VideoHelp Forum

+ Reply to Thread
Results 1 to 20 of 20
Thread
  1. Is there a way to normalize AC3 audio without re-encoding? I know this is possible with MP3. I'm curious because I'm trying to make a DVD with varying clips but each video has a different volume. I'd like them all to have the same volume without re-encoding, if possible.
    Quote Quote  
  2. AC3 Normalizer was the closest thing I could find to what I wanted so far but it re-encodes the audio. The huge deal breaker is that the results produced major buzzing noises and crackling.
    Quote Quote  
  3. I'm a Super Moderator johns0's Avatar
    Join Date
    Jun 2002
    Location
    canada
    Search Comp PM
    I use audacity with ffmeg plugin to normalize ac3,never get any buzzing or crackling that wasn't there.It re-encodes but does a good job.
    I think,therefore i am a hamster.
    Quote Quote  
  4. Member
    Join Date
    Mar 2008
    Location
    United States
    Search Comp PM
    I rememmber playing with this a year or two back, I used Audacity to WAV, then the Aften Gui back to ac3.

    It allows you to add Dialnorm and/or set optional dynamic compression
    Last edited by davexnet; 27th Jul 2022 at 17:59.
    Quote Quote  
  5. If there is a way to losslessly do this with MP2 maybe I'll consider that.
    Quote Quote  
  6. https://forum.audacityteam.org/viewtopic.php?t=104273
    DVD Doug - If you don't know this already - Normalizing files doesn't necessarily make them all sound equally-loud.
    Is there a program that can do this or is this something that has to be done manually?

    Also I tried this and it did it's job. https://forum.videohelp.com/threads/258002-Mini-How-To-Increase-Your-AC3-Volume
    Quote Quote  
  7. I'm a Super Moderator johns0's Avatar
    Join Date
    Jun 2002
    Location
    canada
    Search Comp PM
    Unless you have the best audio system in the world you won't hear any difference when you properly encode ac3 with the same bitrate or higher.Also audacity will save the file as ac3 so no need for wav and then ac3 encoding.
    I think,therefore i am a hamster.
    Quote Quote  
  8. Originally Posted by johns0 View Post
    Unless you have the best audio system in the world you won't hear any difference when you properly encode ac3 with the same bitrate or higher.Also audacity will save the file as ac3 so no need for wav and then ac3 encoding.
    I encoded one 192kbps ffmpeg AC3 file. Then I transcoded that one using the same settings and compared the two in a blind test. I'm no audiophile but I could tell the difference between the two. I don't have any set up either I'm just using airpods. I know ffmpeg isn't a good codec so I'll try one with aften too.

    In the grand scheme of things it doesn't matter a ton, but I want to make a DVD with music so I'm trying to avoid re-encoding a bit.


    foo_abx 2.0.6d report
    foobar2000 v1.6.7
    2022-07-27 20:53:57

    File A: 192.wav
    SHA1: d42410b08e49be8063744f3d08dbfca3ae455a90
    File B: 192 transcode.wav
    SHA1: 77ed3584f1cf96e9821c2067c5282aad6e1ad1e9

    Output:
    Default : Primary Sound Driver
    Crossfading: NO

    20:53:57 : Test started.
    20:56:20 : 01/01
    20:58:53 : 02/02
    20:59:12 : 03/03
    21:00:19 : 04/04
    21:00:32 : 05/05
    21:00:44 : 06/06
    21:00:55 : 07/07
    21:01:15 : 08/08
    21:01:29 : 09/09
    21:01:41 : 10/10
    21:01:41 : Test finished.

    ----------
    Total: 10/10
    p-value: 0.001 (0.1%)

    -- signature --
    40d80805ef6e8eddd02269922a1f972c7209354d
    Quote Quote  
  9. Here's 256kbps aften. This test was way harder and I took breaks.


    foo_abx 2.0.6d report
    foobar2000 v1.6.7
    2022-07-27 21:35:01

    File A: 256.wav
    SHA1: df919e9e3a15b561f0ae57a736626a03889b4a38
    File B: 256 transcode.wav
    SHA1: 66e71008842b2cec0a91a37d2f2838d59db3bdc2

    Output:
    Default : Primary Sound Driver
    Crossfading: NO

    21:35:01 : Test started.
    21:35:46 : 01/01
    21:45:09 : 02/02
    21:46:10 : 02/03
    21:47:38 : 03/04
    21:48:11 : 04/05
    21:49:05 : 05/06
    21:51:16 : 06/07
    21:51:51 : 07/08
    21:52:08 : 08/09
    21:52:25 : 09/10
    21:52:25 : Test finished.

