Hello, I use DeepL Pro API to translate srt files and for most languages I do not experience this problem. It is when I attempt to translate to ZH-HANS, Japanese, or Korean, and also Turkish that I often get many lines of srt text all concatenated onto one line further down in the transcript.
It results something like this where it looks like the text in those timestamps all get pushed to a later timestamp:
39
00:00:50,973 --> 00:00:52,533
毕竟 比糟糕的网络摄像头视频要好看些
40
00:00:52,533 --> 00:00:53,703
41
00:00:53,703 --> 00:00:54,873
42
00:00:54,873 --> 00:00:56,313
43
00:00:56,313 --> 00:00:57,813
44
00:00:57,813 --> 00:00:58,863
。 但归根结底,如果 网络摄像头 视频 的内容 本身 够精彩,人们 照样会看 。
就像现在还有无数人观看 过去 伟大 音乐家和艺术家的 演出录像 。
Let me know what steps I could take to try to resolve this. I am calling this endpoint: https://api.deepl.com/v2/document