5.1 audio in Video Enhance AI

We really need Multi-Channel audio as part of the program. Its a long process to use one program to rip the audio then Video enhance AI to fix the video then another program to merge the audio.


I agree. What program do you currently use to merge the Audio?

Doesn’t Video Enhance AI use ffmpeg? Seems like this should be doable.
ffmpeg -i input.mov -i audio.ogg -map 0:v -map 1:a -c:v copy -shortest output.mov

To remap in 5.1 audio using ffmpeg could just use the transcode option flag:
-af “channelmap=4|0|1|2|3|5:5.1”


There is software that will multiplex audio without you having to extract it from an existing file. I believe avidemux and mkvmerge both do this, not to mention ffmpeg.

MKVToolNixGUI :slight_smile:


Yes, this works. One single file at a time. This is a no go for batch processing multiple seasons of a show for example. Unnecessarily tedious when this can be easily (but apparently not?) implemented within VEAI.

I agree it should just be implemented. I read that they has corruption when using 5.1 audio so they removed it but it just makes unnecessary steps if they can just work it out and add it

I thought my video result having no sound is a bug. So now they removed the sound but leaves the “keep sound” option intact? Since what version? I was using 1.8.1 and the audio came out just fine.

2 Channel Audio works the problem is with 5.1 channel audio. (Dolby, DTS and Atmos) They removed the feature a long time ago because it was causing corruption. I just with they would rip the audio then process the video then re-add the audio all within the process.

1 Like

So in 2.2.0 version the output will always lose its audio because thats the way the program works? Even with the “keep audio” option turned on? So why would they keep the “keep audio” option still available in the UI?

If you are using a video with 2-channel audio it will keep it. That is my frustration they need to add 5.1 audio support


Dealing with 5.1 audio in Windows is simple. A really simple Powershell script can call either ffmpeg (if you also want to use x265 for the video - something that is advisable for 4K upscales) or mkvmerge (if all you want is to select audio from the original file and video from the VEAI processed file) whilst automatically processing all files in a directory with automated renaming. I’ve done this for processing of entire series.

That said, a simple option for Topaz would be to allow the user specify the ffmpeg options in Preferences.


I have a hard time comprehending how the audio handling of this program is still in the state it’s in.


I disagree. It’s a needlessly convoluted process to drag audio across various stages of video preproccessing. And it introduces many opportunities for errors trying to convert it. Or you leave the audio ‘as is’, in which case you’re better off just muxing it in yourself, at the end of your process. Let me put it this way: if you have time to wait more than a day for your video to be processed, you have time to demux your audio first.

Besides, leaving say, a DTS-MA stream in, only makes for unnecessary I/O (back- and forth seek times between frames for any type of temporal processing). In all my 20- years of video processing, I have literally never done so with the audio still residing in the stream.

I’m upscaling / converting an tv show episode, wich play perfectly on my pc, but after AI is finished…no sound…
ATM another episode is running and…maybe when im uploading it to my NAS server it wil play with ’
sound ?
For upscaling i use Prometheus ( auto settings ) keep audio.

Absolutely. I’m currently processing MASH upscales, and the 96kbps is OK for that as its almost entirely spoken word. However, for high quality shows/movies/videos, we need the ability to select “Audio Pass-through”. This way, we’ve prepared a file in say Handbrake and chosen a quality audio format, whatever that is. Rather than having VEAI reprocess the audio, it should just reattach the unaltered high quality audio from the original file onto the outputted, upscaled video. - Audio Pass-through.


Since VEAI uses ffmpeg, and ffmpeg has the -c command to simply copy the audio… it really makes no sense to me why VEAI has to reprocess the audio.


I agree with meimeiriver in general. For one project I am doing a fair but of color grading so find adding the audio later is appropriate. BUT, it would be nice to have a “pass thru” of audio rather than down grading the quality. Given this, simple upscaling of Music Videos now have to have a round trip in FCP or FFMPEG.

1 Like

I have heard that more robust audio handling is coming to VEAI, but I’m not aware of the timeline.

I doubt that the “-c” directive would work in this case, since the encoding is being done after extensive video processing, and it would appear that VEAI makes the encoding call every 8 frames (i.e. not as a stream). Just too many opportunities for it all to go out of sync.

Honestly, this is a non issue. Download MKVToolNix and remuxing the original audio with the VEAI processed video is extremely simple. I even remux stereo audio in DVD rips since I also want to bring in subtitles and chapter marks.