Performance problems

[bronze13]

Since 2 or 3 versions I experience unusually long calculation times by D-O-M, especially when analysing audio. A 100 minutes film (.mov) with stereo audio took almost 8 hours to analyse audio. During this time my 16 GB RAM was used between 30 and 60 % and my Intel 4-core CPU with 3.4 GHz had a workload between 8 and 20 % (OS = Win 10 Pro 64-bit).
Does any one have a clue what I could do (except buying a new computer)?

[carl]

2.15.x versions or 2.14.x versions?

[bronze13]

I have 2.14.38

Cheers

Ulrich

[carl]

And you're finding performance worse there than with a previous version? Do you know what you were running before?

[bronze13]

Sorry, I don't remember.
Right now I lifted the peak level and added a C channel. Re-calculation ran for 1 hour.

[Carsten]

I learned that these calculations can take a different amount of time depending on how the audio is formatted/organized. I confirm that sometimes audio analysis can be painfully long, compared to J2K compression, especially on modern multicore CPUs. Sometimes it appears as if audio analysis runs longer than J2K compression.

That said, I have not experienced what you describe. Are you sure you haven't switched between 'simple' and 'complex' audio analysis (R128/LUFS)? The latter takes much longer.

I have a couple of older versions on my machine, I could try to run a few comparisons.

Level change should not cause a recalculation. Adding a channel will. Are you using separate files for multichannel DCPs, or interleaved (single file multichannel)?

- Carsten

[bronze13]

I updated to 2.14.55, and - behold - all performance problems have vanished.
So this thread may be labeled as "solved".
Cheers
Ulrich