You might get a warning about "-acodec vorbis" being experimental. You can use "-acodec libvorbis" instead.

Be careful with the "-aq" option, the value 100 (as suggested by a previous post in this thread) is definitely too high. The supported values are listed in the "Technical details" section of the Vorbis wikipedia page.

I used "-qa 4" (roughly 128 kbit/s):

for wma in *.wma; do ffmpeg -i "$wma" -acodec libvorbis -aq 4 "$wma"_aq4.ogg; done

Wma is lossy, ogg vorbis is lossy, lossy-> lossy transcodes will make the files sound worse.

Its better to rerip as ogg vorbis, and if thats not an option, then maybe its possible to find the same files in MP3, ogg vorbis or a lossless codec.

Wma is lossy (and I dont trust wma 'lossless' to be lossless). Transcoding a lossy-> lossless file is just a good way to use space, its not going to sound any better.

For sane transcodes (lossless->lossy, lossless->lossless) you can normally use soundconvter (GTK) or soundkonverter (Qt). Some of the linux media players will transcode as well. No need to go poking around the command line to transcode, unless you want to.

I have not noticed any difference between the wma files (that I found in a torrent download) and the ogg files that I converted them to (using ffmpeg -acodec libvorbis -aq 4).

Maybe other people's ears are finer or other people are of the "audiophile" type and have stereo systems worth thousands of dollars that make those subtle differences noticeable.

For me, using a relatively cheap compact stereo and my laptop as the aux sound source, this doesn't matter.

And I just wanted to have the songs in my collection so I did not spend time hunting for the same tracks in other (lossless) formats.

Personally I dont need thousands of $$$ worth of stereo to be able to tell the difference between low bitrate lossy files and lossless. I cant deal with any lossy file at lower than 160-192K, to me they sound flat, lifeless and bland.

Most wma files I've ever seen have been 192k/sec, vorbis -q 4 is 128k/sec. AFAIK the ffmpeg 'aq' is a 0-100 quality level, not the q-2 (32k/sec) to q10 (500k/sec) of standard vorbis. It thats right, I'd love to know what bitrate those vorbis files encoded at, it could be a tiny bitrate.

http://howto-pages.org/ffmpeg/

Even if wma and ogg vorbis used the same method for discarding information from the audio track (they dont) the ogg vorbis file should sound worse than the wma.

If you cant pick the difference, thats fine for your use. But I wouldnt assume that the limit is your ears, laptops tend to have pretty lame sound chips, and the 'bookshelf'/compact speakers wont be helping.

There is a difference between
-acodec vorbis (-aq values 0..60)
-acodec libvorbis (-aq values -1..10) <= This is what I used

See http://www.linuxquestions.org/questi...6/#post3827342

It is more of a sidenote than any great help but you can always pipe ffmpeg output directly to oggenc if you like: