itch.io

An independent app that does full translation of subtitles and text from any language to any other one, is indeed very interesting!
It should of course have GPU support as well as CPU support.
It would be truly awesome if it could also have the support for converting .SRT into .VTT (WebVTT). That format has a header (3 lines) and all dots in the time-stamp unlike .SRT that has one comma in their timestamp. It also has the ability to place text at different positions on the screen, but the normal/default captions/subs position we're all used to is:  align:middle size:95% line:95
The incrementing block-numbering before each block of "timestamp+position+text" as one has in .SRT, is optional in .VTT
Example (WebVTT) .VTT Subtitle/Captions File:

WEBVTT
Kind: captions
Language: en

00:00:00.000 --> 00:00:08.380 align:middle size:95% line:95
So how do you make a web page that looks like this where you have a video and it shows a

00:00:08.380 --> 00:00:14.680 align:middle size:95% line:95
nice picture where you can put the title of the video or information about what the video

00:00:14.680 --> 00:00:23.560 align:middle size:95% line:95
is about and it has obviously this video player which has this hamburger menu over here

So as you see, very similar to the .SRT format, but the benefit of this format is that it can be used with HTML5 players, so you can upload it directly to youtube (or other services supporting sub/caption) and on your own homepage, if you need video with subtitles.
PS: I've made a python script doing this .SRT to .VTT conversion, that i can share it with you, if need be. (not C++ code, but just as easy to read, he he)

I found a bug (i think) in the new 241 version.
When selecting the most accurate (biggest) model, it outputs random garbage text for quite a while, then after a few minutes into the video it starts picking up the correct language, and translates it correctly.
On the default middle setting it translates the video fast and correct!
PS: on the most accurate setting it also often crashes the app at the very end of loading the video, before it begins transcribing it (.mp4 720p vids).
The video file itself is not the issue, since the middle setting does the whole video correctly without crashing.
The program has no error-log when it crashes, so I cant be helpful in providing that info, so maybe an idea to implement such a log in the future? (see image below for results)

Hey! Great work!
The new GPU version (v241) is waaay faster than the older (v214)
(PS: important question below the picture, so view the rest of this post)

Is it possible to have it translate to other languages than "English" ?
I assume the AI is first transcribing the audio to text in original language, then the AI translates the sentence to English also using the AI, right?

If that is the case then a dropdown list of languages the AI supports would make it possible to translate a german video into norwegian subtitles, or as now getting the german video into an english subtitle.
Can you possibly add this? Or would it be more complicated than I think?

Nice! Looking forward to your next update!

Nice! Looking forward to the newer version then! Any approx. E.T.A for that?

Does this app unload the image model each time it has produced the number of images you set it to make (1 by default)?
Because after testing the demo it seems like that to me, and my suggestion would be to find a way to keep the model loaded for a while longer, so one may edit the text (within a few min max) and press Start (to generate it with the changed words), without having to reload the model to memory. Is that possible? Maybe display a countdown until the model unloads from memory?

What are the minimum requirements for CPU and RAM, and/or free space available on HDD, to run this app?

And in the future, will this app also support small and bigger GPU's for those that have those?

Very nice app! Love the idea of a fast offline transcriber/translator!
I am currently testing the demo,  and after doing so I paid and bought this software!

I have two questions I'd like to ask: Do you have an approx. e.t.a (month/year) for when the GPU / CPU switching possibility will arrive in an updated version?

And also I wonder if the app is temporarily writing the text it translates from speech to txt, to a txt file, like every minute or so, just in case the app might crash for unforeseen reasons. So that you may read some of it, if you have to restart the procedure due to a random app-crash?