I’m seeking help with matching video to audio (from a DAW) where the wave-forms are off shorter than what the timeline allows for nudges. Basically, with a full nudge (say :04 to :05), the video is behind. And likewise, the video is ahead if I nudge the other way. Is there a way to nudge or even cut by 1/2 a ms?
Hi, and welcome to the forum and community.
I’m afraid this is not possible because the smallest unit is the frame. If you know how much the audio needs to be nudged use Audacity to remove the 1/2ms and then bring it back to Kdenlive.
How noticeably out of sync is it really in practice when watching the result?
Even sharp transients in audio tend to be “quite long” in comparison to the duration of a single video frame, so slight (sub frame-duration) misalignments usually aren’t all that noticeable - and perceptually our brains quite often expect sound to slightly lead or lag visual action anyway.
Aesthetically, I find it tends to be more important to carefully match the duration of a visual transition to the duration of an audio transient than it is to have them perfectly synchronised.
What Bernd says is correct, if you really need alignment with audio sample granularity, you’ll need to pad or trim the start of your audio.
But the bigger problem you’ll have if what you’re doing really is that sensitive, is that no matter how accurate you get it in the editor, all bets are off as to how well whatever it is played back on maintains the synchronisation between audio and video. It can easily vary by (many!) 10’s of milliseconds across the whole spectrum of playback devices.
Some players allow the audio delay to be tweaked to account for hardware and other buffering inserting variable delays - but that still means you aren’t guaranteed to get the result you’re hoping for from all that work to get sub-millisecond synchronisation unless you also give viewers a way to “calibrate” their player first …
(and yes, I have a test clip with a sweep clock and audio pips because I’ve seen set top boxes that can be 100’s of ms out of sync and that definitely is glaringly noticable).
Thank you both for the input.
I’m basically recording drumming vids of myself. The wave form output from the DAW will almost match stride for stride with the WAVE form output from the GoPro audio, so it’s very easy to align the Audio/Video of the MP4 (GoPro) to that of the MP3 (DAW). But, I get stuck in these situations where the cut based on framerate will force me to shift the vid just ever so slighly forward or backwards giving the appearance that “its off.” It’s not too drastic but I still notice it … so I was hoping there was a way to nudge by custom increments vs whole ms framerate. That not being possible, I’ll run with your suggestions and add that micro-ms to the DAW audio export.
Thanks again to both of you!
Yeah, if you’re a drummer, especially watching your own drumming, you’re going to see things most other people won’t ever notice …
If you record a “long enough” session, you may also run into a secondary loss of sync due to the two devices running on separate clocks and Not Quite having the same idea of exactly how long a millisecond is - so you may need to trim or pad more samples every few minutes to keep the precision you initially align with. I’d expect the DAW clock to be quite good, but gopro timekeeping can be pretty sloppy.