An example with the Back to the Future clocks, with different colors of the text for each speaker, with location. They show date and time when the podcast originally aired.
Also in the radio station speakers could seat and show only the upper half of the body (easy to complete).
This picture is also accidentally nice for the characters staring at their creators (when they will be there ) trough the glass.
To emphasize the current speaker, as besmaller asked, the three (exceptionally four) spotlights could be used. When a speaker talks, the spotlight above him becomes brighter.
Itâs not a bad idea. Though, the items on the console (and possibly the crack in the window as well) would have to be removed. Elsewise, the faces would be partially covered by them.
By the way, I partially agree with @Gffp that a split screen would be better, because they occasionally talked about their different locations during the podcasts. But the selection (or creation) of appropriate room screens would be difficult. Also, Ron has moved in between. Would we really like to use different screens for different podcasts, depending on who was where at that time?
I wouldnât dislike seeing them in the broken radio booth, even if the cracks might partially cover their faces. I like the idea of putting them in the decadent Thimbleweed Park.
But we could just test a few different things and see what looks better.
Well, I guess that it would be quite feasible with the layered version of the radio station. Are we able to overlay everything with the transparent blueish window glass there?
I still think that we should rearrange some items on the console, if we really decide for the radio station. In my opinion, the lever is wasting too much space, if there are three or four people in the studio. Also, everyone ought to have an own microphone.
Moreover, I think that the audience should not stand as near in front of the window, because it would make the audience in the foreground more eye-catching than the people in the studio. This is actually a disadvantage of the radio station, because the people in the studio are pretty much in the background - in contrast to the Diner, where the audience would be in the background, looking in our direction.
Ron has provided the layered images from the Radio Studio (thank you, Ron!), so I hope to include some initial background using this for the next demo. It will take some time to put this all together, but I think it will be worth it.
Itâs nice and very helpful to hear all these creative ideas. Based on my current availability for this project (an hour or two a night at best), some of the ideas are impractical, but perhaps can be added in the (distant) future. Remember that my personal goal for this project was a fully automated flow, where I would just provide the script the the podcast number, and it would use only the:
Original podcast mp3
The transcripts from Sushi (with help from the YouTube captioning and additional scripts from Sushi and myself)
Rhubarb lip-sync data (automatically created by above 2 inputs, with some scripts)
The character spritesheets (provided by Ron, with some effort to locate appropriate images within the sheets and extract them)
Background image (provided by Ron, also will effort to pull layered components and assemble them onto the image)
With this only this info, the script would fully automate the animation (no special case code for specific podcasts). So theoretically, if I get the script working well for a single podcast, it will then work automatically the 66 other podcasts with no changes needed.
This sounds like a cool idea (changing the image in some way like this for this sort of âintermissionâ segment). If the manual transcript annotations would add a âscriptâ item to indicate the intermission start time in some standard format, this could be done in the podcast_animate script automatically. I seem to remember something like this happening in more than 1 podcast, which would make this feature even more worthwhile.
No. Not a music detector. In fact the intro music seems to confuse it (as I mentioned in an earlier post) causing some problems.
If anyone knows of a âmusic detectorâ type of program I could run on the podcast, to output time stamps for when music starts and stops, it would actually really help.
YouTube did that already. These parts are marked in the (converted) SRT files, for example:
1
00:00:06,560 --> 00:00:12,780
[Music]
2
00:00:09,650 --> 00:00:15,360
hi I'm Ron Gilbert I'm here winning and
3
00:00:12,780 --> 00:00:17,640
this is our first stand-up meeting
You could write a script that removes all rows except the ones above a line with â[music]â. With this approach you have the times for the music.
If you would like to have the SRT files, please send me a PM.
Interesting. Thatâs a good point. I could use the YouTube original, unaltered transcripts as input to the podcast_animate script, for certain information (like this).
@Sushi, do the other original files from YouTube, (not the retimed ones after your cleanup) work like this as well?
