Doing all the recordings in lossy Ogg Vorbis is a bad idea from the start, as was explained in this forum to you before. Record them as WAVE, VOC, FLAC, anything which is lossless. Then you can do conversions do arbitrary formats from those, you can do editing, cutting, noise shaping, etc..
Actually, I was warned straight after I got so many lines.
Anyway, they've been instructed on keeping the sound clean to keep editing to a minimum. Justin even made a cover (to decrease "PUH"s on Ps and breaths) out of tights (or pantyhose to Americans).
iPwnzorz wrote:Actually, I was warned straight after I got so many lines.
Anyway, they've been instructed on keeping the sound clean to keep editing to a minimum. Justin even made a cover (to decrease "PUH"s on Ps and breaths) out of tights (or pantyhose to Americans).
So convert the ogg's to mp3 or wave, what's the problem?
I think I read that the problem with OGG in general is that it needs floating point operations, and afaik the DS doesn't have that built into the hardware, so it'd be horribly slow.
And good luck with the speech project! Don't let the criticism get you down - while I also think that using lossless compression makes more sense until release time, even if you lose quality by saving a dozen times I assume it'd still sound far better than the speech in games like BASS or FOTAQ.
Still tremor is a fixed point implementation of ogg, but still it requires abit of cpu, and I guess more than the poor DS can handle.
And even if all lines are already recorded, I guess,converting it to wav again would n't be to hard, what bitrates are the oggs in anyway? On mobile platforms I usually recommend very low bitrates, to lower the amount of data to be be read and decoded.