QUOTE
Originally posted by john33
Yeah, sorry about this but I can't seem to find where this occuring at the moment. The odd thing is that the frontend to oggenc is unchanged from yesterday when oggdrop and vorbis.dll all compiled and ran fine. However, the backend did change, obviously in a way that didn't affect oggenc, but did affect the other two. I just see can't where it is. If I still had a copy of yesterday's programs, I could make a comparison but they were overwritten in the CVS update! I haven't given up, but I'm kind of shooting in the dark!!
john33
As far as I know, current oggdrop source code seems to be obsolete when compared to oggenc especially in encode.c (ie. oggdrop one doesn't use vorbis_bitrate_addblock() in oe_encode function. ... but I don't know how this function works

)
And the main difference between 20011218 and 20011219 is that ABR mode is fully supported normal/minimum/maximum bitrate setting by bitrate management on VBR.
ABR encode on 20011218 is only passed through to VBR encode function, transforming normal bitrate to VBR base quality.
I compiled oggdrop today and checked that vorbis_encode_init(&vi, opt->channels, opt->rate, -1, opt->bitrate*1000, -1); resulted in extraordinary high bitrate, while vorbis_encode_init(&vi, opt->channels, opt->rate, 192000, 128000, 96000); resulted in about normal bitrate.
But ABR mode on 20011219 seems to be unstable at the point encoding result get too low even when I encode the same song continuously with proper and same minimum/maximum/normal bitrate setting (not -1). This affects oggdrop but not oggenc.
I'm looking forward to updating tomorrow.
Thanks,