Topic: MD5 for FLACs

[blu666z]

Is it pointless to create an MD5 for a FLAC fileset?  Gives you a way to verify you have an original set right?

-Kevin

[JNT]

You don't really need a MD5 for FLAC files they have their own internal verification system.  There is the FLAC fingerprint for comparing your copy to a known copy. 

The only reason I can think of for using MD5 with FLAC is to create a "packing list" for the files.  This way if you download a show and one of the files is missing running the MD5 will notify you that you don't have all the shows files. 

The problem with using MD5 for FLAC files is that changing the text portion of a FLAC will not alter the internal check but the MD5 will fail even though the music has not been altered.

Joe

[John Kelly]

Yeah just put a fingerprint in your info file.

[blu666z]

Good point.  The problem with fingerprints is that there isn't a good way to compare them.  I wrote a program to do it for me but have had problems getting it to compile to distribute.

-Kevin

[JNT]

Yeah I remember reading your posts about the FLAC fingerprint compare program.  Sounds like a good idea to me. It's not much fun comparing long strings of text.  Perhaps someone here can help you get it going.

Joe

[macdaddy]

you only need to check the first few characters (or last few).

as for md5s - i include them because i can run them after i burned a disc, and check that the files ON the disc match the md5 - one click burning verification for me...

ditto for downloads/uploads...

just make sure you tag/apply replaygain before making the md5 because all that non-essential info changes the header, which also changes the md5 signatures...

[leegeddy]

lately, i've been seeing more md5 with flac downloads. i hear many people use SummerMD5 to generate the checksum and works fine with mkwACT for verification.

marc

[macdaddy]

lately, i've been seeing more md5 with flac downloads. i hear many people use SummerMD5 to generate the checksum and works fine with mkwACT for verification.

marc

i use the command line of md5summer (md5summer.org), copy the text in the command window, and paste it in the md5. it will make md5s of anything...

the gui puts that crap with the web address, but it isn't commented out properly, so you get the error message rather than the all OK message, and that drives me nuts.

but yes, mkwACT will verify md5s made with any program out there (and there are others)...

[twatts (pants are so over-rated...)]

I'm not sure about renaming FLACs creating errors with the related MD5s...

However, this whole thing issue can be eliminated with the creation of WAV MD5s to go along with SHN/FLAC sets.  Once the conversion to wav is done, check the WAV MD5.  Who cares what the original format was or checking it etc., if the final WAVs check out, then its all good right???

I've thought about this and found the no MD5 for FLAC a minor issue (no way to tell if your FLACs are same as original FLAC siles except by physically reading each FFP, or using a home-made GUI.  Sure, the FLACs may be error-proof, but I can EAC>FLAC an audio and make it error-free - doesn't make it the same set as the original.  I think MD5s give a quick and easy way to differentiate between sources.

However, if you make WAV MD5s, then we're all good...  I've started including them on all my archives...

Terry Watts

[Craig T]

no MD5's for FLAC - its pointless.

you can do MD5's for the .wav's.  that way you can verify after you decode (for you anal freaks who don't trust the internal error checking in the FLAC process).

[John Kelly]

I'm not sure about renaming FLACs creating errors with the related MD5s...

However, this whole thing issue can be eliminated with the creation of WAV MD5s to go along with SHN/FLAC sets.  Once the conversion to wav is done, check the WAV MD5.  Who cares what the original format was or checking it etc., if the final WAVs check out, then its all good right???

I've thought about this and found the no MD5 for FLAC a minor issue (no way to tell if your FLACs are same as original FLAC siles except by physically reading each FFP, or using a home-made GUI.  Sure, the FLACs may be error-proof, but I can EAC>FLAC an audio and make it error-free - doesn't make it the same set as the original.  I think MD5s give a quick and easy way to differentiate between sources.

However, if you make WAV MD5s, then we're all good...  I've started including them on all my archives...

Terry Watts

Problem is my tapes never see a "wav" stage.  AIFF is the standard on the mac, and it's what my software saves to by default.  Then I convert right to FLAC...

[blu666z]

Its not so much trusting that it decoded to WAV correctly.  Its that the FLACs you decoded are the original FLACs.  

[ChrisV]

For sharing and distrubution I always just include a .ffp fingerprint file...no md5's.

But for my own purposes on my recordings I'll keep a md5 for the .Wavs just because I like to do a check before I burn an audio CD that the .wavs are indeed correct.  And I'll keep an md5 for all the .flac files because I myself like to have an easy clickable verification when moving files from one comp to another or to CD/DVD.   Yeah, I am pretty anal ;p

[Cooker]

MD5s for compressed FLAC files are bad. that's a no-no.

you can change encoding levels, id3 tags, all sorts of things that will not effect the source wav files and will cause the flac files to fail MD5 verification.

use flac fingerprints, and if you want to check a packing list, you can just open the text file and manually count the result files.

as for fingerprint compare, it sucks there is no easy way, but I don't really see the point... why not just use the verify feature?

[dnsacks]

I find that a md5 for a flacset that's been bittorrented or ftp downloaded is a GREAT tool as it's the only quick one click confirmation that everything came through ok that I'm aware of.  Of course, it's no substitute for providing a flac fingerprint list in the text file so the integrity of the flac set can be confirmed. 

On the other hand, running a verify of the flacs is not as useful as, with a 30 file flacset, it's a pain to manually scroll through the dos window output to be sure each file verified ok -- imho easier to include an md5 of the flacset for a quick and dirty way to confirm that all's ok -- per my thoughts above, if the md5s fail, the flac fingerprints provide a backup method of determining the integrity of the flacset.

[blu666z]

why not just use the verify feature?

ie.  I get original FLACs from a taper.  I decode one track to wav, do some kind of editing and reFLAC it. I send you changed track.  You do your verify and it checks out ok, but its not the same as what the original taper put out. 

I just really need to figure out what my problem is with ther verifier I wrote. 

-Kevin

[greenone]

Why not use shntool md5's? They just take the md5 of the WAVE *data* (not the .wav files) contained within your shn/flac/ape/whatever.

http://forums.etree.org/viewtopic.php?t=1948

[dnsacks]

but what does that solve -- the thread you posted indicates this is pretty much the same as a flac fingerprint.

[greenone]

Aren't they easier to verify than ffp's? Hmm. You might be right.

[macdaddy]

ffp and shntool md5 are the same things...

that is why if you eyeball it to compare - the ffp included from the taper, and the shntool md5 of the wav you created. if the match, you are in business...

all i was saying that as a quick, one-touch way to verify if a disc has been burned 100% accurately, md5s are handy. ditto for a one-touch way to verify is a download fineished successfully.

but in a seed, i always include the shntool len output, the WAV shntool md5s, and the ffp. the latter two should match exactly.

but before i even read a txt file, i would double-click on an md5 in the top directory of a bit torrent download in order to verify that i received all information from a download...

but ffp/wav shntool md5 is the way to go to verify data integrity with finality.the uses are not mutually exclusive.