Dithering internally???

If my project is at say, 24 bits, and I want to render a a midi track with Dimension Pro into a Wav file, why is the dithering box there?

Because if your rendering is set to 32bit in sonar, like it should be, then youll be bouncing and freezing to that bit rate. But if your rendering is set to 16bit, and your project is at 24bit, then you need to dither when bouncing or freezing.
i hope this clears it up for you
Cj

...then again, if I were to have everything at 32 bits, then I wouldn't need to dither internally right? Or, if I had the rendering tab at 24 bits on a 24 bit project?

Ok, thanks CJ!

can I still share these 32/48 WAV files with other DAW's?

Its 2 different things here. Its an internal conversion to 32bit. No sound card as of right now can record or playback over 24bits. the 32bits for rendering that cake offers and you should take also is for giving you extra headroom per say for effects and processing of effects, like bouncing and freezing. If you set your render to 24bits, you missing the boat on sonars 32bit rendering for effects and processing.
To answer your question, your files will need to be exported to 24bits anyway. No soundcard can play 32bits or 64bits
Cj

To put that into perspective, picture a stack of sheets of regular printer paper 2 miles high. It may or may not be short by one sheet. Do you care? Would it make any sense to randomly add a single sheet to the stack to correct the potential error?

Good analogy. you know if they didn't offer the dithering for this specific thing, you know people would have made a huge fuss over it and maybe condemned sonar as not being professional. There damned if they do and double damned if they
dont.
Cj

ORIGINAL: bitflipper

Assuming your render bit depth is 32 (that's the default and that's what's recommended), any internal bounces will result in a 32-bit file, even if the original file was 24 bits. Obviously, no dither is necessary when going from 32 to 32 or 24 to 32. So it doesn't matter.

Bit, I too, like another poster mentioned in this thread, run all my stuff at 24 also and don't use the 32 bit option. I couldn't tell much of a difference when I used it and felt it may have been processing for the sake of processing? If my card can't support this bit rate, what are the advantages of using it and what differences should I see or hear?

Sonic, that would be correct Sir. :) I dither in WaveLab 6 where I do all of my mastering. When I export out of Sonar, I export at 24/48 with no dithering options enabled. When I have in the past, pwr 3 seemed to work the best for me. But I've been taught to always master in the same bit/sample the project was recorded in to preserve the best audio. When I am completely done, this is when I change sample rates and then dither to 16bit. Though I love Sonar, I've never liked mastering inside it so I never dither while exporting out of Sonar.

Danny, it looks like we have the same work habits

Matson, good question. I'm not sure on that. My original issue was dithering internally within Sonar, but I read somewhere around here that Sonar would automatically apply the dithering in your case...

You definately have a point. I guess it just depends how anal we want to be about it and if it's actual worth being anal - that's anal and not analyst. If Paul McCartney were an anal person, he would have never used a Hofner or an Epiphone!

Here's what I do:

Set audio driver to 24 bit. This will deliver 24 bits to Sonar when recording, and play out 24 bits from Sonar.
Set record bit depth to 24. Since that's what the soundcard is delivering, that's what should be recorded. Having the record bit depth set to 32 offers no advantage, since the soundcard only delivers 24 bits.
Set audio mix engine to 64 bit. If you have the CPU horsepower, then there's no reason not to use the higher resolution processing.
Set render bit depth to 24 and...
Set dither to triangular.

Regarding Sonar's dither: Always set it to triangular. Sonar is programmed intelligently enough to not apply dither when it's not necessary. So when you're recording at 24 bits, and laying down 24 bit files, Sonar won't apply dither. And if you bounce to 64 bit float files (or 32 bit float files, last time I checked) Sonar also will not apply dither. When you bounce down to 24 bit files (from the 32 or 64 bit float audio engine) Sonar will apply dither, as it should since you are reducing bit depth in those cases. If you mistakenly set something to 16 bit (which I've done...) Sonar will dither appropriately.

When you want to deliberately export a 16 bit file, the export dialog lets you choose a different dither if you so desire.

Just set Sonar's dither to Triangular, and you're good. Sonar knows when and when not to dither.

Here's a good thread/read on dithering.

RE: Pattor says: let your ears decide whether dither at 24 bits or not

Make sure to have snacks and coco/coffee/beer handy.

ORIGINAL: Danny Danzi

But I'm still wondering and hoping for an answer from someone about this....If my card can't support this bit rate, what are the advantages of using it and what differences should I see or hear? The only thing I notice is a bit more processing use as well as hard drive use going up a few notches. There is no sound difference at all and nothing that I can physically hear or see. So why would I or wouldn't I use 32 bit? I did 2 small projects using 12 tracks of audio with a nice batch of plugs. One recorded at 24/48, the other at 32/48. The only difference I *see* is 3% increase in cpu and hard drive. There is no difference in what I *hear*. Is it something I'm supposed to hear? If someone replies "no, you probably won't be able to hear it" then why would we use it? If it's not audible enough to make a difference, obviously there has to be some benefit from using this setting? Anyone?

This has been discussed already. It's also in the help file. If you don't understand, slow down and re-read the posts.

Rendering at 32-bit (or 64-bit) is recommended to prevent round-off errors.

No, you can't hear the difference. I can't hear the difference between 16-bit and 24-bit audio, but I use 24-bit audio during engineering tasks to maximize head room and minimize error, as already stated.

To output files for listeners, bounce down to 16-bit using the best dithering algorithm available.

To output files for use by other engineers use 44.1Khz, 24-bit.

And when reducing audio file resolution, use dithering to reduce aliasing.

Hey guys - I'm hoping someone can help me with a newbie question regarding Audio Driver Bit Depth. In Sonar 7 I'm using an ASIO driver for my Edirol UA25EX. I've got everything set up and seemingly working. Under the General tab in the Audio Options dialog I have the UA-25EX set for my Playback and Recording timings; the Audio Driver Bit Depth is defaulted (uneditable) to 24 bits. My Default Settings for New Projects is set to a sampling rate of 44100 (which matches the switch on the back of the unit).

However, after creating a new project I happened to look at the file info stats (File>Info) and see the following:
Editing time: 00:18:45
Revision: 1
Events: 0
Sample Rate: 44100 Hz
Bit Depth: 16
Pan Law: 0dB center, sin/cos taper, constant power
File Version: 16.0.2.243

Am I missing something here? Even though I have it set to 24 bits is it really 16 bits? How do I know I'm recording at 24 bits. Not sure I'm even asking this question the right way.

Any help would be appreciated.
Thanks.
Rob

Thank you, Jim! I must have looked through all the settings so many times that it became a blur.

Much appreciated.
Rob