I’m working on the transcript for #5 in the Felix Garcia series, and at the same time, working on the video for #6. For today’s post however, I thought I’d give you some insight into the tech work behind these videos. Bringing you this interview is really a pretty big job.
The files came to me originally on 2 DVDs, each one-hour in length. I extract them from the DVDs in the MP4 format. They are about 4Gb in size, 1280 by 720 frame size, at 30FPS. They are progressive interlaced. The audio track is 16 bit stereo – 44.1 sample rate.
I convert them into WMVs with no compression or changes in frame size or rate. I then cut them into 4-15 minute segments. The uncompressed segments are rendered and copied over onto a private file server, where the interpreter can download them. She watches them in real time, while recording a separate audio file, which she then uploads back onto the file server.
Her audio file is mono, and contains her translation voice-over, + all the audio information from the original. Using Adobe Audition on my audio editing machine, I copy her entire file and make a duplicate track, so the file will be in stereo. I then do some post-production mastering on the new stereo file, including audio compression and equalization. Finally, I normalize the two-track mix to –6Db.
Still on the audio editor, I extract the original audio tracks from the segment and master them as well, including the –6Db normalization.
Now, I switch over to my video-editing machine. This is my most powerful computer. Core 2 duo, with 8 gigs of DDR-3. I use 3 video editing programs on this machine – Adobe Premier, Video Edit Magic, and my personal favorite, Serif Movie Plus. Premier is the most professional, and it can do more heavy lifting, but it’s insanely priced and cumbersome to use. Video Edit magic is cheap but limited. Movie Plus is fast, easy to use, does almost everything I want – and doesn’t require a second on the house.
I now have 3 independent files to work with – the video track, and two separate audio tracks. I sync the original audio track back to the video track as a reference. I pan this track all the way to one side. I then insert the voice-over track and pan it all the way to the other side. I turn both tracks up full, and using the stereo separation, manually sync the voice-over track to the reference track. Once I have them synced up with no perceivable doubling or echo, I re-pan them back to center and reset the mix volume to –14dB.
Almost done. After readjusting the aspect ratio to allow for the captions, I have to insert yet another video track for them. This is actually a separate video file that gets created in real time and synched to the other 3 files. I have to listen to the audio tracks, and manually type in the individual captions. This gets separately rendered into a file and re-synched back into the mix.
Finally, the file must be compressed for uploading to the site. This is tricky, because I need to make it much smaller, but if I make it too small, the quality suffers. It can become choppy, garbled and hard to watch. After some experimentation in both what I would accept for quality, and what WordPress will accept for size, I ended up with a file that is 50Mb, down from 5 gig – 512 by 288, still at 30FPS. The audio is still 16 bit, 44-1 and other than loss due to the algorithm most of the trimming is in frame size.
I re-render it to this spec, and upload onto the site. All together, it takes a little over a week to make each segment.
I hope you enjoyed this little foray into my sick, sick world.
Filed under: Editorials | Tagged: Audio Engineering, BitcoDavid, Captioning, Deaf in Prison, DeafInPrison.com, Interpreter, postaday, Video, Video Editing and Software, Video Engineering | 6 Comments »