should I dither ...

[hao nao]

if I'm using (a) 24-bit drum loops or (b) a plugin (ManyStation) that has 24-bit samples, but am rendering an edit to a 16-bit wav?

thanks!

[Lunch Money]

Yes. Without dither, you are going to simply truncate bits. Often, this won't be noticeable at first, so you'll think all is peachy. But all is not well in 16-bit land, my friend. Dither!

Greg

[dealwithit]

How to dither though? Use T's built-in dither? I've heard to use mdaDither, but I'm not sure. I've never dithered anything with T, and it's usually rendered to 16 bit for conversion to mp3. I have mdaDither, but the settings make no sense to me, and I'm not even sure where to put it (last in Master plugins, I assume).

Any help?

[aMUSEd]

Stop dithering and just deal with it!






sorry - couldn't resist

[Todd24]

I could use some help here as well.

Thanks in advance

[IIRs]

I haven't done much testing yet, but I seem to prefer the sound of the mda dither to Tracktion's own. To use it properly, make sure T's own dither is turned off (settings page -> audio devices -> select the input -> "dithering enabled") and insert mda dither as the last plug in the master section. Make sure the master fader is at unity!! If you need to adjust the level, add another Vol / Pan before the dither plug. If you're going to 16-bit, you can probably just stick to the default settings, but feel free to experiment with the different dither types. (I would expect the "N.SHAPE" option to be best though). The word length should be set to your destination bit depth, and "dither amp" and "dc trim" are probably best left well alone.

[nuisance sonore]

I find this issue confusing as well. T own dither settings seems to be linked to the audio output settings, while everything I read seem to hint to the fact that audio card settings shouldn't have any bearing to the rendered output. And I can't find any actual project settings in T, just audio card settings.
For the moment I leave all the audio settings at 24 bits and render as such, then I use Voxengo R8brain to downsample/dither if I need to, but I'm gonna have a go at Platinumears method soon.
I feel this is one aspect of T that could really use some clarification thought.

[Mr. Tingle]

settings page -> audio devices -> select the input -> "dithering enabled"

Did you mean "dithering disabled"?

[IIRs]

settings page -> audio devices -> select the input -> "dithering enabled"

Did you mean "dithering disabled"?

I was quoting verbatim the name of the button that should be switched off. So yes, I did really..

[Lunch Money]

Make sure the master fader is at unity!!

How come? I mean, I put mine at unity anyways, because isn't that the ideal "target" for a recording? But specifically, why is it absolutely necessary to do this?

Greg

[IIRs]

I feel this is one aspect of T that could really use some clarification thought.

Agreed, hence this thread on the other forum..

[IIRs]

Make sure the master fader is at unity!!

How come? I mean, I put mine at unity anyways, because isn't that the ideal "target" for a recording? But specifically, why is it absolutely necessary to do this?

Greg

because the gain of the noise added in the dither process is critical (it must be just loud enough to modulate the lowest bit of the target file) so any gain changes you apply between adding the dither noise and shortening the wordlength will "break" the dither effect.

I'm not one of those people who say the master fader should be left at unity anyway btw.. its good practise in an analogue desk (esp. a cheap one) but a bit pointless with a 32-bit mix engine IMO. The only time I strictly make sure its at unity is when I have a dither plug in the master section, and I often drop a Vol / Pan into the master section aswell to use instead of the real master fader. (you can even use the normalise functions: just remember to transfer the gain value from the master fader to the new Vol filter you add, and zero the master.)

[nuisance sonore]

Agreed, hence this thread on the other forum..

Yes, I've read this thread. What still confuses me is what happens if you have different audio output with different audio dither settings, some enabled and some not? For instance, I can route my main track to the SPDIF output connected to an external decoder for monitoring (on which I tried dither on/off and couldn't hear the difference so I left it off figuring I'd save CPU), but have a few tracks routed to another track (I believe it's called sub group?) which goes out to analog output for outboard processing (send?) with dither enabled and then come back in to a track routed to the main track.
What would Tracktion do in such a case? Dither the rendered output or not?
I believe the method you propose has the merit of at least being clear about when and where the dithering is applied.

[Beno]

Tracktion adds dither at the mix bus for that audio output. So this is why you can enable dither for each output independantly. It is applied to that output's mix bus during playback and rendering.

Whether you use the Tracktion dither or use a third party dither is a matter personal preference. But keep in mind that the subtleties of dithering algorithms has very little to do with your final output compared to the myriad of other options available to you. Mics, Preamps, converters, soundcards, plug-ins, even electrical power, all have much more impact on the final product.

So for me, I simply enabled dither for all my outputs and forget about it. No need to worry about settings on a dithering plugin, and I can use the Tracktion Master fader too! But a quick read through everything above and I'd have to say that platinumears is spot on with his knowledge of the subject. So lets throw out a bonus questions. For 10 points: Master Fade Out before or after the dithering plug-in?

And one final tangent as I saw it referenced slightly. In DAWs with a floating point mix engine (32 or 64), the master fader is your friend. You do not want it at unity if you are clipping the main output. Unlike analog, ther is virtually no danger of clipping the mix bus (especially with a 64 Bit mix engine) so you can leave channels faders where ever they end up. But you must lower the master signal coming out of the mix bus before it hits the file during a render or the output during playback to bring it back down bellow 0dB. I usually set it at -6 to start, lower it as I mix if I see a large number of peaks on the main meter(I don't trouble myself with a few here or there), and then when I"m ready to export, I'll choose the master fader option to set the master fader based on the max level of the edit. Easy.

I hope this helps.

Ben

[gizmo]

Thanks for this, haven't felt clear about dither in T since dither was explained so much in Ozone. Will experiment using ideas mentioned here.