Topic: Dithering + bouncing confusion

Viewing 8 posts - 1 through 8 (of 8 total)
  • #2531
    thcilnnahoj
    Participant

    Say I have a limiter plug-in that has a dither algorithm built in, and I want to make use of it directly inside Podium. Please help me see through this confusing stuff! 😉

    Let’s say I set this very last plug-in in the chain to output 16-bit (dithered). Podium will still render a 32-bit file (or even 64) from this, right? Does this simply leave the other bits unused, or what happens here?

    If I would then use the export file dialog to create a 16-bit file from the master bounce render, won’t this essentially do the bit conversion twice – the second time just by truncating…?

    Is there even a way to do this right currently?

    #20362
    Trancit
    Participant

    These are 2 independent things:

    If you use a plugin with dithering settings, you only tell the plugin by activating the dithering for i.e. 16 bit, how much low level noise the plugin should add to the audiooutput of the plugin…

    The render settings of Podium don´t know anything about what the plugin adds to the audio and renders the file as you told…you get a file with a bit resolution like you told in Podium preferences > engine > Mixer engine bit resolution, so 32 or 64 bit…and all of the bits are used like you haven´t had any plugin with dithering…only difference: you get a 32 or 64 bit audiofile, which contains beside the normal audiodata that low level noise added by the dithering algorythm.

    This means, adding dithering before bouncing doesn´t make any sense…
    Dithering should only be applied before finally export the master output to a file of 16bit…and before doing this, there is no bit conversation.

    Hope this is not too confusing 🙂

    Trancit

    #20373
    thcilnnahoj
    Participant

    Thanks – even though I’m not completely sure what’s summed up where and how, I assumed it wouldn’t make sense to dither before bouncing to 32-bit or higher.

    However, since there are quite a few plug-ins that offer dithering, and even plug-ins dedicated to it (think UV22HR for Cubase), I think there must be a way other hosts can utilize it (i.e., not processing the signal further after dithering).

    The question is whether it’s possible to correctly use dithering done by a plug-in at all in Podium currently, or in the future…? :-k

    #20380
    Trancit
    Participant

    Not sure, if I understand you correctly…

    Again, afaik, any plugin offering dithering options doesn´t do (and I think cannot do) any bit conversion on it´s own.

    The dither option in such plugins doesn´t do anything but adding a noise to the audiosignal to mask quantization errors, which arise out of the conversion from higher resolutions to lower.

    Definition here: http://en.wikipedia.org/wiki/Dither

    So dithering in Podium should be always applied, if you use “Export to soundfile” with a lower resolution than 32bit floating point.
    And if it is the last plugin in the master chain, Podium will not change the audio afterwards, so there is no need for a special or additional communication between the dither plugin and the host.
    Just add the plugin as the last instance in the master, export to soundfile as i.e. 16bit and your are done…and safe. Even Normalization applied afterwards will not change the dither result in a noticeable way, expect if it has to raise the final audio by 10db.
    But…it is overrated by using it in modern music style with a very low dynamic. It is more related to music with a very high dynamic, i.e. classic, where the quantization errors are really noticeable in the very quiet section.

    In a modern dance track, hiphop or rock it doesn´t make any difference, it is simple not noticeable. In theory it´s there, but you cannot here it.

    If your post is about, if it makes sense to offer dithering options in Podium, i.e. not only offering normalization but dithering as well, I don´t know how hard it is to programm such a feature.
    I guess, it is not that easy to make a “good” dithering option, on the otherside, even FL Studio has it as well as Reason/Record and nearly every other DAW/ Sequencer out there, so perhaps Fritz can tell us more on this subject.

    Hope I missed your post not completely 8)

    Trancit

    #20398
    thcilnnahoj
    Participant

    Hmm, as far as I understand, dithering doesn’t simply add noise before the conversion, but is used right during the bit conversion. The result, however, does sound like noise (i.e., random).

