear hair cells (A/D) > acoustic nerve > brain stem]. To my knowledge, the parallel paths of hearing conduction and body cavity resonance are paths associated with strong low pass filters. Since much of our concern with conversion lies in the upper audio band, such parallel paths might be incorporated into the measurements of "hearing" without much trouble. Any thoughts here?

John-

[love666]

I can only agree. My mixes sound better undithered, to be sure. But you'll probably also agree, Love, that it's not the benefit those extra 8 bits are affording. Rather it's the process of lopping those 8 bits off (to yield 16) that's destructive. Certainly we can't actually use the dynamic range 24 bits can convey in a final mix, given listening environ noise floors and loudspeaker system limits. At the rate we're compressing things these days, I wonder if we truly need even 16 bits in the end product.

John-

yeah. the problem is dither. but what is the solution? better dither? dvd-a? i don't know. i don't really care any more. i plan to assemble tracks of no more than 8 channels in pro tools at 24/96, import the files into sonic, src, dither, sum, burn and that's it. the 24/96 archives can probably be used when i'm dead, if anyone cares.

[love666]

What are the playback levels? And have you tried POW-r type III?

Nika

playback levels - full range
POW-R type 3 - yeah, don't like it
thanks
- dave

[Nika]

I can only agree. My mixes sound better undithered, to be sure. But you'll probably also agree, Love, that it's not the benefit those extra 8 bits are affording. Rather it's the process of lopping those 8 bits off (to yield 16) that's destructive. Certainly we can't actually use the dynamic range 24 bits can convey in a final mix, given listening environ noise floors and loudspeaker system limits. At the rate we're compressing things these days, I wonder if we truly need even 16 bits in the end product.

John-

16 bit material, on its face, conveys 96dB of dynamic range. If you had your music turned up to 96dB SPL (peak), the noise floor would be at 0dB SPL. With a little bit of noiseshaping the noise level can be dropped precipitously below that. The noise floor can be shoved below the threshold of hearing for all frequencies in the audible spectrum. Why is this inadequate? Why do we need to maintain 144dB of dynamic range when the ear's dynamic range is hardly 100dB? Why do we need to preserve lower noise floors than we can hear? Why is the process of dithering to 16 bits audible destructive?

Bob Katz has said that when he uses good noiseshaping and reasonable listening levels (K-14), 85dB SPL RMS, he can't hear the difference between 16 bit and 24 bit material so long as it is shaped properly. What are you hearing that he isn't hearing?

Nika

[love666]

16 bit material, on its face, conveys 96dB of dynamic range. If you had your music turned up to 96dB SPL (peak), the noise floor would be at 0dB SPL. With a little bit of noiseshaping the noise level can be dropped precipitously below that. The noise floor can be shoved below the threshold of hearing for all frequencies in the audible spectrum. Why is this inadequate? Why do we need to maintain 144dB of dynamic range when the ear's dynamic range is hardly 100dB? Why do we need to preserve lower noise floors than we can hear? Why is the process of dithering to 16 bits audible destructive?

Bob Katz has said that when he uses good noiseshaping and reasonable listening levels (K-14), 85dB SPL RMS, he can't hear the difference between 16 bit and 24 bit material so long as it is shaped properly. What are you hearing that he isn't hearing?

Nika

psw mastering dither thread

[Nika]

Our goal was to characterise how an audio device must perform to be measurably as good as, or worse than, some standard related to sound quality. As you are aware, the middle ear is the transducer, your A/D converter, which represents only one of the inputs to the acoustic nerve.

The middle ear is one of the transducers, as is the tympanum on one end and the round window on the other end. The basilar membrane is also a transducer. The actual A/D conversion is done past the middle ear, at the inner ear.

The acoustic nerve which receives conductive input from the body as a whole, and the skull most of all. Further, the acoustic nerve represents only one in the brains inputs which contribute to sound perception.

Do you have citations?

So the thought is that delineation of the physiology of hearing needs to entertain these pathways which operate in parallel with the predominant path of [outer ear > middle ear mechanical components > middle ear hair cells (A/D) > acoustic nerve > brain stem]. To my knowledge, the parallel paths of hearing conduction and body cavity resonance are paths associated with strong low pass filters. Since much of our concern with conversion lies in the upper audio band, such parallel paths might be incorporated into the measurements of "hearing" without much trouble. Any thoughts here?

First, are you assuming that in measuring the boundaries of the ear that the scientists (otologists, audiologists, and researchers at Bell Labs, AT&T, etc) over the past 100 years have failed to take into regard any other senses that contribute to the perception of sound? FWIW, I have never seen any evidence that the acoustic nerve is subject to any consequential amount of "conductive input from the body as a whole." Where would this "conductive input" come from? I have indeed seen literature about alternate ways in which sound can get to the inner ear without passing through the middle ear, but this isn't about conduction to the auditory nerve - it's about conduction to the inner ear, again relegating the inner ear to be the physiological filter of our hearing. There are bone conduction studies that have been done, as well as analysis of tissue conduction. In each of these cases, though, the frequency spectrum is heavily low-pass filtered, in addition to being extremely low in amplitude. As such, when coming up with a model of the ear's frequency response it seems hardly relevant - even IF we were to ignore the fact that it has already been factored into tests previously performed. Since all of the bone-conduction, etc. gets transmitted to the brain through the actions of the basilar membrane (just through a different conduit of getting TO the basilar membrane) we are still restricted to the physiology of the ear. The VIIIth auditory nerve does not feed from other parts of the body. The auditory cortex, as far as I am aware from my readings, does not feed from any other than the VIIIth auditory nerve.

There seems to be a pretty large auditory-research-denial industry about these days, especially in advance (or in lieu) of understanding the research already done. I'm perplexed by this...

Nika