Crossfade Pops and Clicks

I've posted about this problem before but never found a solution to it or know what the cause is?

I use Audio Snap to find the transients of my bass notes, split the clips at the audio snap markers, and then Quantize with the Xfade box selected.

The edited clip plays back without clicks and pops, but when I "bounce to clip" all the sudden I'm getting clicks and pops on some of the crossfades. I tried "bouncing to a track" with the fast bounce option disabled but there were still pops and clicks.

Anyone else have this problem and how did you fix it? This is really driving me crazy.

ORIGINAL: daveny5

Do you have Snap to Zero Crossings selected?

Does Snap to Zero Crossings matter if you're using Audio Snap. I've never noticed the Audio Snap markers snapping to a zero crossing nor have I seen the automatic crossfades snap to a zero crossing. Would this be the problem if I'm not hearing the clicks and pops until after I bounce down to a clip?

ORIGINAL: CJaysMusic

The ends of your clips need to be cut at the zero crossing. If you stop a signal abruptly in mid flow it will cause farts. This is digital audio not analog and the rules are different
Cj
Edit, if you abuse audio snap and drag the transients to much, this will cause pops also. audio snap is for minor timing screw ups

This is a problem that comes and goes for me. The first song I edited (Drum & Bass) went really smooth...not a single click.

The seconds song (the one I'm on now) is a completely different story. I edited 10 tracks of drums without any clicks but when I started editing the bass, my luck ran out. After locating the first 30secs of bass transients with audio snap, I split the track at the markers and quantized with automatic Xfades set to 5ms. Between every crossfade there was a click, and this was before bouncing down. I finally solved the problem by saving, closing Sonar, and reopening the file....all the clicks were gone on playback.

I did the same for the next 30 secs of bass track and this time there were no clicks between the Xfades upon playback. But when I bounced down to a clip, I had clicks.

I think this problem has more to do with settings in Cakewalk then how I'm using AudioSnap. In both songs I've edited the same way (without moving the markers to zero crossings) so something else must be going on here. Might it have something to do with buffer settings?

I'm wondering why he would split the clips in the first place. Why not just quantize the audio with autostretch enabled on the intact clip? Unless you are drastically stretching the audio to fit a significantly different tempo, I would expect this to work more seamlessly.

And once again, i'm getting pops and clicks. To answer your question, i'm editing multiple bass tracks from a single performance. Using audiosnap would cause nasty phasing issues.

Out of curiosity, I tried recording a quick bass track (actually low notes on a smooth E.P. patch) with some 16th notes in it, and quantizing it both ways - by autostretching with AudioSnap, and by splitting and quantizing with crossfades. After much experimenting, I got both to sound reasonably good. Here are some key things I learned:

1 AudioSnap often locates superfluous transient markers in the middle of sustained bass notes, especially with this particular patch which has some built-in chorusing and tremolo; I couldn't get the Sensitivity or Threshold controls to get rid of them without also throwing out good transient markers, so I had to go through the track and disable them.

2. I was sort of deliberately sloppy with my timing during recording to make the job more challenging. As a result, one of the transients was just over halfway to the next 16th beat, and moved the wrong way when I quantized, so I had to undo, and manually drag-stretch that transient a few ticks earlier to get it closer to the intended beat before quantizing.

3. Once I got the track prepped for quantizing, using AudioSnap was by far the easiest way, and sounded pretty good after bouncing down with the Radius Solo Bass algorithm, but it had one place where two notes at a 16th interval weren't handled well, and produced a sort of double-hit on the second note. I think with a little tweaking of that transient marker location before quantizing, I could probably fix it.

4. To make a long story short on splitting and quantizing, I found that quantizing within AudioSnap after splitting did not work well, because the crossfades created with the spits did not follow the transients when I quantized. Not surprisingly, this produced lots of pops and clicks, and they only got worse when bounced down to a single clip. What I ended up doing was using AudioSnap only to split the clips at transients without crossfades enabled. Then I disabled AudioSnap on the track, and used the regular Process/Quantize command to quantize Audio Clip Start Times . I started with crossfades at 5ms, but got some bad transitions. 10ms was still a little iffy, and I ended up using 15ms. All in all, this method yielded the best result with the fewest artifacts, and wasn't a lot of work once I figured out not to quantize within AudioSnap (I think not having the crossfades follow the transients on quantize qualifies as a bug).

5. I should mention a helpful trick I found along the way to figuring out what was going on with crossfades. This was to Ctrl-click select every other clip in the split track after quantizing, and Shift-Drag them down to the next track so that I could easily see how the overlaps and crossfades were working from one clip to the next.

I don't know if any of this helps, but I thought I'd share anyway.

P.S. Actually the fastest and best result was obtained by quantizing the MIDI I recorded along with the audio, and re-recording it. I sure am glad I don’t have to deal with multitrack MIDI "bleed".


Edited to add an epilogue: I went back to an intermediate save of the test project I was working with, and found to my surprise and displeasure that it is now exhibiting the dreaded "disappearing transient markers" problem I have reported elsewhere. Most of the transient markers in the first half of the track are gone, and I can't get them back. Moving the threshold slider away from All, actually makes some of them come back as others go away, but many of those that come back are disabled, and cannot be selected to re-enable them. Re-detecting transients in the clip properties doesn't help. This is my number one most troublesome bug with AudioSnap, because the original state seems to be unrecoverable.