    This is the signal flow as I understand it:

    Audio from Podium’s engine running at 32-bit -> dither plug-in (outputs 16-bit) -> bounce rendering puts it into a 32-bit file.

    If I would then export this as a 16-bit file, Podium would do the bit conversion a second time! So that doesn’t work for me in this case…

    I could be completely wrong, of course! 🙂 I’ll send Frits an e-mail to check the forum again soon, maybe he can clear it up.

    #20399
    Trancit
    Participant

    Ok, now, after you made me unsecure, I did the test 8) …

    2 things are really for sure:

    1. Dithering is for sure overrated till it comes to recordings in natural rooms (especially recordings with noticeable ambient reverb of the room) with very high dynamic …

    2. dithering is really nothing else than adding a dither noise to the audio to mask quantization correlation. This correlation, which occurs by downconverting a high resolution signal into a lower, forces the waveforms more and more into a rectangle form. The dither noise smoothes the stairs, which appears by the converting process and let you recognize the downconverted waveforms better like there were sounding before…

    About the third point, does the plugin change the bit depth by itself, I was not sure, so I did following test:

    I played a softsynth (IL Sytrus) in Reaper and added a bit meter (Schwa Bitter). Bit meter showed me 32 bits were used.

    Now I loaded a limiter with dithering option to the signal.

    The limiter works internally with 64 bits, so the bit meter showed my 64 used bits after inserting the limiter.

    I activated the dither option (to 16 bit) and yes, you are right: the bit output dropped down to exactly 16bit.

    I am pretty sure, even if you convert the bit depth 10 or 20 times in a song, you won´t hear any difference in 99% of all music.

    But if you want to keep the full signal flow for your personal feeling, you should avoid using such dithering plugins before the final master.
    I´ve tried other plugins with dithering options and some doesn´t have even an option to turn it off, so the maximum bit output of those plugins (f*ck Waves, this shall be studi standard…) was only 24 bits…

    Again, I don´t think, that anybody on this world could really hear this, but to be safe, one should not use plugs like these in the “normal” signal flow…
    I wasn´t aware of this, so thx for forcing me to test 😀

    Trancit

    #20415
    Zynewave
    Keymaster

    @thcilnnahoj wrote:

    Say I have a limiter plug-in that has a dither algorithm built in, and I want to make use of it directly inside Podium. Please help me see through this confusing stuff! 😉

    Let’s say I set this very last plug-in in the chain to output 16-bit (dithered). Podium will still render a 32-bit file (or even 64) from this, right? Does this simply leave the other bits unused, or what happens here?

    The float values should ideally be value-quantized by the dithering plugin so that a later conversion to 16 bit integer values (by Podium) just removes/truncates unused bits. It’s important to not apply any processing or gain-scaling after the dithering plugin.

    If I would then use the export file dialog to create a 16-bit file from the master bounce render, won’t this essentially do the bit conversion twice – the second time just by truncating…?

    You can open the sound properties for the master bounce file, and change it from floating point to 16-bit integer. That way it is not needed to do the bit conversion when you later export the arrangement as a wav file. But it is ok to just leave the master bounce as a float file.

    Is there even a way to do this right currently?

    Once Podium supports dithering natively, it will be an option in the export dialog. It will probably also be a global option for the audio interface, so that dithering is automatically applied each time Podium outputs audio to the audio interface, ensuring that realtime monitoring also is dithered. Of course, if you are using a dither plugin as the final step in the master chain, you should then not enable the native Podium dithering.

    #20418
    thcilnnahoj
    Participant

    Thanks, Frits!

    About the usefulness of dithering – I don’t claim to actually hear a difference, at least certainly not with the equipment and room I use. I have a lot of 32-bit floating point WAVs around which I want to finally finish and probably archive to FLAC (which doesn’t support floating point anyway).

    I just need a way to make sure I get the dithering job done right before I could even start doing listening tests. 😉

Viewing 8 posts - 1 through 8 (of 8 total)
  • You must be logged in to reply to this topic.
© 2021 Zynewave