Pitifully minor bug reports Options

[h]

Just a couple things I've noticed, but am too lazy to actually fix the code myself:

- Command line encoder progress reporting goes crazy if the input file name length is too large - try oggenc "This is a friggin long file name blah blah blah blah.wav", or even longer if needs be, and you'll see a nice scrolling effect.

- WAV input module can't seem to cope with the (seemingly valid) output from WavTrim if silence has been trimmed from the file's beginning. I don't know much about the WAV format, so I assume it's either setting a cue point that Ogg doesn't like, or messing around with the header length to make whatever's reading it skip over to the first non-silent sample. Not really major, but it's annoying to trim songs with WavTrim, then open and resave them all with CoolEdit just so Ogg can encode them.

-h

[Nic]

Similar problem with the Wav in module in OggEnc (I got so pissed off with it, I wrote my own little one which will accept any stereo wav - www.freewebz.com/nixstuff - Maybe if OggEnc is really pissing U off try mine) (AudioGrabber 1.61, which I use to rip my CDs, creates wav files it doesn't like....

There is no need to load the complete header for audio Encoding. Just grab the samplerate and channels (& im not sure but is stereo the only one fully implemented at present????? Could be completly wrong about that, but just by browsing through the source i.e. modes.h)

Also, why cant Vorbis be forced below 63kbps???

Cheers,
-Nic

[Jon Ingram]

Also, why cant Vorbis be forced below 63kbps??? :)

You can force Vorbis to go below 64Kpbs quite easily by resampling the input. In Linux you'd use something like (from memory):

sox input.wav -t wav -r 22050 - | oggenc -o output.ogg -b 64 -

and that will get you about 45Kpbs (your milage may vary). I suppose there's an argument for including a resampler into the encoder frontend like LAME -- in fact they could probably just use the LAME code as the licenses are similar. There are probably more important things on their minds though.

[Garf]

Originally posted by Nic
There is no need to load the complete header for audio Encoding. Just grab the samplerate and channels (& im not sure but is stereo the only one fully implemented at present????? Could be completly wrong about that, but just by browsing through the source i.e. modes.h)

If you have any idea what causes or triggers the problems with OggEnc, WAV or stdin please report them to the Vorbis lists/people.

Specifically, I also saw the Audiograbber problems and Vakor (oggenc maintainer afaik) had no idea what caused them. If you do, please mail him.

Also, why cant Vorbis be forced below 63kbps???

I will make a FAQ item for this (and similar questions)

--
GCP

[Nic]

I looked through the source of OggEnc a little way back. The way the Wav files are loaded is very safe (as the whole header is read including the 'fmt' 'data' etc). I may email him and suggest in addition to the force raw parameters there is a /force wav reading. (which simply reads the channels & sample rate and then skips the 0x44 byte header)

The FAQ would be really good, as I have only just started toying with Vorbis and I havent found a good source of info on it (apart from wading through r3mix.net & the source

Cheers & good luck with the FAQ Garf,

Cheers,
-Nic