Looking for a solution to create a video w/o re-encoding

[eddman]

Hello,

I'm looking for a free or paid software or any workaround that will help me output a video without re-encoding.

I'm planning to make a video by importing a silent video stream (.h264) and an audio stream (mp3).

The video stream isn't going to be edited at all, and only the audio stream will be edited.

Obviously there will be no point in re-encoding the video in this case, because it'll lower the quality.

Is there a way to achieve the above?

[poisondeathray]

There are many slightly different ways to do this.

The specific options available and software you would use depend on what types of audio edits you plan on doing, what the final format goal is (what container, how is it going to be viewed).

The general workflow is to do your audio edits, export audio only then multiplex with the video into a container using a muxer

[eddman]

There are many slightly different ways to do this.

The specific options available and software you would use depend on what types of audio edits you plan on doing, what the final format goal is (what container, how is it going to be viewed).

The general workflow is to do your audio edits, export audio only then multiplex with the video into a container using a muxer

The problem is, I cannot edit the audio independently. Let's see if I can explain what I'm trying to do.

Source/input streams: a video stream (.h264), first audio stream (english, will NOT be edited), second audio stream (armenian, will be edited, some parts will be cut).

The video and first audio stream come from an MKV file. The second audio is an MP3 file.

Desired output: MKV

I have to edit the second audio stream within a non-linear editor, so that I can match the parts that I want to keep with the correct scenes in the video.

This image might help explaining it:

[poisondeathray]

The problem is, I cannot edit the audio independently. Let's see if I can explain what I'm trying to do.

Source/input streams: a video stream (.h264), first audio stream (english, will NOT be edited), second audio stream (armenian, will be edited, some parts will be cut).

The video and first audio stream come from an MKV file. The second audio is an MP3 file.

Desired output: MKV

I have to edit the second audio stream within a non-linear editor, so that I can match the parts that I want to keep with the correct scenes in the video.

This image might help explaining it:

It's not a problem. This is what video editors like NLE's do You are only using the video as a visual "guide"

Most NLE's won't accept MKV as a container, and many have problems with a raw AVC stream, so you would need to remux to MP4 container or other suitable container, if the video is compatible (not all AVC streams are compatible with all NLE's)

Or if it's just simple cuts, simple audio edits such as in your diagram, you can use the freeware aviutl, which does support MKV and many different types of AVC natively

[eddman]

It's not a problem. This is what video editors like NLE's do You are only using the video as a visual "guide"

Yes, that's what I need to do; use the video as a reference.

Most NLE's won't accept MKV as a container, and many have problems with a raw AVC stream, so you would need to remux to MP4 container or other suitable container, if the video is compatible (not all AVC streams are compatible with all NLE's)

Hmm, I suppose I could remux to MP4; not an issue. Do you know any NLEs with a pass-through output option a.k.a no re-encoding? I know, for example, QuickTime has it, but of course it's not an editor.

Or if it's just simple cuts, simple audio edits such as in your diagram, you can use the freeware aviutl, which does support MKV and many different types of AVC natively

Yes, mainly simple cuts, although I'd need to adjust the volume too. The second audio is very loud compared to the first.

Thanks for the recommendation. AviUtl seems to be the right tool, although I'm kind of overwhelmed by it.

Am I doing this right? First, I import the video file (which also contains the first audio) through "File -> Open"; then I right-click on layer 1 and add the second audio through "New media object -> Audio file".

or should I extract the audio from the MKV and use layers for both audio streams?

Let's see if I could figure out the cutting process.

Suppose I finished the editing, how do I save/export the end result?

[poisondeathray]

It's probably easier in aviutl to use the advanced editing timeline (filter+ => checkmark advanced editing ) . You can drop and drag tracks onto the timeline. The layer order is reversed compared to most programs (layer 1 is on the "bottom"), not that it matters for your specific project, it would only matter if you had multiple layers for video

You can drop and drag the mkv onto the track 1, and it will open up both audio & video. Then drop the 2nd audio track onto the next layer. Look at the aviutl tutorial thread for basic usage, there are video tutorials by racer-x and maverick tse. I don't think you can passthru video (stream copy, but it doesn't matter for your project anyways), but to save audio you use file=>export as wav. If you want to compress it farther use another program like taudioconverter, or lamexp. Use Mkvtoolnix to combine your original video, new audio track.

The GUI can be a bit difficult to use for some people - if it's not your type of thing, there are other professional NLE's like premiere pro, sony vegas etc.. that have a lot more control over audio functionality, and the GUI's are more intuitive and easier to use, but not free

[eddman]

The GUI can be a bit difficult to use for some people - if it's not your type of thing, there are other professional NLE's like premiere pro, sony vegas etc.. that have a lot more control over audio functionality, and the GUI's are more intuitive and easier to use, but not free

Just wanted to thank you again. Now that I got the hang of AviUtl, it's quite easy to use. Too bad there is no 24-bit or multichannel output for WAV or W64.

I suppose I'd have to use something like Vegas for those cases.

[poisondeathray]

The GUI can be a bit difficult to use for some people - if it's not your type of thing, there are other professional NLE's like premiere pro, sony vegas etc.. that have a lot more control over audio functionality, and the GUI's are more intuitive and easier to use, but not free

Just wanted to thank you again. Now that I got the hang of AviUtl, it's quite easy to use. Too bad there is no 24-bit or multichannel output for WAV or W64.

I suppose I'd have to use something like Vegas for those cases.

