Protools 192k

[Joz]

A recording engeneer from a big studio in LA, told me that digidesign is going to skip doing protools 96k, instead they are working on 192k of resolution. I'd like to know if this is true, cause that would solve alot of the problems that digital recordings have.

[peteri]

Sure, one of the things it would certainly allow is better EQ curves (which no-one seems to mention).

BUT it will also cause a lot of problems, both storage and DSP wise, and I'd be interested to see how the PCI buss will deal with this amount of traffic (64 tracks at 192k anyone?)

Pete

[emilano]

Well, I'd like to open a discussion on the proposed benefits of higher sample rates. Or should I just do a DUC search? (pause for tomato throwing)...

Since many monitors aren't designed to playback frequencies much above 20hz, what is the point? Same goes for many microphones....Not so for many pre-amps though, many of the higher end models profess to be good way above that. But if your mic and speakers aren't recording/playing them, what's the point? Do well have to upgrade everything in our signal path to be able to reproduce these freqs once the new systems come out?

The only thing I can think of is that perhaps the way higher frequencies make things sound better is by having some sort of "influence" on the lower ones. Like maybe somehow causing them to resonate in some way or something...just going on intuition here...If this is true in someway, then you wouldn't need to upgrade everything, but then again, then I guess we wouldn't even need the 192khz systems...

I've read a study that showed that people did show more brain wave activity when played music with the higher freq's but the music was played back on monitors that were proven to be playing back freqs up to 50khz.

Anyone know more about this? Okay, okay, I'll do the search! (wipes tomato from eye).

[mpayne]

http://duc.digidesign.com/cgi-bin/ub...&f=16&t=010660

Has a fair bit on this topic

Malcolm

[CO2]

Hello Everyone:

The future of digital???

web page

Hopefully, we will know in a bit.

Best Regards

[F Umminger]

emilano-

I am extremely skeptical that anything, let alone anything of importance, can be heard above 20khz . Nevertheless, higher sampling rates can be beneficial for signal processing even if ultra-high > 20khz frequencies can neither be heard nor even reproduced by the playback system. This is especially true for highly nonlinear signal processing,

A simple example of this is the process of squaring a signal. If one squares a sinusoidal wave then the result is some dc and a sinusoidal wave at twice the original frequency. So if your sample rate is 48khz and the input sinusoid has frequency 15khz then the output is at 30khz, which is 6 khz above Nyquist = 24khz. Therefore the output aliases down to 24khz - 6khz = 18khz and one hears a spurious anharmonic 18khz tone when one should be hearing (or rather, not hearing) a 30khz tone. Even low-pitched tones can generate strong high-frequency harmonic content, that will then alias, when nonlinear processing is applied.

This is a real problem when designing digital distortion effects. It also is a real (albeit lesser) problem for dynamics processing. The problem can be solved at lower sample rates given enough engineering effort and enough dsp cycles, but it is much easier to design a good, efficient nonlinear effect at higher sample rates.

It is also much easier to design virtual analog synthesizers with low aliasing noise if one is working at higher sample rates. In this case the powerful high harmonics of the traditional saw and square waves will alias badly unless one is careful and clever in designing the digital oscillators. Less care and cleverness is needed at higher sample rates.

Another issue is the design of simple linear eq filters. These are often designed by applying the bilinear transform to an analog filter transfer function. The bilinear transform only accurately reproduces the analog response in the range of roughly 0hz to 1/2 Nyquist, and the result is that at a 48khz sampling rate the frequency response in the range 12khz-24khz is not quite right, while at a 96khz or higher sampling rate the frequency response is accurate over the entire audible range. The better eq plugins will not exhibit this problem because they will have been designed using more sophisticated techniques than the bilinear transform. As a result they will show less improvement at higher sampling rates.

Finally, fractional sample delay interpolators, which are needed for modulating delay lines and chorus effects, are difficult to design for low sample rates and thus they often suffer from some amount of aliasing artifacts. At higher sample rates these are much easier to design and the aliasing artifacts will generally be much reduced.

In summary, I believe that higher sampling rates will lead to improved audio but that, paradoxically, the benefits will accrue primarily to the most highly processed audio and not to the kinds of unprocessed audiophile recordings that people usually think about when thinking of
higher sampling rates. Similarly, the most amateurishly designed plugins will be the ones the most improved by higher sampling rates. The improvements have nothing to do with any purported abilities to perceive ultrasonic frequencies but instead have to do with more accurate signal processing calculations.

This is all based on theory. I have not done any listening tests to confirm any of this.

[emilano]

Interesting...Even more interesting to me is that I actually could follow what you were saying...Though I had to read it three times [img]images/icons/smile.gif[/img].

[lwilliam]

One argument I've heard in favor of higher sampling rates as far as the PERCEPTION of audio (vs plugin design, etc.) is that, for instance, when a violin section plays together, frequencies above 20K will still be able to interact with those frequencies below 20K to create certain subtle "beat frequencies" that would only occur when recording an ensemble.

However, I haven't fully grasped why this wouldn't also be the case with an ensemble recorded together at 48K hz. Technically, those sub-20K hz "ensemble-created" interactions (from the ultrasonic harmonics) should still appear on a 48K recording given that the group was recorded with half-decent equipment.

Anybody care to explain this to me, or is this just hogwash?

[snoopy]

My edjumactud guess is that sample frequency will have an affect on track count. In other words, threy will probably have it work like :32 tracks of 192k, 64 tracks of 96k, 128 tracks of 48k. I have no idea of the actual track count or sample freqeuncies they will use though. And I imagine the same would hold true for DSP. The higher the sample rate, the less DSP you get per chip/card.

I wish there was a way they could just avoid the PCI buss all together. Have a box that does everything externally and the computer is nothing more than an interface (even hard drives would have to be connected to external box). AS plugs would be done in the box, etc, etc. But who knows, there's probably 100 problems associated with that idea...

[Joz]

I think that its not so important that you cant hear above 20k, the important thing is that with 192k the recorded sound will be more truth to the original source. Wich in theory you'll have a better recording