PDA

View Full Version : Whole file MD5s (afps)


Erich
2005-01-17, 03:30 AM
I just saw this quote from RD, and I was wondering if this is still the case?

No problem...we've decided to hold off on requiring shntool md5 stuff for the time being, as it's really just too much of a pain to require everyone to learn this command line app. Sorry for bugging you about it, we're still trying to iron out all the rules.

At some point in the future when there is a decent graphical frontend for shntool, we may bring this rule back, but for now, wholefile md5 for shn sets is fine. For FLAC, ffp is still required.

If so, is it possible to just make referance to a batch file someone can use for shntool? for example, if someone using winXP puts whatever.bat into their flac folder, double clicks it, and the batch file has this:

----

@<hidden> off
shntool md5 *.flac > afp.txt
shntool md5 -c *.flac >> afp.txt

----

shouldnt it work without the person having to think too much, making it easier for intergration of this checksum into the seeds? and it can be changed to work with shns or whatever else.

(PS - the -c switch allows for the composite MD5, creating an MD5 for the entire set of files. This should also be used, in case the show ever needs to be reseeded retracked, or with simple SBE corrections. If done correctly, the single file checksums change, but the composite stays the same, further prooving that lossless is lossless)

I mean, its the absense of audio fingerprints (afps) that causes people to have hoplessly ridiculous "disclamers" such as "Do neither re-flac or re-ape", "do not convert to other lossless formats", and "do not convert to flac!" on their txt files. if shn, flac, ape, pac, wav, aiff, ofr, and the countless other usuable lossless codecs have the same AFP, then they should all be treated the same way. But you guys know and understand this.

Until the frontend is ready, shouldnt you guys still require this, but have step by step instructions with help so as to make it easy for even the most technicaly deficiant seeders?

ssamadhi97
2005-01-17, 06:40 AM
(PS - the -c switch allows for the composite MD5, creating an MD5 for the entire set of files. This should also be used, in case the show ever needs to be reseeded retracked, or with simple SBE corrections. If done correctly, the single file checksums change, but the composite stays the same, further prooving that lossless is lossless)
Only if the total number of samples remains the same. But if you correct SBEs, I'd say chances are that samples will be cut from the first or the last track (or either of those might be padded with zero'd samples) to ensure sector alignment of the first/last track.

RainDawg
2005-01-17, 08:29 AM
Yes, I would like to require it, but when I tried to do this, there was an uprising of sorts from people who were unable to get shntool setup on their systems. I was forced to back down and volunteered to create a frontend so that people would no longer have an excuse. Needless to say, my time that I have been able to devote to finishing this program has been extremely limited over the past month, so very little progress has been made. Nonetheless, as soon as it's out, audio fingerprints will be required on this site again.

As for -C, if people pad the start or end of the show when fixing SBEs, this command will not work. I personally like to pad shows so that EVERY track on the set is SB aligned, which will casuse the -C option to longer work. I just think it's a poor fixall, and will work in few enough cases that using it an expecting it to be valid would be a headache.