Mini Guide: Normalizing Audio for Multiple AVI files

[Nologic]

Ahh I see what the problem is...in the _Main function I don't account for empty folders...nor do I store a count for all the files found under each folder listed in the GUI.

Good catch.

[fonzo]

I don't have all the info in front of me, but I wanted to give you the gist of something else I found. Xvid/Mp3, Avi file is "MPEG-4 (XVID), MPEG Layer 3 (mp3). While processing, the video extract ends up being an .mp4 file and is MPEG-4 BASIC, which loses the XVID part. I'm not sure what the technical term for that is, maybe the fourcc code or something. Using the old style batch file process, you get an .avi file for the video stream which retains the XVID part. That is using the ffmpeg with command line options you supplied a while back in this thread.
So, the merge fails, but even if it did work correctly, you'd lose the XVID designation and it would not work in DVD players.
I hope this helps you to track down that issue. I'm not sure if that or the FLAC issue has anything to do with the merge fail though.

[Nologic]

I actually held the FourCC tag "Xvid" in a variable which would upon merge add it back...I produced MP4 temp containers since all the tools I use can read from MP4...and a few don't like AVI containers...I know it's puzzling but true.

So the MP4 Temp containers were to work around another problem...and the FourCC tag was still being preserved...well it was for AVI containers...I discarded it for MKV & MP4 final containers...since it made little difference to those...far as I'm aware of.

The Merge failing has more to do with me using FFMpeg switches that don't work with AVI...to preserve stream offsets...since AVI containers do not contain a method to keep track of general time. The switches work fine with MP4 & MKV and probably a few other containers...but absolutely fail with AVI.

FLAC doesn't matter since the file gets encoded to MP3 prior to merge...or at least should be...if an MP3 stream isn't created (it may be held in a MP4 container)...and there is only the FLAC stream...a merge will fail since AVI from my understanding doesn't like FLAC...actually I think the only one that does is MKV...but I could be wrong here...not a guru by any stretch of the imagination.

Anyways right now when possible I'm trying to get use to Git...once that is done I'll start addressing issues as best I can.

Sorry for the delay...just things have been a cluster **** around the household as of late.

[fonzo]

Ok,thanx for the update. I don't know if you need to fix audio/vid sync problems in avi files in your prog for any reason, but I know there is a program called winx hd video converter which is a front end for mencoder. It has an option for syncing offset audio and video streams so I'd assume mencoder has a way to do that in case you need a way to do it and ffmpeg doesn't work. Just a thought.

Sorry to hear about the ruckus. I wish the neighbor girl would come over there and help you out.

[Nologic]

New build

v1.46 Beta 6
==================================
Switched back to FFProbe from MediaInfo
Buffed Up Progress Dialog (for now)
Cosmetic Fixes
MKV Stream Track ID Fixes
Prune Already Completed Files
Added Variable To Count All Files Processed
Corrected Stream Delay Milliseconds
Added Progress Bars to nearly everything
Added Lots of Error Reports.
Resolved Audio Delay issues with MKV & MP4 and maybe others...AVI still hosed.
Resolved the basic MetaData issues with FFMpeg...AVI still doesn't hold any tags, but isn't failing anymore.
Resolved all Known Progress Bar Issues
Stopped Nuking External Subtitles, when Direct Replace wasn't selected.
Updated FFMpeg
Updated MKVToolsNix
Updated x264
Removed Subtitle Workshop using FFMpeg instead (maybe regressions)

Only known issues...well outside of a lack of features that I wanted to complete....are Adjusting the Aspect Ratio can sometimes make things go very wrong...and Audio Delay in AVI containers still isn't resolved. I'll probably do another update soon, that will skip files where a Audio Delay is detected...and the final container will result in a AVI...till such time as I find a reasonable method to resolve this issue.

[Belini]

@Nologic thank you for continuing the work and for sharing with us.

[Nologic]

You are very welcome, and this should be a rather stable release...out side of the issues mentioned above. Which hopefully I'll get resolved shortly.

Hoping to add command line support, hard subtitles...and maybe a scheduler in the next few releases.

[Nologic]

Growl...seems I just can't have a release go right.

Anyways here is a quick update that resolves an issue where I managed to delete a variable that was...well needed.

