Compressing audio for mahara

Re: Compressing audio for mahara

by Mark Penrice -
Number of replies: 0

To build on what Bob said, Lucy, if the linked guides haven't already solved it for you - how did you record your file in the first place, and how long is it in terms of minutes and seconds?

If it was e.g. using a voice recorder app on your phone, or a more up to date digital dictaphone, it's possible that it's already quite well compressed (if it literally IS already an MP3, this is even more likely), at a space-saving "telephone" quality level that would be no good for music recording but is perfectly acceptable for vocal content.

As far as I'm aware, the relatively simplistic MP3 conversion function in Audacity defaults to near-CD-quality stereo, with file sizes to match, unless you specifically go into its options and change this, and it doesn't perform any "sanity checks" or warnings to prevent you doing inadvisable things.

So, for example, you might have a 10m20s voice recording at a relatively compact and slightly oddball (but not unheard of) 18kbit quality setting, which if it's MP3 would sound about the same as a mobile phone call with a good signal (somewhat worse with a more primitive recorder and format, better with something like AAC as used in iTunes and various phones), taking up 1.46mb of disc space ... which when imported into Audacity and then saved out again using the defaults of its MP3 converter would become a 10.4mb monster at 128kbit quality ... even though it wouldn't actually sound any better, as it's the equivalent of recording an old wax phonograph cylinder onto CD - it merely preserves a very accurate impression of how the original sounded!

Guides which cover converting your audio to work in presentations or for submission can sometimes be a bit out of date, and/or have to cover a complex range of possible "inputs" with a very wide brush... ancient dictaphones using poorly compatible formats, portable audio recorders that record at raw CD quality (10mb/minute), etc, in which case converting to 128kbit MP3 is a definite improvement... but it means they have to make some assumptions that may now be incorrect, e.g. "no-one records direct to low rate MP3". It might just be you're the one person who's come along with her file already in optimal condition and has been sabotaged by reading it! :-)

Certainly, unless it's some rather esoteric file type that only your own computer can read and convert, I'd say your original is good to upload as-is. 1.5mb is barely larger than the capacity of an old floppy disc, and an almost insignificant amount of data these days. You're likely to have downloaded more through your internet cable just getting to and reading this page.


Hmm, just skimming back over that, a different possibility has come to mind - if your recording is instead quite short, say on the order of about one to three minutes... it's entirely possible that it was originally near-CD-quality MP3, and an error within Audacity (malfunctioning converter, a missing "codec" file for it (a quite common problem), or even its slightly misleading menus tricking you into saving the file the wrong way) has led to the output file itself being raw uncompressed HQ audio data... The one hole in that idea is I've had difficulty finding a combination of common recording settings that would cause an inflation from exactly 1.46 to somewhere between 9.50 and 10.99mb. I can do 1.46 up to either 8.76 or 11.68 quite easily... but even though there _are_ a couple of options that come out at very nearly 10.00, you have to get there from an uncommon starting point...