When changing clip speed via properties, there is audio glitching where the clips start and end, which shows up in the final export

I tried using proxy files to improve the situation, but it made no difference. The glitches are inconsistent when I play them back in the timeline, yet they make their way into the final export. Sometimes it’s hardly noticeable. Sometimes it’s a quick screech. And then when I hit export, they end up screeching at every point where the clip was split and sped up. The process I was using was to just split the clip twice, scoot whatever came after down if I’m slowing that bit down, change the speed, then if needed scoot back whatever came after. It’s making me really sad cuz I spent a while learning this app and I like it a lot. But this happened last time, and it was a pain in the a** to remove the audio glitches in Audacity, adding gaps in sound where the glitches were. I dont want to go through that again. Help would be appreciated but I doubt Im getting any, ig a i7-7700k just isnt good enough for this app? The audio from my video is 256kbps and 48khz sample rate. It says the codec for the clip is ATSC A/52A (AC-3) and format fltp. I tried exporting as both mp4 and AVI and both retained the audio glitches.

Ok, quick update, it’s definitely tied to using the pitch compensation. I mean, I need that. But yeah I went back and disabled the pitch compensation for all the speed adjusted clips and the audio plays back properly, but obviously its all distorted now. I also tried applying 1.00001x speed to the normal clips and then checking pitch compensate to those to see if it had to do with switching pitch compensate on and off and it didnt help. Version 24.01.28. Tried adding minimal transitions between the clips to no avail.

I did not reproduce it. It might be due to some combination of the audio in your clips and the speed values. Here are some tips. Instead of splitting clips and changing the speed of sub-clips, use the Speed: Forward Only Time filter. Instead of manually adding minimal transitions between the clips, use track filters (click the track header to select it and then Filters) Track Auto Fade Audio with a small duration like 0.05 or Track Seam Audio filters.

This documentation illustrates how to use the Speed: Forward Only filter with keyframes to change the speed up and back down without splitting the clip

I will try this soon and update, just wanted to return your lightning fast reply with my appreciation. Y’all are goated for being so incredibly helpful in these forums. For being a free app, I think this level of support is unrivaled worldwide. I’m not so sure this reply truly “improves the conversation” but you deserve to hear it.


Alright, speed forward only was more than just an effective workaround. The final product is superior to adjusting the speed of subclips because the temporal adjustments are buttery smooth this way. I laughed really hard thinking about what the audio waveforms SHOULD look like using this technique. Like yeah the way that the video looks on the timeline is kinda confusing at a glance, but actually makes sense when you think about how subclips work in shotcut, and I have zero complaints about needing to add or subtract to the clip length, but the audio waveforms are on a whole 'nother level of blatant lies xD It seems like such a convoluted way of getting the job done, but its really not and the workflow is actually really not bad whatsoever, and it’s very functional. This straight up works better and is far less buggy than the PS4’s video editing temporal keyframes. I still had some of the same audio glitches in timeline playback, but it came out clean on the export, and it actually rendered a good deal faster because of the lack of sub-clips which was cool. I have 2 main complaints. 1 being the audio quality of moderate to heavily slowed down clips, but I truthfully have no idea how competitive the quality is there. And 2, the fact that you inherently cannot set the end point of what youre temporally editing is kind of a pain. Even with the most basic setup, by just setting 3 keyframes, 2 at 1.0x speed and something else in between them, will cause the 3rd keyframe to change what video frame it was anchored to, and it changes further dynamically as you adjust that middle keyframe. My best workflow was to predictively offset the 3rd keyframe by a small amount and call it good enough. I would like to suggest a feature that anchors 2 keyframes to each other, maybe making them both purple or something, such that adjusting any keyframes in between dynamically shifts the later keyframe forward and backward on the timeline to compensate for all the timing changes in between the two purple/anchored keyframes. You’d still have to work left to right (even if you did the extra work to make it reversible you’d still have to do the whole temporal edit in one direction), which I really don’t love, but at least you wouldnt technically have to continuously readjust everything to get the exact results you want. Seems like a simple enough solution to take the feature to the next level. I don’t know if this would ever be precise enough to adjust EVERYTHING that comes after the final anchor point, but I’d imagine it would work as long as there’s no more keyframes to the right of it and you just focus on adjusting that one anchor point, so you can at least effectively set the beginning and end of a section to temporally edit before you start changing the actual playback speed. It’d be kinda like making subclips except also not at all. This feature as it stands today actually creates a very smooth final product, such that it’s very easy to create something that’s hard to tell it was even adjusted, aside from the audio. This mainly comes down to the fact that it just works, very well, despite a pain point or two. Smells like good coding to me. And in case any users read this, I’d suggest using the Blend setting. Thanks again for all the help!!!

I would be interesting if someone could do an A/B comparison with some other tools. The problem with slowing down audio is that you are effectively reducing the sample rate. For example, if your audio is 48Khz, but you slow it down to 0.5x, then Shotcut only has 24Khz to work with. The quality will seem lower for sure.

You should try the Time Remap Filter sometime. This is the most advanced form of speed change that Shotcut offers. It is a bit of a “head change” for many people. But if you master it, it provides all the detailed control you could want - including anchoring source frame numbers to output frame numbers.

This topic was automatically closed after 89 days. New replies are no longer allowed.