WAV vs MP3 quality effect on Neural Mix separation

  • Device model macbook air M4 100% Neural
  • Version of operating system macOS 15.5
  • Version of djay 5.3.2
  • Hardware/controllers used DDJ-REV5

Does the quality of file have a positive/negative effect on separating elements on songs?

I am jumping between WAV and MP3 versions of songs and can’t really tell which has a better result. I am using BBNO$ - Two as a reference, as I noticed some artifacts while playing with it before. But the same results exist in both versions that I have from my pool, which offers both. I am also comparing it to max quality on tidal, with no real change. I get that with this level of mixing, in real time, the quality will be affected, but it almost seems 1:1 once separation is involved.

Does Neural Mix default to a specific format/quality, regardless of the original file? I guess maybe that is my actual question. I have been on the fence about switching to WAV recently, but it doesn’t seem like it’s necessary overall.

Hi @SpEd, sorry for the delayed reply. I wanted to confirm with engineering before answering. File format can have a small effect on separation quality depending on the compression, but it’s probably not very noticeable in most cases. Uncompressed/lossless (e.g. wav/flac) will in theory always be better than lossy (e.g. mp3). I hope that helps.