@Sushi has a program he uses to pull these down as the start to his transcript updates, and we already share a git repo, so he could just start checking these in.
1 Like
Iām not Sushi but: You mean the original VTT files? If yes, then yes, they include the informations as well:
...
00:00:06.560 --> 00:00:12.780 align:start position:19%
[Music]
00:00:09.650 --> 00:00:15.360 align:start position:19%
hi<c.colorE5E5E5><00:00:10.650><c> I'm</c></c><c.colorCCCCCC><00:00:10.889><c> Ron</c></c><c.colorE5E5E5><00:00:11.099><c> Gilbert</c></c><c.colorCCCCCC><00:00:11.309><c> I'm</c><00:00:11.880><c> here</c></c><c.colorE5E5E5><00:00:12.150><c> winning</c><00:00:12.480><c> and</c></c>
...
In the exported TXT files there is only a hint to the music without the times:
[Music]
hi I'm Ron Gilbert I'm here winning and
this is our first stand-up meeting
...
Thanks. What I meant was - does the initial music seems to be detected in all the podcast original āauto-transcriptsā that Sushi has used so far. Iām wondering if YouTube is reliable with the music detection or not, or if we need to depend on a manual step to deal with it (which Iām trying hard to avoid). I believe Sushi has downloaded all 3 caption formats (or could), for all the podcasts already with the script/program he uses for that.
1 Like
This is brilliant!
The only extra effort might be caused by the podcasts with a special guest then.
1 Like
The podcasts are containing the names of the speakers. So a script could pre-process the podcasts and ādetectā how many speakers are there. 
2 Likes
So, the way Iāve written podcast_animate is that I pre-parse Sushiās transcript (which has all the speaker names), and then I know all the Speakers for that podcast. I determine the speaker layout depending on how many there are, and load their image data based on their name (directories are named after the speaker name used in the transcripts). This next part isnāt yet implemented but the plan is, if I canāt find image data, I will default the āGuestā speaker using some default mechanism, so the animation will work, but have some āplaceholderā speaker image, or maybe make it funnier, like have the guest speakers talk from off the edge of the screen, or somethingā¦
2 Likes
Well actually, I got these srt files from @Someone, so ican add them on git. No need to rerun those.
2 Likes
Oh cool, I didnāt realize (or remember) @Someone was the source for those captions for you. Thanks, @Someone!
1 Like
Iāll add them to git in a minute
2 Likes
How about we do our own weekly podcast? Use something like a google hangouts group, talk about mighty interesting stuff like californian weather and someone records it and puts it online? Would that be in any way doable? Anyone interested if that were the case?
After a bit of a hiatus (things got pretty busy at work and home), I have Demo 4 available for the podcast_animate script. Itās now able to play any of the 8 podcasts we have transcripts for. I still havenāt coded the eye movement, but you will notice many other improvements. I used podcast 7 for this run.
10 Likes
THIS IS AWESOMELY FANTASTICALLY ABSOLUTELY GREAT!!!
1 Like
To me itās very pleasant to listen, to readā¦ and to understand!
and the soft lights on their heads, itās a nice touch! Really congratulations!
1 Like
ThatĀ“s amazing. That pan at the beginning is fantastic!
One thing though. Would it be possible to make the font look more like the one in the game. The way it looks now just feels a bit off to me, but everything else is brilliant!
2 Likes
Iāve only been sporadically ogling this thread but this is amazing. Iām in awe of you guys 
I just finally clicked on the video link on this post. Yes, I had completely forgotten about this excerpt from the end of the Kickstarter video. I seem to remember now having an instant uncontrollable urge to increase my backer contribution when I saw this.
1 Like
I love it because that kind of joke where someone gets all worked about about an imaginary situation he himself created reminds me of this scene from Duck Soup
1 Like
Yes, that worked out better than expected. There wasnāt much I needed to do specifically to get that, other than just properly connecting the RadioStudio layers with alpha transparency (included in the PNG files). So I put down the background layer first, then the speakers, than the foreground, which includes the transparent window glass which has the light rays painted with transparency in the same image as the glass, so it gives the effect of lighting, essentially āfor freeā.
That all said, I think I need to make some changes in that area, since it isnāt practical to position each speaker under the fixed position lighting. I might need to alter the window layer to remove the lighting, and use the standalone ālight coneā sprite on the same RadioStudio sprite sheet to get the same type of affect, but where I can position it directly above wherever I need to place the speaker.
There was some earlier discussion on this from @Gffp :
I think this could be done with the changes I mentioned above - Iām going to look into it as one of the changes for the next version.