Digitizing my old tapes was one of the most rewarding side projects that I did over the last year. I managed to get in under the wire (pun intended) of Firewire compatibility on Sequoia and a long daisy-chain of adapters. But it was clear the days of this approach were numbered. I'm optimistic these 3rd party accessories will become more standardized into self-contained cheap boxes where people can easily transfer over their stuff before camcorders degrade.
My pipeline went camera -> dvrescue -> ffmpeg -> clip chunking -> gemini for auto tagging of family members and locations where things were shot.
We now have all our family's footage hosted on a NAS with Jellyfin serving over Tailscale to my parents Macbooks. I found the clip chunking in particular made the footage a lot more watchable than just importing the two-hour long tapes although ymmv.
I used dvgrab to ingest my old tapes, and ffmpeg and avisynth/QTGMC to de-interface and encode files for easy viewing (though I keep the original .dv files).
The biggest issue I ran into was that while the audio and video were properly synced up in the original .dv file (due to it being an interleaved format), when I re-encoded the videos, the audio and video would drift out of sync as the video went on.
I was able to fix the sync issues by using dvgrab to split the original dv file into a bunch of 3 minute chunks. I then wrote a script to extract the audio track from each chunk, pad the end of the audio with milliseconds of silence to the exact length of the video track, combine the padded audio tracks, encodes the combined track, and muxes the fixed audio track with the encoded video. This worked really well; the silence padding is imperceptible, but the audio and video are still in sync - even after 2 hours.
A final point that needs making is that doing anything with dv files in ffmpeg (even -c:v copy) destroys the SMPTE timecodes embedded in the original file, making it much harder to split by scene.
Any idea how long the tapes will last? I have a bunch sitting in a box that I should probably verify have all been imported already before they turn to dust.
It would be fun to use this to build something like the old capdiv firewire hdd recorder I have. then you can just capture direct to disk and not worry about the 1:1 transfer time later.
12 comments
[ 2.9 ms ] story [ 30.4 ms ] threadMy pipeline went camera -> dvrescue -> ffmpeg -> clip chunking -> gemini for auto tagging of family members and locations where things were shot.
We now have all our family's footage hosted on a NAS with Jellyfin serving over Tailscale to my parents Macbooks. I found the clip chunking in particular made the footage a lot more watchable than just importing the two-hour long tapes although ymmv.
The biggest issue I ran into was that while the audio and video were properly synced up in the original .dv file (due to it being an interleaved format), when I re-encoded the videos, the audio and video would drift out of sync as the video went on.
I was able to fix the sync issues by using dvgrab to split the original dv file into a bunch of 3 minute chunks. I then wrote a script to extract the audio track from each chunk, pad the end of the audio with milliseconds of silence to the exact length of the video track, combine the padded audio tracks, encodes the combined track, and muxes the fixed audio track with the encoded video. This worked really well; the silence padding is imperceptible, but the audio and video are still in sync - even after 2 hours.
A final point that needs making is that doing anything with dv files in ffmpeg (even -c:v copy) destroys the SMPTE timecodes embedded in the original file, making it much harder to split by scene.
I'll have a shot of this I think.