On Fri, 2006-01-13 at 14:44 -0600, John Lightsey wrote: <snip> > Mixing formats is something that should be avoided if possible. <snip> That certainly complicates affairs, yes. Also there was one talk last year (number 6 on the 16th) that appeared to have a 14:9 frame ratio. I'm guessing this was due to accidental selection of the wrong format on a camera or perhaps some incorrect post-processing that cropped PAL frames to NTSC size. <snip> > > The sound level was not consistent enough. We should figure > > out some guidelines for level control, to keep the dynamics > > within a rather narrow range. Some compression and limiting > > may be required. > > Normalizing all of the audio prior to compression would be a nice > improvement. Yes, and maybe compression (that is, dynamic volume normalisation) too. Loud applause seems to have skewed the initial volume adjustment, and also hampered my attempts to normalise the volume between talks. I have yet to find a tool that will do compression automatically, though Nattie has had some success doing this to audio files in Audacity. > Chopping videos at the beginning and end could be minimized. If the > moderator gave a signal to start the talk, waited 3 seconds and then > began, we wouldn't need to crop the front. On the tail end, stop the > recording as the applause dies down. You'd have to convince tolimar to > limit his introduction to "Here's Blah with a talk entitled Blah" > though. :) This didn't seem to me to be too difficult, except that I didn't have a good video editor with which to get precise timings. > And the DVDs are spectacular. Thanks! > I hope copies of the 2005 DVDs will be sold at Debconf. I had thought of that, but am not sure of the logistics. > What more can be done to help/improve the production > of DVDs? 1. I and anyone else who wants to be involved in "post-production" should be able to take home good-quality copies of the recordings (like the MPEG-4 AVIs you sent me for this year). This of course requires that they are all converted from DV by the morning of 22nd May. 2. I/we should settle on a reliable batch video encoder/processor (mencoder worked, but was a real pain to use; I know the rest of you use other tools). Similarly for remultiplexing if necessary. 3. We should be wary of losing audio/video synchronisation. There were a few videos where the audio was between 100 and 300 ms late (my estimate). 4. I should get WebDVD ready for production. These are the fairly essential items yet to do: * Integration with video recoding. * Fix resolution of relative filenames in VOB lists. * Provide a means to specify video aspect and menu & audio language. * Allow explicit ordering of videos. 5. We should make a note of any changes to session titles. The files provided for download were named according to the titles stated in COMAS and on the schedule. However, a lot of speakers used somewhat different titles in their introductions or on their slides. 6. For two-part sessions, if they are happening this year, it would be nice to have the parts joined with a fade in and out rather than distributed as separate files. (Although using separate files makes chapter marks automatic in dvdauthor.) I should also send a sample DVD image to Rich Wareham (author of libdvdnav) so he can try fixing the bug that last year's DVDs triggered. Ben. -- Ben Hutchings In a hierarchy, every employee tends to rise to his level of incompetence.
Attachment:
signature.asc
Description: This is a digitally signed message part