    ----------
    Total: 9/10
    p-value: 0.0107 (1.07%)

    -- signature --
    cbd14b52f93a5ba85136e4fe5789932a64ac8e83
    Quote Quote  
  10. Member
    Join Date
    Mar 2008
    Location
    United States
    Search Comp PM
    Never used that tool. Are you saying the Aften encode sounded closer to the original than the Audacity encode?
    If that's so, use Aften.
    You have to re-encode regardless, there doesn't seem to be any other way.
    Even changing or nullifying Dialnorm requires a re-encode
    Quote Quote  
  11. I'm a Super Moderator johns0's Avatar
    Join Date
    Jun 2002
    Location
    canada
    Search Comp PM
    Not sure what settings you used but i did a test of one ac3 and re-encoded to the same bitrate with audacty and no other settings changed and it sounded exactly the same through my high quality headphones.
    I think,therefore i am a hamster.
    Quote Quote  
  12. For ffmpeg
    -ab 192k

    For aften
    -b 256
    Quote Quote  
  13. Member
    Join Date
    Mar 2008
    Location
    United States
    Search Comp PM
    Originally Posted by shampistols69 View Post
    Still involves re-encoding. If you have a point to make, speak up
    Quote Quote  
  14. I think that is the best and easiest way to do it but involves re-encoding.
    Quote Quote  
  15. I'm a Super Moderator johns0's Avatar
    Join Date
    Jun 2002
    Location
    canada
    Search Comp PM
    When you said 'I know ffmpeg isn't a good codec ' just proves you dont know anything about encoding.
    I think,therefore i am a hamster.
    Quote Quote  
  16. Member Skiller's Avatar
    Join Date
    Oct 2013
    Location
    Germany
    Search PM
    There is ChangeDN which changes the DialNorm value present in any AC3. This works without re-encoding but is not exactly normalization in the usual way. And if the value of DialNorm is already -31 db, it cannot be made any louder with this.

    Note that this should not be done to 5.1 mixes that may use DRC during playback.
    Quote Quote  
  17. shampistols69,
    ffmpeg's ac3 encoder is Aften. It was incorporated into ffmpeg a long time ago, I think after the Aften author stopped developing it.

    A sample Aften command line for foobar2000:

    -readtoeof 1 -pad 0 -v 0 -b 192 - %d

    ffmpeg:

    -i - -ignore_length true -c:a ac3 -b:a 192k %d

    Are you aware foobar2000 has a ReplayGain scanner, so you can scan the files, save the info to tags, then apply the ReplayGain info to adjust each file's volume while converting?
    Quote Quote  
  18. I am aware of foobar's tag scanner but the audio is for DVDs.
    Quote Quote  
  19. The ReplayGain scanner is built in. I'm still using an older foobar2000 (running on XP) but it looks like this:

    Image
    [Attachment 66335 - Click to enlarge]


    Once you've scanned your files and saved the ReplayGain info to tags, you can adjust MP3 and AAC audio losslessly (Apply track ReplayGain to file content), and for any other audio you can adjust the volume using the ReplayGain info while converting. You can configure that under the Processing section in the converter setup.

    It's getting a bit old now and I should replace it with the latest version at some stage, but there's a portable version of foobar2000 here you can play with if you're interested. https://forum.videohelp.com/threads/396860-foobar2000-portable-(for-audio-encoding)
    You can't see the Volume tab's contents in the screenshot in that post, and I think I've changed what it displays a little since then, but once you've scanned your files and saved the info, the Volume tab looks something like this:

    Image
    [Attachment 66336 - Click to enlarge]


    -18 dB or -18 LUFS is the ReplayGain target volume, hence most of the files in the screenshot being close to that. TrackGain is the volume relative to the ReplayGain target volume, which is fixed in stone. You don't have to adjust to the ReplayGain target volume, but TrackGain is always relative to it.

    In preferences, the default ReplayGain target volume is probably still shown as 89dB, because it was originally based on a SMPTE standard... but to cut a long story short, ReplayGain's 89dB target volume is the equivalent of -18dB on an output meter. So if you lower it to 86dB it's the same as -21dB on an output meter etc. For most music files, a volume of 89dB (or -18dB) has enough headroom to prevent peaks being clipped. For TV/movie soundtracks, which can be more dynamic, the industry standard volume is about 6dB lower.

    PS After you've scanned files and saved the info, by default foobar2000 will use it to adjust the volume on playback so the scanned files always play at the same volume, which is what ReplayGain was originally designed for, so to hear the volume difference between scanned files the ReplayGain volume adjustment on playback needs to be disabled in preferences.

    Image
    [Attachment 66337 - Click to enlarge]
    Last edited by hello_hello; 14th Aug 2022 at 15:11.
    Quote Quote  



Similar Threads