Adding dither before importing to Logic

[Antiphones]

I need to dither from 24 bit 88.2 Khz down to 16 bit 44.1 Khz. I would like to use iZotope's MBIT dither as, for the sort of music on this album, I like it better than any of the POW-r dithers. However the only way I have of applying this dither is in an offline stand alone application (RX3).

However, if I dither to 16-44.1 and then bring the files into Logic (actually I'm using Waveburner for this but it applies equally to Logic) and top and tail or adjust the level of a track (when laying out the CD), presumably these files will be re-rendered when I bounce (to CD or PPD image). Processing and re-rendering the 16-44.1 files will result in lower quality (due to alaising) than if I'd done the top and tail and levels with the original 24-88.2 files. Am I correct?

One option that occurred to me was to add the MBIT dither but keep the files at 24-88.2 and then bring the into Waveburner (Logic). Then I can top and tail and adjust levels at 24 bit which would be much better in terms of processing quality than 16 bit.

My thought is that since the dither noise is in the 24 bit files, when I bounce to 16-44.1 the dither would still work its magic. But am I correct in this thinking?

I'm presuming this because limiters from Waves and PSP allow you to add dither in the plugin. So in this case it seems that the dither noise is added to the signal before bouncing (when truncation occurs).

On the other hand, these same plugins allow you to select a bit rate. This I've always found confusing, because presumably they don't do the truncation themselves, Logic does it when you select the bit rate in the bounce dialog. Or do they do the truncation and somehow magically tell Logic not to try and do it again? I've always found this last part a mystery.

Thanks for any help someone more knowledgeable than me can offer.

 

[bayswater]

You only want to dither once, and that should be the last thing that happens to the audio. Is there some reason you cannot apply the dithering process as the last stage in Waveburner? I'm assuming you have MBIT as a component of Ozone or some other package, so why not just put that in the chain?

Also, if I understand you, you also ask about applying dither to a file at 88.2/24 but leaving it at 88.2/24. As far as I know that will do absolutely nothing to the file. The point of dithering is to reduce the bit depth while minimizing the resulting quantizing noise, so if you don't reduce bit depth, the output will be the same as the input.

 

[Antiphones]

Thanks for your reply.

You only want to dither once, and that should be the last thing that happens to the audio. Is there some reason you cannot apply the dithering process as the last stage in Waveburner? I'm assuming you have MBIT as a component of Ozone or some other package, so why not just put that in the chain?

I don't have Ozone I have RX3 and that does not work as a plugin in Logic (though I think the next version will). But RX3 does have MBIT dither. So I was thinking of adding the dither using RX3 but not truncating to 16 bit there.

I don't want to process 16 bit files in Waveburner for obvious reasons, so I was thinking of leaving them as 24 bit but with added dither. The idea being that the dither would do nothing until the bounce to 16 bit from Waveburner

Also, if I understand you, you also ask about applying dither to a file at 88.2/24 but leaving it at 88.2/24. As far as I know that will do absolutely nothing to the file. The point of dithering is to reduce the bit depth while minimizing the resulting quantizing noise, so if you don't reduce bit depth, the output will be the same as the input.

Yes I realise that. But my thought was to put the dither in at an earlier stage (because I can't do it in Waverburner) set levels in Waveburner and then bounce to 16 bit. I know that some engineers don't use dither if the audio already has quite a bit of noise in it (analog hiss etc...) as this works as dither. So its the same principle.

 

[bayswater]

Thanks for your reply.

You only want to dither once, and that should be the last thing that happens to the audio. Is there some reason you cannot apply the dithering process as the last stage in Waveburner? I'm assuming you have MBIT as a component of Ozone or some other package, so why not just put that in the chain?

I don't have Ozone I have RX3 and that does not work as a plugin in Logic (though I think the next version will). But RX3 does have MBIT dither. So I was thinking of adding the dither using RX3 but not truncating to 16 bit there.

I don't want to process 16 bit files in Waveburner for obvious reasons, so I was thinking of leaving them as 24 bit but with added dither. The idea being that the dither would do nothing until the bounce to 16 bit from Waveburner

Also, if I understand you, you also ask about applying dither to a file at 88.2/24 but leaving it at 88.2/24. As far as I know that will do absolutely nothing to the file. The point of dithering is to reduce the bit depth while minimizing the resulting quantizing noise, so if you don't reduce bit depth, the output will be the same as the input.

Yes I realise that. But my thought was to put the dither in at an earlier stage (because I can't do it in Waverburner) set levels in Waveburner and then bounce to 16 bit. I know that some engineers don't use dither if the audio already has quite a bit of noise in it (analog hiss etc...) as this works as dither. So its the same principle.

Putting dither at an earlier stage makes no sense. As I mentioned, dither reduces the bit depth. Usually, it is used to reduce bit depth from 24 bits, which most people use for mixing, to 16 bits for burning CDs. If you apply a dither process to a 24 bit file and leave it at 24 bit output, it will do nothing -- not then or at any later stage.

How about this:

- process the file the way you want at 24 bit in Logic.
- bounce the file to disk from Logic at 24 bit
- load the result into RX3, apply dither to go from 24 to 16 bit, and output the 16 bit result.
- load the 16 bit output into Waveburner to make the CD.

 

[Antiphones]

Putting dither at an earlier stage makes no sense. As I mentioned, dither reduces the bit depth. Usually, it is used to reduce bit depth from 24 bits, which most people use for mixing, to 16 bits for burning CDs. If you apply a dither process to a 24 bit file and leave it at 24 bit output, it will do nothing -- not then or at any later stage.

