Grid offset in certain tracks in rekordbox

I have a big music collection. A big part of my collection was managed in Traktor 2 (beatgrids, cue points), and some tracks have been managed in Traktor 3. I notice that all tracks that was made with Traktor3, after i used Rekord Buddy, and import the collection in Rekordbox, have this “cue and grid shifting” problem. I know that when the first version of Traktor 3 was released, they have the same problem. What i mean to say is that on the first version of Traktor 3, all tracks that are in minor key, even it was with a proper beatgrid and cuepoints, when you play them, using sync function, the tracks was shifted on the right. (here is a example: https://www.youtube.com/watch?v=1JfIO_JK0Ys). Meanwhile, they resolve this problem, but seems that all tracks that are managed in Traktor3, then scanned with Rekordbuddy, and imported in Rekordbox, still have this problem. I say all this because i saw here, that you think that this problem comes from the mp3 file, but i think that this problem comes from Traktor.

I’ve moved your reply to its own thread. It is easier to keep track of issues if we keep the replies related to the original post.

Are the tracks displaying the problem you mention MP3 tracks or another format?

Mp3 format.All MP3’s that was from Traktor2 have good beatgrid and cuepoints into Rekordbox, after was converted with Rekord Buddy, but all mp3 that was from Traktor3 collection have bad beatgrids and cuepoints in Rekordbox after conversion with Rekordbuddy. My reply was like a observation, not a issue. I understand that alot of Rekord Buddy users have this issue. The think is that into first version of Traktor3, i have the same issue with all tracks that was in a minor scale. See the video example. After a week, Native Instruments release a update that fix this problem. Maybe this Rekordbuddy problem have some things in comun with their issue from he first version of Traktor3

If those are MP3s then this is a known issue. The current 2.1beta deals with a lot of those but some tracks may still fall thru the cracks.

Could you send me two or three of your track files (sendspace link via a DM) so I can make sure that I’ve got those covered and/or add a fix for them if not?

I cannot use the new 2.1beta, because all my music is on external drive (i replace the dvd-rom with a hdd)

Does this beta not work?

Thanks. It works. With crashes from time to time, but works.

OK. At what point does it crash?

Following.

I’ve had the same issue in rekordbuddy2, but contrary to the OP, I’ve seen it in both TP2 and TP3 analysed tracks:

I’m currently waiting for the 2.1 release to work with Damien to identify examples that still need fixing in rekordbuddy2.


Independently, In the last few months I’ve manually analysed my collection, and came up with algorithms to do the same.
The current iteration fixes this problem in my 100 files dataset with very few outliers (AKA false positives):

if ! (mp3guessenc sees Xing/INFO tag):
    case = "A"
    correction = 0ms

elif ! (eyeD3 sees correct LAME tag):
    case = "B"
    correction = 26ms
         
elif eyeD3 sees correct LAME tag:
    case = "D"
    correction = 0ms

This issue was also seen in MIXXX when a) reading Rekordbox collections and b) transfering cues from macOS MIXXX to windows MIXXX


I have some files to look at from another user. Could you send me one or two of your false positive so that I can take a look? You can DM me a link on the forum here.