noticed that you've been getting more attention lately, and i just wanna say you totally deserve it :> ive been watching you for a while but i never commented anything until now. i really enjoy your content and im glad to see you doing better than ever, keep up the good work!!
Really cool project! UI question: Do you think it would be good to have the subtitles as a 'timeline' above the sound's, and being able to zoom and scroll might be more useful?
Hm, that's not a bad idea. The specifics get a little confusing though, e.g. what if the timeline is too zoomed out, what do we do the words? You could truncate them, but when you're fully zoomed out even one character would be too fat. I suspect that there's not enough value there for it to be worth figuring that stuff out. I can just click on the start and the end of a sentence and look at where on the audio waveform that ends up. Skimming on the timeline would be a little rough as well, so I think I would end up with something resembling the current UI as well as the words on the timeline at the bottom
Silly idea, you could speak a certain keyword during livestreams and then draw a blip on the timeline so you can find the interesting spots later on. Additionally you could do the same with some "onion skin" type chunk where if you know you want something in the summary you say something like "record this" and "end recording" and the start end frame would happen right after and right before so those aren't included etc. Similarly it could be used to highlight words you can mass cut out after if you search the video for that word
Of course, but we have to start somewhere. I actually fixed this off stream yesterday github.com/sphaerophoria/video-editor/blob/1dab734bc24ab5f3798f116c7953fac6c800b031/src/WordTimestampGenerator.zig#L63
github.com/sphaerophoria/video-editor/commit/144b192b302d3c0250fecc2d2d5813e35a9dda84
noticed that you've been getting more attention lately, and i just wanna say you totally deserve it :>
ive been watching you for a while but i never commented anything until now. i really enjoy your content and im glad to see you doing better than ever, keep up the good work!!
Can’t wait for this to work so I don’t have to listen to the pitch again 😂
Really cool project!
UI question: Do you think it would be good to have the subtitles as a 'timeline' above the sound's, and being able to zoom and scroll might be more useful?
Hm, that's not a bad idea. The specifics get a little confusing though, e.g. what if the timeline is too zoomed out, what do we do the words? You could truncate them, but when you're fully zoomed out even one character would be too fat. I suspect that there's not enough value there for it to be worth figuring that stuff out. I can just click on the start and the end of a sentence and look at where on the audio waveform that ends up. Skimming on the timeline would be a little rough as well, so I think I would end up with something resembling the current UI as well as the words on the timeline at the bottom
I am going to be sad when this is over
Silly idea, you could speak a certain keyword during livestreams and then draw a blip on the timeline so you can find the interesting spots later on. Additionally you could do the same with some "onion skin" type chunk where if you know you want something in the summary you say something like "record this" and "end recording" and the start end frame would happen right after and right before so those aren't included etc. Similarly it could be used to highlight words you can mass cut out after if you search the video for that word
Maybe there’s a way to add some kind of metadata to the video while recording?????
Like bookmarks that can be added by a hotkey…
@@creator-linkhe already does that on the scratchpad
Now make the text selectable to determine the clip segment :)
Maybe you should resample the audio to 16k/s mono on the fly, to feed to Whisperer, instead of preparing a resampled video.
Of course, but we have to start somewhere. I actually fixed this off stream yesterday github.com/sphaerophoria/video-editor/blob/1dab734bc24ab5f3798f116c7953fac6c800b031/src/WordTimestampGenerator.zig#L63