no dither please

[dcboucher]

Anyone know how to make the exporting of files dither free? I dither with an outboard digital path (UV22HR) and need to not add the digi dither when exporting the files as 16 bit (for CD Authoring). Any way I seem to do it (other than recording to a DAT or something), I get digi dither all over it. HELP!

[Robert U]

Paul,

So you are saying that it's applied when it's gets converted from 24 to 16??? Does that mean that we don't have to put dither on the masterbus????

Now I'm getting http://duc.digidesign.com/ubb/images/icons/confused.gif.........


-Robert

[Paul Greyson]

Sorry. I wasn't clear. I'm referring *only* to conversion from 24->16 bits when using the "Exported selected regions as files" command.

The reason the dithering is automatic in this case is there's no other mechanism for doing dither at this stage (unlike during bounce where you can use a dither plugin.)

Go it?

------------------
Paul Greyson
Digidesign Engineering

[666]

Does PT appy dither only if i bounce to 24 bit file and THEN convert to 16 bit?

------------------
somewhere in the planet

[Paul Greyson]

As I mentioned in another thread, the conversion from 24->16 bits that may occur as part of bounce does not dither.

The 24->16 bit conversion that may occur as part of "Export Selected As Files" does dither.

------------------
Paul Greyson
Digidesign Engineering

[gaz]

PT's dither isn't automatic, I believe - it is selectable in the "Default Dither" option, allowing you to choose the Dither type and depth, or to choose no dither at all. The location of this option momentarily escapes me - perhaps somewhere in Preferences, or Peripherals?

Although setting the option to "No Dither" doesn't guarantee that PT won't dither it anyway, without your knowledge - some testing will be in order.


==gazzer==

[Paul Greyson]

Dither is always applied to exported files when converting from 24->16 bits.

The "Default Dither" options apply only to Audio Suite.

------------------
Paul Greyson
Digidesign Engineering

[Robert U]

Thx for the clarification!

http://duc.digidesign.com/ubb/images/icons/smile.gif

-Robert

[tld]

Yes..thanks very much for the clarification. When I bounce to disk as 16 bit after applying IDR or UV22 I always wanted to be sure that no other dither was being applied by the bounce. This finally clears it up. Thanks!

Tom


------------------
http://www.digitalaudiorock.com

[dcboucher]

The problems are: 1 bouncing to disk does not sound as good as recording the mixes to two new tracks. I know I'm not the first to discover this. Bounce to disk is taboo in the pro world. 2 recording into a 24 bit session with dither applied at the converter stage (i.e. UV22 on an AD8000) is not applicable with bounce to disk.

[blairl]

Originally posted by dcboucher:
The problems are: 1 bouncing to disk does not sound as good as recording the mixes to two new tracks. I know I'm not the first to discover this. Bounce to disk is taboo in the pro world.

Although many have claimed that bounce to disk degrades the sound quality, it is absolutely untrue and scientificly impossible. Do your own scientific tests to decide for yourself. Sometimes peoples ears seem to hear what they want them to hear. I'm not trying to be rude or anything, I'm just trying to clear up some of the hype that's out there. Try this test and make your own decision. Try it with 2 tracks or 32 tracks whichever you prefer.

1) Bounce a session to disk and import the results back into the same session. (make sure the bounce is the same bit depth as the original session), (don't use any delay based plug-ins such as reverbs, choruses, etc.. because these are all random and will never reproduce the same thing twice)

2) Assign all of the original outputs to bus 1-2, (or whatever bus is available).

3) Create a stereo Aux input and assign the input to bus 1-2 (or whichever you used), and assign the output to Output 1-2.

4) Create two new audio tracks and put the bounce results on those two tracks.

5) instantiate "Time Adjuster" plug-ins on these two new tracks and flip the phase on both of them.

6) Line the bounce up "sample accurate" with the original material.

7) Create a master fader if you already haven't

8) Command click on the display area of the master fader until you get the "peak" option on the meter.

9) Start playback making sure that the bounced material is lined up sample accurate with the original tracks, then control click on the numerical readout on the meters.

The result will be infinity or 100% cancellation. There is absolutely no low level noise or garbabe going on. In other words, the bounce is exactly the same as the orignal material. There is no narrowing of the stereo image, there is no distortion, there is no level change, it is exactly the same as the original, so the sound quality is not worse, it is the same.

Do this test for yourself, then make your own decision.

[cjguitar]

I did the test with a stereo track of music. I bounced one set and recorded another set. With the phase reversed for both I still got around -90db signal. This was a 16 bit session in a PT III system, with no plugins, gain set at 0 and no dither. What does that mean?

[blairl]

There may be several problems. First they may not be lined up sample accurate. Try nudging the bounced tracks 1 sample at a time in each direction and see if that changes anything. If that doesn't work.... Are you using the 24 bit optimized mixer or the 16 bit optimized mixer with your PT III hardware? Do you do any "convert after bounce"?

[Robert U]

What was your set up for the recorded version? Didi it go AES/EBU, SP/DIF or analog?

-Robert

------------------
"Are we going to do it with the computer or are we in a hurry?"

[cjguitar]

This isn't a phasing issue at all. The tracks were all lined up to the sample (I zoomed in as far as I can go and nudged the tracks so the waveforms lined up exactly). I'm using the 16 bit mixer, only because the 24 bit mixer doesn't work on PT III systems (24 bit isn't supported at all). The song came off of a CD that I recorded (import audio from other movie). I could try this test on a Mix plus system later today.

[Paul Greyson]

I wrote the code, so I can tell you with certainty that the data that gets written when you bounce is EXACTLY the same data that would get written if you recorded from a bus. The only difference between bouncing and recording from a bus is that the CPU demands for bounce are slightly different so automation performance may not be identical.

I can't think of any reason that you wouldn't be seeing 100% cancellation.

------------------
Paul Greyson
Digidesign Engineering

[Paul Greyson]

Well Speer, maybe I shouldn't bother since you say you can't be convinced, but here goes:

- You'll notice that bouncing to stereo interleaved requires a conversion. I believe that's conisistent with Pro Tools' lack of support for stereo interleaved files.

- I don't understand your "recording doesn't have conversion options" comment. Bounce doesn't have conversion options either if you bounce to the internal format (24 bit, current sample rate.) (There's an additional subtlety here in that you can record to 16 bit files without "conversion". I believe there's a mixer mode that does the 24-16 bit truncation as data is being written from the mixer into host memory. We could probably take advantage of that in bounce as well and avoid using the CPU to do the truncation. From a "sonic quality" perspective though, there's no difference. Truncation is truncation.)

- wrt your blind tests, what can I tell you? The data's the same. It comes off the mixer in the identical way that recorded data comes off the mixer. The only difference is mechanism by which the data gets written out to files. That mechanism is different to allow more convenient on-the-fly conversion to your desired final format. That's it.

------------------
Paul Greyson
Digidesign Engineering

[This message has been edited by Paul Greyson (edited April 26, 2001).]