I appreciate your reply but you are incorrect. Dither does not affect the bit depth. Dither is a type of noise added to the signal which covers the quantisation distortion caused by reducing bit depth. Reducing the bit depth (truncation) is a separate step. You can do either one without the other.

How about this:

- process the file the way you want at 24 bit in Logic.
- bounce the file to disk from Logic at 24 bit
- load the result into RX3, apply dither to go from 24 to 16 bit, and output the 16 bit result.
- load the 16 bit output into Waveburner to make the CD.

I don't think working on 16 bit files would be a good idea. I'll need to use Waveburner to top and tail the tracks as well as adjust the volumes, possibly even apply some EQ. This would mean processing 16 bit files, which is never a good thing.

 

[bayswater]

Putting dither at an earlier stage makes no sense. As I mentioned, dither reduces the bit depth. Usually, it is used to reduce bit depth from 24 bits, which most people use for mixing, to 16 bits for burning CDs. If you apply a dither process to a 24 bit file and leave it at 24 bit output, it will do nothing -- not then or at any later stage.

I appreciate your reply but you are incorrect. Dither does not affect the bit depth. Dither is a type of noise added to the signal which covers the quantisation distortion caused by reducing bit depth. Reducing the bit depth (truncation) is a separate step. You can do either one without the other.

How about this:

- process the file the way you want at 24 bit in Logic.
- bounce the file to disk from Logic at 24 bit
- load the result into RX3, apply dither to go from 24 to 16 bit, and output the 16 bit result.
- load the 16 bit output into Waveburner to make the CD.

I don't think working on 16 bit files would be a good idea. I'll need to use Waveburner to top and tail the tracks as well as adjust the volumes, possibly even apply some EQ. This would mean processing 16 bit files, which is never a good thing.

Dithering does add noise, but it does so usefully as part of the process to reduce bit depth. Adding noise to a 24 bit file and outputting as a 24 bit file is pointless. The dither process is used to mask quantizing noise that occurs when you truncate a file, usually from 24 to 16 bit. It has the effect of shifting the noise that occurs from truncation from easily audible spectra to those that are not as audible. Noise is normally applied to the 16th bit in such processes. I guess if you did it to a 24 bit file and output a 24 bit file, noise might be applied to the 24th bit, but again, that would be pointless if you later truncate to 16 bits. Yes, you can truncate without applying the noise if you want.

As for working in 16 bit, you wouldn't be doing it. If you want to top and tail in Waveburner, just use that instead of Logic in Step 1 above. The only thing that happens in 16 bit is the burning.

 

[Antiphones]

Dithering does add noise, but it does so usefully as part of the process to reduce bit depth. Adding noise to a 24 bit file and outputting as a 24 bit file is pointless. The dither process is used to mask quantizing noise that occurs when you truncate a file

Yes that's what I just said above 🙂

I guess if you did it to a 24 bit file and output a 24 bit file, noise might be applied to the 24th bit, but again, that would be pointless if you later truncate to 16 bits. Yes, you can truncate without applying the noise if you want.

I'm pretty sure (not 100%) that the noise contained in the 24 bit file would still work as dither when truncated to 16 bit later.

As for working in 16 bit, you wouldn't be doing it. If you want to top and tail in Waveburner, just use that instead of Logic in Step 1 above. The only thing that happens in 16 bit is the burning.

Not if I bring in 16 bit files.

 

[bayswater]

I'm pretty sure (not 100%) that the noise contained in the 24 bit file would still work as dither when truncated to 16 bit later.

No, it won't. Think through the signal path.

As for working in 16 bit, you wouldn't be doing it. If you want to top and tail in Waveburner, just use that instead of Logic in Step 1 above. The only thing that happens in 16 bit is the burning.

Not if I bring in 16 bit files.[/QUOTE]
I'm not suggesting that you bring 16 bit files into anything that processes the files. Bring the 24 bit files into whatever you want to process them. Export 24 bit results. Bring that into MBIT. Produced dithered 16 bit files. Export. Bring that second export into whatever you want to use to burn the CD.

As I said before, you have to dither after all other processing. You're not going to get some sort of magical memory imprint of the MBIT processing that carries through to truncation that you do later.

 

[Antiphones]

I'm not suggesting that you bring 16 bit files into anything that processes the files. Bring the 24 bit files into whatever you want to process them. Export 24 bit results. Bring that into MBIT. Produced dithered 16 bit files. Export. Bring that second export into whatever you want to use to burn the CD.

Thanks for the suggestion. That would work as long as waveburner doesn't process the files when it makes a PPD image (assuming as you say that I've already done the top and tail and levels before bringing it in).

As I said before, you have to dither after all other processing. You're not going to get some sort of magical memory imprint of the MBIT processing that carries through to truncation that you do later.

You may be right. Unfortunately there are a lot of people who disagree with you on this point. My problem is that this is where my knowledge ends. I've read strong arguments by respected industry people who say that dither does carry through if added earlier. And I've also read what you're saying elsewhere too. But I don't feel I have the knowledge to know which is right here 🙂

However if you are correct that Waveburner will not re-process 16 bit files when creating a PPD or burning a CD (assuming I don't change any levels or anything) this would be a good way to deal with the problem. Thanks for your help.

 

[mt100uk]

Hi my ears tell me to do src and dither (in that order) after all sonic manipulation. From what I can hear, waveburner doesn't do anything to the audio if you feed it 16/44.1 and make no changes. Hope this helps


Sent from my iPhone using Tapatalk