It dealt with SubTitles and would case a crash of the app...so no harm to any files...just a lot of wasted time is all.

[fonzo]

Using Levelator on xvid/mp3 avi file.

Autoit error
Line 11749
Error: Variable used without being declared.

[Nologic]

v1.46 Beta 8
==================================
Resolved Bug reported by Fonzo, issue with a var not getting renamed
Added Crude Command Line Interface, simply pass the executable a proper INI file.

.\VideoGain.exe "<SomePathSomeWhere>\<NameOfSettings>.ini"

Added CloseVideoGain.exe to be used to do cleanup duty if VideoGain via CLI needs to be closed early.

Keep in mind when VideoGain is invoked by the command line...it's hidden.

[Nologic]

v1.46 Beta 9
==================================
Added AVI Early Out if Stream Delay Detected
Added Ability to Que Multiple Instances via CLI
Added Sending Empty Path via CLI runs default INI

.\VideoGain.exe ""

Added CLI Secondary Option "CloseOthers" to Close All Running Instances Except This one

.\VideoGain.exe "<SomePathSomeWhere>\<NameOfSettings>.ini" "CloseOthers"

Added check for blank destination path when Alt destination is selected

[Nologic]

New Build

v1.46 Beta 10
==================================
Added Scheduler
Moved Shutdown option to Scheduler Tab
Added Exit option to Scheduler Tab (Just Closes App With Silent Logfile)
Updated FFMpeg
Updated MKVToolsNix
Updated QAAC

Okay the scheduler isn't a hard line...it should allow the current stream action to complete before halting everything...however progressbars will likely be effected in a negative way during the halt.

I am tossing around the idea of only halting the process at the start of a file...rather than anywhere along the line. The only difference is how soon the CPU will be freed. Right now how it is...you are more likely to get it free sooner.

Anyways give me your thoughts on the subject.

[Nologic]

v1.46 Beta 11
==================================
Fixed redraw issue with Scheduler
Updated ReadMe.txt to be current

I forgot to mention that because of the new scheduler INI's must be recreated, so please delete your current or rename it, if there are settings you wish to copy out of it to apply to the new one that will be created.

[Belini]

I tried to normalize downloaded some videos on youtube and only worked in previous versions to v1.42, download this video for testing ==> http://www.youtube.com/watch?v=CMot9oCm8rY

[Nologic]

Odd seems to turn out fine for me...I grabbed the 720p MP4 version of the video you linked to and all was fine. Hehe that was until I enabled AutoCrop...to remove the bars...then things went sideways...as I ended up with super skinny people.

So I do have to look at the AutoCrop code...it removed the bars...but stretched the picture to fill in those bars. Not the desired effect.

Anyways my testing was with ReplayGain, ReplayGain (Tag), ReplayGain (Levelator) all turned out as expected...in this case lowering the gain of the audio.

What settings are you using that are resulting in unexpected results?

P.S. Lovely song even if I can't understand a word of it.

[Belini]

I used the default setting of the program

[spoiler][/spoiler]

After nearly five minutes gives this error:

[spoiler][/spoiler]

[Nologic]

Okay if you still have a copy of v1.46 Beta 9, take it's build of QAAC from it's Bin folder and replace the current QAAC build and see if that resolves the problem.

You will want to replace the whole QAAC folder, as the Quicktime support files were also updated along with QAAC.

Let me know how that turns out.

[Belini]

I replaced all BIN folder and it worked more took 4 minutes and 18 seconds to create the standard file.

[Nologic]

If you would do some further testing...use the latest, but with an older QAAC build...then try the latest, but this time with an older FFmpeg build...the reason for this is I'm not sure if its QAAC to blame or FFmpeg...there seem to be reports that newer builds of FFmpeg are failing on WinXP...but by the same token...the prior version of QAAC was accidentally built to support Vista+...however that was corrected in a point release...but may have happened again.

Just sucks when things break...course it's nicer when it's not my code that is broken.

