Improving the transcription quality (PotPlayer )

[milk052701]

In video recordings, normal dialogue is captured clearly, but whispers are difficult to record due to their low volume. What methods or techniques can improve the quality and accuracy of capturing whisper audio without relying on individual audio tuning for each video?

[Purfview]

Can you share a sample of the problematic audio?
Please post the full command you use for transcription.

[milk052701]

Hi, I’d like to check if the XXL Pro model is ready yet. I’m currently using PotPlayer (64-bit) with the large-v2 language model. Is there a version specifically optimized for Japanese? Also, I’ve been generating subtitles with all default settings. If I still encounter the same issue after receiving the Pro model, I’ll report it again. Thank you! Purfview ***@***.***> 於 2025年10月20日 週一 下午2:21寫道：

…

Can you share a sample of the problematic audio? Please post the full command you use for transcription. — Reply to this email directly, view it on GitHub <#523 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A7KNKGWDI2GJBZMN5JFL66L3YR5ODAVCNFSM6AAAAACJUVJOCCVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTINZSGU4TEMA> . You are receiving this because you authored the thread.Message ID: <Purfview/whisper-standalone-win/repo-discussions/523/comments/14725920@ github.com>

[Purfview]

I’m currently using PotPlayer (64-bit) with the large-v2 language model.

PotPlayer is not the right tool to generate proper subtitles.
As far as I know, it takes text from the progress output, which isn't actual subtitles.