Hi, I just updated to Engine Prime v1.5, and I wanted that the beatgrids of all my library was calculated using the new algorithm. With this aim, I selected all the tracks that I wanted, and selected “re-analyze tracks”. Is this the correct way to do it? (I want to keep the loops and hotcues). What would happen if, instead of “re-analyze”, I click to “analyze”? What’s the difference?
Select a few tracks -> left click and then:
-
Analyze: only tracks in the selection that are not analysed will be analysed. (the • in the “grid” column tells you if a track has been analysed or not)
-
Re-Analyze: all tracks in the selection will be analysed overwriting any previous analysis data.
2 Likes
To add extra
-
If you import tracks that already have bpm EP will use that to make its beatgrids so you may still end up with tiny drifts due to how individual software stores that bpm info. I believe they are stored as whole integers eg 120.4 will be 120
-
Use reanalyse to force engine prime to Analyse all tracks that already contain bpm that wasn created with EP ie EP will recognize the .4 and the grids will be correct
-
Analyse for new imported tracks that don’t have any bpm info will also take advantage of the new algorithm
Hope that makes sense
2 Likes
and now we are in the details 
so to even add extra to @mufasa:
The BPM is the same as the 'wideness" of the individual beatgrid idents. In other words: the grid contains the BPM. The only difference between BPM and grid is, a grid has a “begin point”, usually refered to as the beatgrid marker or beatmarker for short.
There are 2 locations where BPM info is stored:
-
The DJ’s own proprietary data, where the BPM is stored as a “real” or “floating point” value, a value with a deicmal point. Usually acompanied by a beatmarker position together making up (part of the ) the beatgrid.
-
The tags stored in the track’s audio file. There are different standardized methods of tagging an audio track, but they all contain the BPM as a whole number, an integer. Standardized audio tags can be read by any software. If a tag is available then this is what EP will use as the base for its own analysis. But it’s often a rough proxy, because it’s by definition a whole number.
Some software (Rekordbox and EP 1.5) do not store their analyzed BPM value in the track’s tags but only in it’s collection database. For software that stored BPM values in tags (Serato, Traktor, Mixed in Key) it’s a guess if they round a value or chop it off. e.g. 120.5 BPM becomes either 120 BPM or 121BPM. If a 120.8 BPM track is rounded to 120.BPM or 121 BPM makes all the difference for the drift of the grid.
How severe the drift of the grid becomes depends on the real BPM of the track and how far you are out in the track. The higher the BPM the worse the grid, the further out in the track, the more progressive the shift.
An extreme example that signifies this:
The true BPM of a track is 160.9 and that is rounded to 160 BPM in the tags.
The true grid has an ident every 0,372902423865755 seconds (60/160.9)
The grid that is based on the BPM tag has an ident every 0,375 seconds (60/160)
Not a big deal you would say. The “error” is the equivalent of a pitch set at -0,6% (rounded) .
In other words after a minute into the track, the drift of the grid is 0.9 ident, the grid has shifted “0.9” beats to the right (slower BPM=wider grid). The further you get into the track, the worse the drift gets, after 3 minutes the drift will be (0.9*3)=2.7 beats.
Now this is a bit of an extreme example to signify how wrong this can get. Every grid that drifts becomes useless for setting an “auto loop”. And where are auto loops used most?
At or near the end of a track, so where the drift is the worst.
The BPM analysis of EP has indeed improved considerable with version 1.5. But can be still thrown off by a wrong BPM value stored in the tags. I show that here during my AMA live stream (link starts at the correct time):
Because the improved BPM detection introduced with EP 1.5 I would rather like an option to turn off the use of the BPM tag and rely on EP’s own analysis strength.
2 Likes
This topic was automatically closed 24 hours after the last reply. New replies are no longer allowed.