Dithering Question: Trimming song files, do I need to dither?

W

WaxGolem

Logician
I am using Logic Pro X to edit music files downloaded from Bandcamp and burned from CD. Basically, if an audio file has a long lead in or tail that is just silence I open it in Logic and trim up the file to make the album flow smoother.

I download from Bandcamp in ALAC and I also burn CD's to this format. I'm then opening a file in Logic, trimming the file and presently I've been just exporting them (not bouncing). So Export 1 track as audio file. It is my understanding this method does not add dithering and doesn't even offer the option as bouncing does.

I've read that you shouldn't apply dither to files that have already been dithered. Being that these songs have likely already had dithering applied at some point before purchase, am I correct in not dithering these files when I export? Or since in many cases I'm exporting them in 16 bit 44.1/48 kHz would dithering be recommended? I am not converting at any point into a lossy format, which is what I've read dithering is supposed to be of benefit for.

Thanks in advance for any clarification.
 
* Exporting them in WAV. I do reconvert them to ALAC once I open them again in iTunes. Not sure that it matters, but just to add the final steps of my process just in case.
 
Bandcamp sound like mixed and mastered finished songs. I would not dither these - all the more so if they are already 16 bit, as would be the case with your CD files.

kind regards

Mark
 
Markdvc said:
Bandcamp sound like mixed and mastered finished songs. I would not dither these - all the more so if they are already 16 bit, as would be the case with your CD files.

kind regards

Mark
Click to expand...

Thank you for your response.

I want to clarify the answer a little further if possible. As you say many of the files are already 16 bit. However, when I open them in Logic they are opened as 32 bit floating point correct? So I'm wondering if this increasing the bit depth/upsampling of the sample rate then reducing the bit depth at export introduces the quantization errors that dithering is supposed to combat. If I opened it as a 16 bit file and it remained a 16 bit file at export I can see that wouldn't be a thing that would require dithering.
 
WaxGolem said:
However, when I open them in Logic they are opened as 32 bit floating point correct?
Click to expand...

I don't understand. Exactly what format and audio resolution do these files actually have? Or are you referring to Logic's Audio Engine?
 
It varies, when you download from Bandcamp ALAC they retain the original resolution they were uploaded in as WAV I suppose. So I've had 16- 24 bit and 44.1-48kHZ. Some albums may not have all songs in the same exact resolution, just depends on how the artist uploaded it. The only requirement is lossless, not uniform bit depth and sample rate.
 
Lossless sounds like some some sort of compressed format, not WAV? Perhaps the most pragmatic approach is to try with and without dither. If you can hear a difference then you will be able to make a clear choice.
 
Markdvc said:
Lossless sounds like some some sort of compressed format, not WAV? Perhaps the most pragmatic approach is to try with and without dither. If you can hear a difference then you will be able to make a clear choice.
Click to expand...
Here is Bandcamp's requirement: Your audio files must be at least 16-bit, 44.1 kHz lossless files (WAV, AIFF, or FLAC).

So they upload it in one of those formats at whatever bit depth and sample rate. When you download the files from Bandcamp you choose your download format. In the case of lossless formats it merely gives you a clean conversion from say WAV to ALAC. Same bit depth, same sample rate as the artist originally uploaded it in.
 
Sorry above I missed when you asked was I referring to the audio engine. Yes I was. Does that only apply to when you record into the DAW? I thought any files you put into a session were that resolution until you export or bounce them.
 
OK, we need to differentiate here between the audio file format and the software that is playing back the audio file, those are two separate things. If you import a 24 bit WAV into logic, you still have a 24 bit WAV. If you import a 16 bit WAV, it stays 16 bit. The audio engine at 32 bit floating point means that any edits made are more precise, levels can be changed without risk of overloading and so on. Might be worthreading up on this, there is plenty of info around, for example: Q. Should I bounce to 32-bit floating point or 24-bit fixed point audio files?
 
You must log in or register to reply here.
Back
Top