Page 1 of 1
timestretch snafus
Posted: Mon Dec 11, 2006 3:44 pm
by Electric Lotus
Hello all,
A) desire : Change a whole song (24 tracks or so over a 4 minutes period) to 128.5 BPM from 127 BPM.
B) method : Merged all the bites (new clean files), changed the BPM in conductor track, select all, Hit the 'adjust all soundbites to sequence'
C) results : some files yield bezerk renditions, where the audio has NOTHING to do with the intended compression, 2/3 through, they will speed up dramatically, and the end points won't match their 'sister' files.....
The warbled sonic effect is appalling, phasing issues, you name it...
D) workarounds : Reiterate the merging, copy project to another drive, the usual common sense workarounds, no success.... Tried 'trimming' the merged files by 0 dB to clean them out, no success....At his point, I'm using other programs to alter my files.....Peak does not improve things, PT is just a mess...but their endpoints do match up....
E) : Conclusions : There is a bug either somewhere in the merging process, or within the time expansion procedure......i've had similar problems with the Pure DSP pitching thingy....
Why do some files do fine, while others go bonkers ???? It's a math thing !!! or did I chose the 'evil' tempi for that project ?????
Cheers !
Posted: Mon Dec 11, 2006 5:14 pm
by oldguitars
I have never seen any software adjust tempo across multiple tracks (especially 24) with out lots of issues. I would complete the project as is, then after you have bounced it to a stereo pair, then adjust. serato pitch and time is the best (me thinks) and live is really damn good. with that many tracks, there is way too much that can go wrong with timing. another reason i miss good old varispeed (aside from changing the pitch : )
Stereo Bounce
Posted: Mon Dec 11, 2006 9:28 pm
by Electric Lotus
That seems the way to go, or into my MX-2424, then varispeed the bastard....
Hoewver, it still does not explain why files get 'corrupted' by the timestrecth, they all ought to share common endpoints after the action, if they shared it before the operation....
...Thanks for taking the time to reply, I love the/my Gretsch Tennessee Rose too !!!!

Posted: Mon Dec 11, 2006 10:16 pm
by HCMarkus
In my experience, timestretching individual tracks prior to mixing yielded much better results when using DP's native tools. In particular, when I have sped up vocals to fit MIDI arrangements, the individual approach gave me very useable results, wheras the "shift-en-masse" approach was aweful.
I did not experience the endpoints issue, even working with about 20 tracks of vocals. Then again, the segments I was working with were all quite brief.
Posted: Mon Dec 11, 2006 10:29 pm
by supersonic
It also depends on the nature of the tracks. Time shifting 12 tracks of drums is not the same as time shifting two tracks of completely different nature. Although I have done both (in Digital Performer) with satisfying results...
Re: timestretch snafus
Posted: Tue Dec 12, 2006 9:07 am
by muzishun
Electric Lotus wrote: some files yield bezerk renditions, where the audio has NOTHING to do with the intended compression, 2/3 through, they will speed up dramatically, and the end points won't match their 'sister' files.....
The warbled sonic effect is appalling, phasing issues, you name it...
A couple of things I've tried that sometimes works:
1) If a soundbite was recorded too loud (very dense waveform) try bouncing the track at a lower volume.
2) Try cutting the soundbite where the tempo takes off. Usually merging files helps, but sometimes DP seems to be able to handle smaller files better.
It's never awesome, though.
Hope this helps.
Posted: Tue Dec 12, 2006 11:42 am
by magicd
I've seen this happen before.
There is something specific to the audio file that is confusing DP. If you spend the time, you'd be able to find just one or two beats in a given audio file that are causing this problem. For example, if the audio file is properly stretched up to one minute and then it goes nuts, there is probably something at the one minute point in that audio file that is causing the screwup. Don't bother lookinmgh for the screwup. You won't find anything obvious.
Here's how to work around it. After the first time you do the global time change and notice the problem, look at the Sequence Editor and examine the tracks. You'll notice that some tracks changed all the way through with no problems. Make a note of which tracks had the problem and where. Undo the time change.
Make cuts in the audio on the tracks that had the problem, a bar or so before the problem occured (so you can do crossfades later if you need to smooth out an edit). So if you had a three minute audio track that time stretched with no problem up to one minute, make the cut just before the one minute mark.
Now merge the sections that had the stretch problems. In our three minute example, you'd be selecting and merging the last two minutes of the audio.
Select all.
Copy Sequence Tempo to Soundbites.
Make your tempo change
Adjust SoundBites to Sequence tempo.
In most cases you will now get a successful time stretch all the way through. In rare cases there may be additional problems further down the audio track. In those cases, back up a step and do more isolation and merging. What I've found is that the new audio created by the merge rarely has any problem with the time stretching.
Magic Dave