Open source guru Randal Schwartz of FLOSS Weekly recently interviewed Georg Holzmann of Auphonic.com about their new online service for post-processing audio files for podcasts. Auphonic.com sounds very interesting indeed.
In the past the routine production of the VUC podcasts involved the use of The Levalator from The Conversations Network. That program, while a potent tool, is run locally and limited to processing uncompressed WAV files. This places a certain burden on the user to know how to create the appropriate source files, and later on to encode the processed files for publication online.
In contrast, Auphonic.com is at present a free web service. It can read and create numerous different audio file types, and leverage various codecs, including Opus. Then it integrates with various online tools for file submission & delivery. For example, a source file can be passed as an http link and the resulting output file delivered to Amazon S3, Dropbox, Libsyn, SoundCloud, YouTube or an FTP/SFTP server.
As a test I created a present intended to process a VUC recording. I enabled the various aspects of the processing as shown in the screen shot below. I then uploaded the raw wav file from the VUC December 28, 2012 where we talked with Chris Mathieu of Twelephone. This was a 160 MB uncompressed wave file so it took some time to upload, but it was the cleanest possible place to start.
(click on the image above for a larger view)
Where Auphonic shows real promise is in the way they have implemented some adaptive processing. The filtering and noise reduction functions learn as files are processed, improving the algorithms in use based upon its experience with user files. This is well beyond anything offered by The Levelator, which deals exclusively with program loudness. It’s an impressive bit of work.
When Auphonic completed processing my wav file and send the resulting MP3 to my SoundCloud account.[soundcloud url=”http://api.soundcloud.com/tracks/80681136″ params=”” width=” 100%” height=”83″ iframe=”true” /]
I was so happy with the result that I ran the process a second time. This time I didn’t need to upload the source file as it was already on the Auphonic server. I extended my preset to have the server create an Opus encoded file at 64 kbps, passing that file to my Dropbox account.
In a cursory listen I think that Auphonic is at least the equal of The Levelator with respect to automatic loudness processing. The additional filtering and noise reduction capability, and file format flexibility makes Auphonic very attractive. Their future work to add de-clipping is going to further extend its utility, which can only make it even better. I look forward to watching the project as it evolves.