Purfview / whisper-standalone-win

Whisper & Faster-Whisper standalone executables for those who don't want to bother with Python.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

--highlight_words true --max_line_width 43 --max_line_count 2

nielka98 opened this issue · comments

First of all, I'm sorry because I think I pissed you off last time, which wasn't my intention. I'm quite a clumsy person and sometimes I have to explain something several times, like a cow in a ditch. This is the additional feature that childhood cerebral palsy gives.

Secondly, I pasted this command because I use it and the results it gives are amazing, but it makes the lines repeat. There are an awful lot of them. Example: when I use the --standard command, there are about 1500 lines. And when I use the title command, there are about 10,000 lines. I have a question whether this is a bug and whether it can be fixed somehow, or is it just the nature of this command and you have to love her with all the benefits of the inventory.

First of all, I'm sorry because I think I pissed you off last time, which wasn't my intention.

Not at all,

I pasted this command because I use it and the results it gives are amazing, but it makes the lines repeat.

That's normal, it makes karaoke subs, if you don't want karaoke subs then don't use highlight_words.
What is wrong with --standard?

--standard It is less accurate...I don't know what you did with the commands I use, but they are really much better

Could you tweak it so that it has a standard command?

All these command mentioned here don't have anything to do with any "accuracy".
So, basically I can't make sense of what you've posted here.

Sorry, I'm very tired. What I mean is that the inscriptions on the title command are Better fit than on --standard

What means "better fit"?

Better placed They match the audio better. I don't know how to write it better, I'm writing it using Google Translator, so the translation doesn't have to be good either

All those commands mentioned in this thread don't have any influence on "match the audio better".

I don't think that you'll go anywhere without posting the examples.

I mean that the subtitles are better matched to the sound.

Nie sądzę, że nigdzie nie pójdziesz bez zamieszczenia przykładów.

😂😂😂
I'll attach the log to you tomorrow

Maybe you want to read this again -> #256 (comment)

That means that those commands have same influences on "match the audio better" like a water tap turned on in your kitchen.

So, don't waste your time.
Probably you observed non-deterministic effect of Whiper model.

Okay, I'll see and experiment... You are going to add your Whisper XXL to the Subtitle Editor?

Me not, I dunno C#, ask nikse.

Can I use this topic as it's for you? I have many questions and I don't want to open new topics all the time

Always open a new topic for a new question.