Cross-pressing verification (range of detection)

[Intothisworld]

Hey guys, so I've been doing some experimenting lately, and I'm wondering, does anyone know what the limit is of how much a disc's data can be offset and have the disc still be recognized as an alternate pressing of another disc? I just manually dumped a disc at about a +10,000 sample offset and AccurateRip (unsurprisingly) didn't recognize it. I suppose I could just keep re-ripping at different offsets to test, but I wanted to go straight to the source. And I have another related question I wanted to ask as well...

The other question is, what is the method that AccurateRip is using here to be able to detect alternate pressings in the first place? I know at least that it's related to the CRCv2, but I'm not clear on what that CRC is actually representing. I've actually had some strange cases where a similar pressing was recognized despite the disc being fairly significantly different. There was one disc that in an older pressing had a big band of high frequency white noise cutting across the entire duration of the wavelength, but then in a later reissue had that white noise edited out. Obviously the values of all samples across the entire disc were slightly different as a result. And yet AccurateRip still recognized that it was essentially the same disc. Pretty danged impressive. I'd love to know what is happening to accomplish that.

Anyway, thanks for the help. Cheers,
- David C.

[Spoon]

It is around 1470 samples I believe, however the drives offset can remove some of this range, it is based on rolling matching with an exclusion zone at the start and end of the disc which allows other drives with different offset to work with accuraterip, and the cross pressing detection to work also.

[Intothisworld]

Good to know. Thanks!