Good news: wallet found!
Bad news: still in video encoding hell.
Since I’m sure someone else on the face of the internet is experiencing this issue, allow me to expand:
My project: Shoot video with my Flipcam while I record audio in my home studio. After mixing the audio, sync it to the video in Adobe Premiere for a studio-quality music video to post to YouTube.
Sounds straight-forward, yes?
The Flip is certainly straight-forward – about the size of a pack of cigarettes and operates with a single button. Its 1280×720 isn’t the crispest, but it does well in all sorts of lighting conditions, and can absorb loud sound at concerts without clipping.
That said, the sound is still through a relatively tinny single mic, so adding stereo multi-track audio from my studio marks a vast improvement.
The problem comes when I import the MP4 into Adobe Premiere. It looks beautiful! However, its timing is every so slightly off – compared to the audio track the video falls increasingly behind. The difference is less than a second, but enough to ruin the visual sync of the audio to the video.
Not only is it visible against the video, but you can hear it via an increasing echo if you turn up the audio from the Flip. And after encoding the problem seems even more pronounced.
I’ve been trouble-shooting this for 72 hours, and I can’t discern the source of the problem. So far, I have:
At the moment I am truly and completely stumped. On one hand, it could be that I’m simply not unpacking the MP4 file correctly into a format that I can edit with.
However, my growing suspicion is that the Flip is dropping and/or inserting some frames, and it would only take one or two “skips” to throw the video off several milliseconds against my audio recording.
I lucked out on Monday with “Icy Cold,” which lags just a hair, but since then I’ve been completely frustrated.
Unless some video superhero comes through with an explanation and a fix it looks like I’ll be hawking my Flip to step up to a more pro-sumer model for my upcoming projects.
MPEG is a compressed format, meaning it uses a combination of dropped frames + keyframes to make up for the lack of real data. When you “decompress” the MPEG, those frames are gone forever, so they have to be recreated. This is an imprecise science. Since the Flip is recording compressed video, you’re not recording with any sort of frame-by-frame accuracy.
The songwriter’s job is never done, eh?
PS: Could it be the audio that’s off? It’s possible, but not probable – I’ve been using Cubase for over two years, and my DAW is customized for it. It’s certainly not a logical explanation