Question about track lengths

[Nisto]

Sometimes, the track lengths are printed on the booklets. Should we use those track lengths for our tracklists, or the track lengths of digitally extracted files?

[Secret Squirrel]

Printed lengths can be in error, and that error could prevent the album from being found in FreeDB lookup, so it's best to use the digitally measured ones.

[Nisto]

While I do agree with that, it can still be good to have the printed track lengths to go after, where it seems accurate. It sometimes happens, that people change track lengths back and forth.

Don't you agree?

[Secret Squirrel]

Well, staff gets to examine the track time edits submitted by anyone -- except for trusted editors. Have you been changing existing tracktimes to match the booklets?

The solution might be to have multiple track times for each track, except that I can see that causing confusion. Someday we'll have a better way of associating information with each individual track, so we can think about this idea then.

Also, note that having the actual DiscID trumps all.

[Nisto]

I only use the track lengths from booklets when I don't have the digital files available. I have seen, in albums history, that some people are petty about a single second flaw, though. It isn't wrong in any way to do so, per say. But let's just say that if there were printed times available, we should probably go after that. I know some tracklists are edited by people who have the digital files -- but it's not always that all digital files have the same length, either.

[Fearin]

One of the problem with track lengths is how they are measured. Depending on the program it may be the same length, but some programs round up over .500 in length to the nearest second, and others like musicbrainz round down. to anything below the next full second.

What that means is that there will most likely be an inherent flaw in whatever system is used. Depending on what people use to playback or tag their music with.

Even if track times were expanded to their illogical conclusion (IE ending 1:32.56524526745289463248362476210 or however long it goes) the programs would still do their rounding, people would still read the booklets which used their own programs with rounding.

Don't forget pressing differences too, that may or may not add a few seconds here and there. Thus it is most likely/best to give a +/- of about 2 seconds.

I guess the practice might be most useful with multiple synced parts. use a disc ID, the full length, a program that rounds up and down then the booklet. Add the info in the forums as a post.

[Blitz Lunar]

another problem with using track times from ripped music is that, at least with MP3 format, many encoders add a small amount of silence at the start of tracks.

[ilef]

Originally Posted by Blitz Lunar

another problem with using track times from ripped music is that, at least with MP3 format, many encoders add a small amount of silence at the start of tracks.

Or they might also cut 2-3 seconds at the end of the track. I'd go with the lenght reported from a .wav rip of the track.... Sometimes the informations reported on booklet/covers is wrong. If I remember correctly, VGMdb does allow a little inaccuracy in the tracklenghts, something around 2-3 seconds, when using it to encode the album in a digital form...

[Datschge]

Just wanted to drop the note here (didn't see any better thread and is not really worth its own one), that http://shop.tsutaya.co.jp/cd.html lists the track lengths of all the CDs it has in the shop with tracklists. Has been a good reliable source for me and is helpful when nobody got a CD yet.