Quote from: sark on January 11, 2024, 12:45:24 PMImproving the ability to accurately read and match parameters between encoded and unencoded sections of video streams could indeed be a promising avenue for advancing lossless cutting techniques. This may involve advancements in encoder technology, as well as improvements in algorithms for analyzing and processing video streams.Quote from: Andrew-R on January 11, 2024, 02:03:49 AMHello, all!
I was looking into this problem (minimal reencode of streams with non-all-I videocodecs)
and today I found patent
https://patents.google.com/patent/US9392279
"Method and system for generating an instantaneous decoding refresh (IDR) picture slice in an H.264/AVC compliant video data stream "
Does it sounds like step in the right direction for smart-copy of h264 files?
That patent was applied for ten years ago. Doubt there is anything new there.
The bottom line is that you cannot reference frames that do not exist. Without major encoder changes, I cannot see how you can get around this.
The current crop of lossless cutting apps work by reencoding the first and last GOP, and concatenating back. The problem is matching the encoded GOP sections parameters to the original unencoded section. infinite craft
I guess the ability to better read, and match those parameters is the way forward. Currently though, I would not trust lossless cutting output that uses this method, even when it appears to work.
Quote from: eumagga0x2a on April 26, 2022, 01:30:05 AMShould be fixed by commit [core] Support paths with non-ASCII characters in ADM_PathCanonize function on Windows, fix simplify_path(), restrict its scope to file, remove tabs heardle please try a future nightly build.Testing the changes with a future nightly build is a good idea to ensure that everything is working as expected. If you encounter any further issues or have additional feedback, feel free to report them. Thank you for bringing this issue to our attention, and I'm glad to have been able to assist you!
Thank you very much for your report.
Quote from: jimmyjim on February 10, 2024, 03:10:15 PMIf you are using python (i've never used it), I recommend asking the user for the csv file and reading the csv file into the script: https://www.w3schools.com/python/python_file_handling.asp
Also, if you are using loselesscut to get the timestamps, why not just cut the video with it?
Quote from: Drakon66 on May 04, 2024, 09:16:28 PMI hope option with reencoding not reduce quality.- did you try this?:
Quote from: Geo_log on May 01, 2024, 02:58:13 PM1) Append 2nd video;How good or bad is the resulting quality?
2) "Video Output" -> "Configure" -> Encoding Mode = Video Size (Two Pass) = Vid1 + Vid2 [Mb];
3) Save video.
Quote from: eumagga0x2a on May 01, 2024, 04:31:07 PMAn enhancement to Avidemux to allow user to rotate video preview is possible
Quote from: eumagga0x2a on May 01, 2024, 04:31:07 PMPlease make sure that "Post-Processing" is disabled in Avidemux
Quote from: eumagga0x2a on May 01, 2024, 04:31:07 PMlease also consider using a lossless codec (Ut Video or (FF)HuffYUV) for intermediate steps.