It supports multichannel up to 7.1, but only 16bit. Push the "audio encoding" button in the export dialog box for more options. There will be drop down menu with more options

[eddman]

It supports multichannel up to 7.1, but only 16bit. Push the "audio encoding" button in the export dialog box for more options. There will be drop down menu with more options

I don't see it. I opened an MKV with a 6-channel (5.1) DTS audio, but when I want to export to PCM WAV, there are only mono and stereo options.

[poisondeathray]

6ch option shows up here, if you click lower as in your screenshot under the "attributes"

I can't remember but I don't think there are any hidden options in the preferences

Are you using L-smash to open the file? If you were using directshow and had a stereo downmix in your directshow chain that might explain it

[eddman]

6ch option shows up here, if you click lower as in your screenshot under the "attributes"

I can't remember but I don't think there are any hidden options in the preferences

Are you using L-smash to open the file? If you were using directshow and had a stereo downmix in your directshow chain that might explain it

How do I do that, exactly? I put it at the top of the priority list but apparently it didn't work.

I guess it doesn't matter after all. I looked around a bit and found an article that says Aviutl cannot output multichannel WAVs form the NLE at all.

[poisondeathray]

6ch option shows up here, if you click lower as in your screenshot under the "attributes"

I can't remember but I don't think there are any hidden options in the preferences

Are you using L-smash to open the file? If you were using directshow and had a stereo downmix in your directshow chain that might explain it

How do I do that, exactly? I put it at the top of the priority list but apparently it didn't work.

Yes, just leave it at the top, or at least above directshow

I guess it doesn't matter after all. I looked around a bit and found an article that says Aviutl cannot output multichannel WAVs form the NLE at all.

Can't? Really? Maybe that article is old ?

Tested and works here

Opened the wav in an audio editor and confirmed all 6 channel distinct waveforms are there

[poisondeathray]

The first one I tested was a MKV with 5.1 AC3, but the 2nd was 5.1 DTS, both worked fine

You should go ask in the AVIUtl support thread. I can't find any hidden buttons or switches in the preferences or options that might have enabled it

[poisondeathray]

Close AviUtl and re-open it.

If you have AVIUtl open, but no video or audio or project open, select filter+ => compresson => audio and that should open up the same "sound selection" dialog box as in your screenshot. Can you select 6ch from the attributes drop down box ? This will eliminate if for some reason it was downmixing your source , since there is no source

[eddman]

Close AviUtl and re-open it.

If you have AVIUtl open, but no video or audio or project open, select filter+ => compresson => audio and that should open up the same "sound selection" dialog box as in your screenshot. Can you select 6ch from the attributes drop down box ? This will eliminate if for some reason it was downmixing your source , since there is no source

Just checked. No file opened, still no multichannel. What's going on?!

Are you using a different version? I downloaded it from here. This is the file I got: AviUtl_ExtraPack1-10[Win7]Release.7z

EDIT: This is the article I mentioned. It's written by MaverickTse, AviUtl's translator.

http://mavericktse.mooo.com/wordpress/archives/1635

[poisondeathray]

I don't fully understand what he's doing there, I think he's talking about AAC, then converting it to WAV. Because the AAC metadata has 2ch instead of 6 - some programs get confused. This doesn't affect you if you export WAV (not AAC)

It might be some of your audio filters are downmixing the audio. I just tested import / export and 6ch definitely works . I tested a couple older aviutl versions, but not the newest one. Try an older version and if that doesn't work, ask in the AVIUtl support thread, maybe Maverick knows what is going on

[MaverickTse]

As I have already mentioned in the AviUtl support thread, if you want sophisticated audio editing, do it in an audio editor.
The info in my blogpost #1635 still holds true : ANY AUDIO LOAD THROUGH Adv.Editing will get DOWNMIXED to 2ch.

There is another tricky place to look for:
Check if the channel layout setting in L-Smash Works(File>Settings>Import plugins> L-Smash Works) is DL+DR.
I have set this in EPv1.10 by default because otherwise some user will say they hear nothing.
You can remove this string if you need to encode Multi-channel audio.

Nevertheless, since you now want to edit muti-channel audio data, and I see that you have a few more post that concerns primarily on audio, I STRONGLY suggest you demux and extract the audio track(s) and edit it in Audacity or some dedicated audio editor of your choice, then remux with the video.

[eddman]

As I have already mentioned in the AviUtl support thread, if you want sophisticated audio editing, do it in an audio editor.
The info in my blogpost #1635 still holds true : ANY AUDIO LOAD THROUGH Adv.Editing will get DOWNMIXED to 2ch.

There is another tricky place to look for:
Check if the channel layout setting in L-Smash Works(File>Settings>Import plugins> L-Smash Works) is DL+DR.
I have set this in EPv1.10 by default because otherwise some user will say they hear nothing.
You can remove this string if you need to encode Multi-channel audio.

Nevertheless, since you now want to edit muti-channel audio data, and I see that you have a few more post that concerns primarily on audio, I STRONGLY suggest you demux and extract the audio track(s) and edit it in Audacity or some dedicated audio editor of your choice, then remux with the video.

I can't because I need the video as a reference. I need to do a lot of cutting on the audio streams.

I knew that AviUtl can't do multichannel, it's just that poisondeathray confused me. I suppose he's using an older version. Thanks for clarifying the situation.

I don't really want to edit multichannel in this particular case. I was just wishing Aviutl had that ability. I'll probably try vegas for those cases.