[Nologic]

v1.46 Beta 12
==================================
Added Vorbis & Opus Audio Encodes (Opus Currently Broken)
Added Ogg Container
?Fixed? AutoCrop & Scaling
Fixed Display Issue when Changing Audio Encoders

There maybe issues with WinXP, if so use an older build of FFmpeg &/or QAAC

[fonzo]

NoLogic, would you mind providing the source script files with your releases? It would be greatly appreciated.
BTW, if I use your program and use levelator(replaygain) or levelator(ebu r128) and end up with a file that's just slightly too low in volume(for my uses), what value would I have to change in the source file to end up with a slightly louder output file?
Of replaygain and ebu128, which ends up with a louder volume level and by how much?
Thanks.

[Nologic]

ReplayGain is loader than EBU R128.

As for releasing source for the sole reason of making audio streams non compliant with RG or R128...no.

If it was for bug hunting...maybe

if it's for adding a feature I personally haven't a clue how to create...more than likely

If its to add a feature I personally don't want to add to my build...then maybe.

I released source for some time...and only Belini made mod's and offered source...to which I still from time to time give him code drops...ether without him asking or upon request.

If you want to tweak to your hearts content I'd suggest toying with R128Gain...it'll let you do far more than what my app does...well when discussing Audio Gain tweaking.

[Nologic]

v1.46 Beta 13
==================================
Fixed Fast Seek in AutoCrop Testing
Added Floating Point FPS to Fraction Function
Refined Duration
Added Frame Count for MKV

[csibo]

hi,

is it possible to download the none beta version?

VideoGain v1.45.7z

[csibo]

i would like to do batch volume normalization without video recompressing.
unfortunately i lost the avi files aspect ratio settings.
is it possible to find and set the video aspect ratio automatically?

[kolpotoru]

I'm releasing a new version of AVI Gain...I may have already done a v0.95 release in the past...but if so..oh well as I'm releasing this as v0.95

So whats this build all about:
This is only for AVI's and will only create AVI's
All audio is encoded to MP3 even if it was already. The reason is its not using Replay Gain tags...but adjusting the volume like Video Gain.
Now unpacks Mpeg4 video streams, as requested.
Now tries to insure Encoder FourCC tag is used rather than general FourCC tag
Now tries to insure proper FPS in video (actually broken in FFmpeg)
Now uses random string names for exported streams to help prevent File system locking do to heavy fragmentation of the MFT.

This doesn't handle Subtitles, or encoding of video streams.

I want to download Latest version of Avi Gain, which will give avi output please share the download Link.

[Nologic]

hi,

is it possible to download the none beta version?

More or less all versions are beta...the last batch was mainly tagged as beta because I had planned to add a bunch of features...but have yet to finish adding all of them.

i would like to do batch volume normalization without video recompressing.
unfortunately i lost the avi files aspect ratio settings.
is it possible to find and set the video aspect ratio automatically?

The current version supports setting the aspect ratio...but it only sets is to one value...so if you have different aspects...you'll have to manually separate them into different folders to batch process. This doesn't depend on re-encoding the video...it's simply setting a meta tag in the container file (avi/mp4/mkv/...).

This can be done with FFMpeg if you like:

ffmpeg.exe -i "C:\InFile.avi" -c:v copy -aspect 1.33 -c:a copy -c:s copy "C:\OutFile.avi"

That should copy all the streams over and set the aspect ratio to 4:3....but that won't adjust the audio.

I want to download Latest version of Avi Gain, which will give avi output please share the download Link.

The latest VideoGain versions should output to AVI, if you make sure to select AVI as the container...and make sure to only use MP3 for the audio stream.

[Belini]

@Nologic Have more than 10.000 files, most of them are normalized to 89dB's and has Some that are not standardized, which program on the command line should I use to know which files are 90db's or more and which files are 88dB's or less?

EDIT: I tried to use more ffmpeg.exe did not return the volume in decibel.

