Automatic Post Production offer

Post by **annise** » September 25th, 2012, 10:13 pm

email received

Hallo!

First I want to thank you (and of course all contributors) for LibriVox, it's really a great project and I am constantly listening to audio books from your network!

I am writing you beacuse we built a free automatic audio post production web service (http://auphonic.com), which should be especially interesting for contributors to your platform.

See e.g. the short intro video:
https://www.youtube.com/watch?v=Y6uXP-MGt8I

A short feature list:
- we classify and adjust volumes between different speakers, music or different regions (see https://auphonic.com/blog/2012/03/27/auphonic-audio-post-production-now-open-public-users/#algorithms)
- we normalize the overall volume of the audio to new, consistent loudness standards (see https://auphonic.com/blog/2012/08/02/loudness-measurement-and-normalization-ebu-r128-calm-act/)
- all compressor/limiters are done automatically as necessary
- we also do automatic noise reduction: we automatically detect where the background levels are changing and extract optimal noise prints (see https://auphonic.com/blog/2012/05/11/experimental-noise-reduction-and-algorithm-chooser/) - this should be especially interesting for your contributors, which don't have professional equipment
- encoding to various files (mp3, opus, vorbis, aac, flac, alac, wav, ...)
- setting metadata and chapter marks in all file formats
- we can automatically split files in chapters (as you have it on your webpage)
- automatic content deployment after post production: FTP/SFTP (e.g. to archive.org), soundcloud, youtube, dropbox, etc.

All this sounds a little bit like magic - the trick is that we always classify the different audio parts (music, speech, different background situations, etc.) as a sound engineer would do and classify which processing is necessary.
And because it is all done on our servers, our machine learning algorithms can also learn from the uploaded audio files.
We have already many users in the german speaking podcasting community,
but are still at the very beginning ...

I think your our service would shorten the workflow of many of your contributors and we would be glad if some of you want to use it - of course for free!
It's one way we can give back something to your great community ...
Also, if you have some additional feature requests or problems, we would be happy to implement them!

If you don't like to use it, no problem - I will continue to listen to your audio books

Thanks for your time and inspiration!
LG
Georg

Post by **annise** » September 25th, 2012, 10:16 pm

Several admins have tried it out and been impressed.

We would be interested in any comments from you if you try it out . They do say that their noise cleaning works better if the file is not too small.

Anne

YouTube · Post by **Darvinia** » September 28th, 2012, 8:25 pm

I tried it on a couple of files that I had already completed myself. I keep a prepostprocessing copy (if that makes any sense) and a post-processed copy of all my work. So I could compare what it did exactly against what I had done. It exceeded my efforts and did it in less than two minutes for a half hour file. It takes me about an hour to do it manually.

Two thumbs up!

Post by **annise** » September 28th, 2012, 8:48 pm

Great - that is just what we wanted , A pre and post file.

Anne
Did you have any problems , difficulties etc using it ?

YouTube · Post by **Darvinia** » September 28th, 2012, 9:09 pm

No problems at all! The interface is straight-forward and user-friendly.

catrose · Post by **catrose** » September 29th, 2012, 1:05 am

You know what? I am pretty impressed with this. I've done a test to show you all. The first one's the raw file, the second is after this thingy. It's completely got rid of my background noise (which is difficult for me, because my background noise varies throughout files, so I normally have to do multiple NCs) And it's sorted out my volume, which normally take me ages to fix, because I get quieter the longer I've been recording.

http://upload.librivox.org/share/uploads/xx/auphonictest_catrose_before.mp3
http://upload.librivox.org/share/uploads/xx/auphonictest_catrose_after.mp3

I can see this helping quite a bit with plays and editing, when it's important to keep the volume the same

Post by **annise** » September 29th, 2012, 2:53 am

The noise cleaning has worked well - and they did say it was even better with longer samples. It would be good to get a dramatic script through it - anyone got one ?

Anne

catrose · Post by **catrose** » September 29th, 2012, 2:58 am

Give me an hour and the first act of Skin Game will be done

The first half of Skin Game will be done. I was having issues with the volume of it, so I was going to have to put it through the levelator, but this should work nicely.

Post by **annise** » September 29th, 2012, 3:21 am

catrose · Post by **catrose** » September 29th, 2012, 3:35 am

Pretty impressed with this too. Here is the first minute, before and after:

http://upload.librivox.org/share/uploads/xx/auphonictest_drama_before.mp3

http://upload.librivox.org/share/uploads/xx/auphonictest_drama_after.mp3

This is the beginning part (the cast of characters). As you can here from the first one, everyone had different levels of volume and noise cleaning. This program made everything consistant

Post by **KiltedDragon** » September 29th, 2012, 6:36 am

Awesomeness!!! This could be a great tool for Dramatic Works projects.

auphonic · Post by **auphonic** » October 2nd, 2012, 1:34 am

Sorry for the late answer - I managed it now to finally join the forum

I am one of the auphonic developers ...

First thanks for all your libri vox contributions!

If you have any problems with our service, for example if it's confusing or you want some other/more audio algorithms or settings - don't hesitate to give us some feedback!
We are happy to expand our workflow for the specific needs of audio books!

If we don't manage to continue the discussion here in this forum, please contact us through another channel:
- mail: feedback@auphonic.com
- twitter: https://twitter.com/auphonic
- facebook: https://www.facebook.com/pages/auphonic/217115551651035
(I am sorry, usually I don't manage to stay up to date in forums)

Have fun and let us know what you think!
LG
Georg

Post by **KiltedDragon** » October 2nd, 2012, 3:27 am

I have used it several times and will continue to do so. I think it is working great!

auphonic · Post by **auphonic** » October 3rd, 2012, 12:59 am

Thanks, that's great!
Then we are happy

Have a good day!

Post by **TriciaG** » October 8th, 2012, 4:47 pm

Hmmm. Since the last time I used this, the volume options have been added. What level was the production set for before the options were added?

(Before, the volume outputted measured around 90 dB on MP3Gain or Checker. I tried the "-23 LUFS, EBU R128" setting this time and got 86.6 dB on my recording. Then I tried -15 LUFS and got 94.3 dB.)