32bit Bounce?

[Numi]

I am curious if the new update allows the export of the final bounce to 32-bits, and before anyone lectures me on if it really matters are not..I am just curious

I know the audio engine is 32-bit and it would be nice to export the final mixdown without any internal dithering done from the engine down to 24-bits.e.g.Plug-ins and Reverbs and such...

It's pretty common with most DAW's now a days..but even if this isn't included in this new update..I still think Pro Tools 8(since I don't have 9 yet) sounds better then any DAW I have used to date. So is this a new feature of 9, or is this not really necessary from AVID's point of view?

Thanks!

P.S. Thanks AVID for listening to your costumers and adding in most of the features we have been asking for e.g.ADC..and the lot. I think the price is FAIR...and most people already have an iLok if they have actually "paid" for any of their plug-ins

[albee1952]

Audio engine "bits" and recording "bits" are 2 different things AFAIK. Upping the bit rate after recording(which most of us do at 24 bits) would, in theory, offer no improvement in sound quality.

[Numi]

64-bit and 128-bit floating point processing
All the performance gains described above give rise to a platform that enables new, more powerful, processing systems. In particular, modern CPUs are making 64-bit double precision floating point processing a reality. 128-bit floating point processing is just around the corner.
64-bit double precision processing provides increased dynamic range and sonic clarity over 32-bit floating point systems. An IEEE 64-bit float has 52 bits of mantissa plus a sign bit, giving 53 bits of precision when doing calculations. This amount of precision means audio fidelity is maintained even with dramatic gain staging within processing elements, and also means more accurate processing for recursive DSP such as IIR filtering.
Double precision mixing is especially important when processing 24-bit PCM audio data. Consider the following test program, which simulates mixing one sample in a 3 track project with the same file on each track. Track 1's gain is +6dB. Track 2's gain is -6dB. Track 3's gain is also +6dB and the track's phase is flipped. All 3 tracks are mixed into a bus whose gain is +6dB. The signal produced on the bus output should be the same as the track data but at unity gain, since tracks 1 and 3 would cancel each other out, and the -6dB on track 2 would be offset by the +dB on the mix bus.
void test()
{
}
const double dScale24 = (double)(1 << 23); const double dGain = 2.0; const double dAtten = 0.5;
const int nPCM24Orig = 0x5000FF; double dSamp = (double)nPCM24Orig / dScale24; float fSamp = (float)dSamp;
float fSum = 0; fSum += (float)( dGain * fSamp ); fSum += (float)( dAtten * fSamp ); fSum += (float)( -dGain * fSamp ); fSum = (float)( dGain * fSum ); int nPCM24FromFloat = (int)(fSum * dScale24 + 0.5);
double dSum = 0; dSum += (double)( dGain * dSamp ); dSum += (double)( dAtten * dSamp ); dSum += (double)( -dGain * dSamp ); dSum = (double)( dGain * dSum ); int nPCM24FromDouble = (int)(dSum * dScale24 + 0.5);
printf("Original: %lX, FromFloat: %lX, FromDouble: %lX\n", nPCM24Orig, nPCM24FromFloat, nPCM24FromDouble);
The output produced by this test program is (surprisingly):
Original: 5000FF, FromFloat: 500100, FromDouble: 5000FF
-4-
Benefits of Modern CPU Architectures for Digital Audio Applications (continued)
What this simple program shows is if X is a 24-bit PCM sample, and the math is done using 32-bit floating point, an inaccuracy is introduced due to summation. In this case the least significant bit is lost. If the gain adjustments are more dramatic, or more gain stages are used, then more bits can be lost.

This is a little excerpt from Cakewalks White Paper on Processor and Bit Depth as well as benefits of 64-bit floating point...but also helps explain my point to some degree. I hope, although I have been wrong before..

Here is the whole article http://www.cakewalk.com/Press/Cakewa...plications.pdf

I know 24bit is what the hardware records at..I know 24bit is what we all usually record at..I know that the audio engine for Pro Tools is 32bit Float..My question is if it would be possible to eventually export the mixdown(bounce) in its most truest form.


The reason I bring this up...it's nice to have this option for Mastering. Although, not really necessary if Pro Tools hasn't ever had this export option before...

[nikki-k]

IIRC...

PT9 (non-HD) is utilizing a 64-bit Float mix engine.
PT9 is a 32-bit application
These two bits have absolutely no relationship with one another.

PT9 is not capable of exporting/bouncing internally to produce a 32-bit file.
There is one reason I am aware of for it, and admire Avid for not wasting time on it.

While yes, our hardware is typically capable of rendering 24-bit (max), the reason(s) for the mix engine to operate at 64-bit float has nothing to do with that. A 24-bit converter is capable enough for 99% of all applications. But, to manipulate those numbers without *losing* information? 64-bit float nearly guarantees it. So, with a nice 64-bit float *number word* sitting there, one could *reduce it* to a 24-bit fixed, or even a 32-bit floating *number word* if they so desired. Typically, we will utilize the 24-bit fixed option for a final delivery method. However, one could argue a (slight?) benefit in saving that resulting 64-bit floating *number word*, and thus have a "more pure," unmolested "file" to then apply further processing to later... say, from PT9 to a mastering app that can accept a 64-bit floating file(s). *Reducing* that 64-bit floating result to a 32-bit floating file? Eh.

I am sure someone else could find reason(s) to argue all that...
And I could be very mistaken. Too early for me, and my puff-puff meds ran out...