$File = @ScriptDir & "\4285.mp4"
$vol = Run(@ComSpec & ' -i ffmpeg.exe -af volumedetect -f null -"' & $File & '"' , @ScriptDir, @SW_HIDE)
MsgBox(4096, "Test 1", $vol)

$File = @ScriptDir & "\4285.mp4"
$vol = Run(@ComSpec & ' -i ffmpeg.exe  -af volumedetect -f null"' & $File & '"' , @ScriptDir, @SW_HIDE)
MsgBox(4096, "Test 2", $vol)

$File = @ScriptDir & "\4285.mp4"
$vol = Run(@ComSpec & ' -i ffmpeg.exe  -af volumedetect -f"' & $File & '"' , @ScriptDir, @SW_HIDE)
MsgBox(4096, "Test 3", $vol)

$File = @ScriptDir & "\4285.mp4"
$vol = Run(@ComSpec & ' -i ffmpeg.exe -af volumedetect "' & $File & '"' , @ScriptDir, @SW_HIDE)
MsgBox(4096, "Test 4", $vol)

Each time you perform returns a different value

[ConverterCrazy]

Hello,

This is about Normalizing audio volume in video files. I have a few observations and questions. I'll be grateful for any suggestion for the following:

- Does NORMALIze mean 89dB level, or, uniform audio level in all files? (Will the audio sound the same in all files, if set to 89dB?)

Note: I find that 89dB doesn't sound the same in all files.

To achieve same perceptible audio volume in all files, I did the following:

- Imported audio in WavePad, a Trialware, applied AGC (Automatic Gain Control), Normalize and Save AS .AC3 file with 224kB bit rate.

There was no perceptible difference in audio volume unlike the difference I noticed using MP3Gain, Audacity, etc.
Since, my preferred video file format is MKV, I used MKVMerge to merge the new file.

However, https://www.videohelp.com/tools/WavePad is not free, is kinda buggy software and the antivirus gives many warnings when installing it. WavePad is also very-very slow on my system. Hence, I don't want to use WavePad.

And,

I don't find the AGC function of WavePad in any other audio editor.

What's the solution?

Thanks

[Nologic]

@ Belini -

The code to do this should be around line 2034.

Basically looking at:

ffmpeg.exe -i ".\4285.mp4" -filter_complex ebur128 -f null - 2>4285.txt

You'll end up with a text file with something like this at the end:

frame=31610 fps=1013 q=0.0 Lsize=N/A time=00:21:58.42 bitrate=N/A    

video:1976kB audio:247204kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
[Parsed_ebur128_0 @ 02532420] Summary:

  Integrated loudness:
    I:         -18.1 LUFS
    Threshold: -29.0 LUFS

  Loudness range:
    LRA:        10.4 LU
    Threshold: -39.1 LUFS
    LRA low:   -25.1 LUFS
    LRA high:  -14.7 LUFS

You are looking for:

  Integrated loudness:
    I:         -18.1 LUFS

Which is fairly good, since we are looking for about -18 to match up roughly with 89.9db, so in this case nothing would really be changed...since rarely will we get exactly 89.9db...do to possible clipping.

However lets try a different one...with say the below value:

frame= 1464 fps=0.0 q=0.0 Lsize=N/A time=00:01:01.18 bitrate=N/A    

video:92kB audio:11444kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
[Parsed_ebur128_0 @ 02c07480] Summary:

  Integrated loudness:
    I:         -24.4 LUFS
    Threshold: -36.8 LUFS

  Loudness range:
    LRA:        18.8 LU
    Threshold: -46.8 LUFS
    LRA low:   -42.6 LUFS
    LRA high:  -23.9 LUFS

In the above we have -24.4...but we want -18, so we subtract:
-18 - -24.4

Which results in us needing to raise the volume by 6.4:

ffmpeg.exe -i "C:\SomeFile.mp4" -c:v copy -c:s copy -c:a libmp3lame -q:a 8 -af "volume=6.4dB" -f mp4 "C:\SomeFile (new).mp4"

Basically if they come in between -17.5 to -18.5 I'd leave them be.

@ ConverterCrazy -

Normalize try's to make every sound have the same level of gain....typically with total disregard to clipping.

Generally you want to use either r128 or ReplayGain...this allows different sounds to have different levels of gain...but tries to ensure that that average gain is set to a certain level...adjusting each sound to bring the over all average to that level, while avoiding clipping..which may lower the average gain (db) to prevent such errors from occurring.

Some folks like normalize...I'm not one of them...I personally prefer a soft babbling brook to have a different gain than an explosion later in the movie...but for each their own.

So from what I take away from your post...that when using my app...you should probably set the Gain method to something like "ReplayGain (Levelator)"