Conversation

Notices

1. roka (roka@pl.smuglo.li)'s status on Sunday, 18-Oct-2020 11:36:03 JST roka
   Here's how Gekkan Mugentan Torico stores every single voice cue: Input #0, film_cpk, from 'z0020_08': Duration: 00:00:03.83, start: 0.000000, bitrate: 177 kb/s Stream #0:0: Audio: pcm_s16be_planar ([1][0][0][0] / 0x0001), 11025 Hz, 1 channels, s16p, 176 kb/s

   That's horribly wasteful. Using IMA ADPCM at this sampling frequency means you can drop bitrate from ~177 kbps to mere 44.8 kbps at no discernible drop in quality. And since Torico is an FMV game, adding an ADPCM decoder would've been probably a better option, especially since SEGA FILM container could handle CRI ADX ADPCM streams. I guess they didn't want to pay a loiceince? It was a relatively uncommon combination at the time and most games did use various PCM streams instead :cirnoThink:

   The combined size of the voice cues in the first disc of Torico (PAL version) is 13M. Using IMA ADPCM makes it drop down to 3.6M.

   toricojp.jpg
   • roka (roka@pl.smuglo.li)'s status on Sunday, 18-Oct-2020 11:40:48 JST roka

     in reply to

     • Tija-nee during calamities
     oi @tija, whatcha think? :rokalife:
   • roka (roka@pl.smuglo.li)'s status on Sunday, 18-Oct-2020 11:59:46 JST roka

     in reply to

     • Tija-nee during calamities
     • an actual bear in a hazmat suit
     @tija (and @guizzy; hi!) samples. I did reformat the original stream from planar to linear PCM for ease of playback but that should mean nothing.z0020_08.wavz0020_08-adpcm.wav
   • roka (roka@pl.smuglo.li)'s status on Sunday, 18-Oct-2020 12:11:30 JST roka

     in reply to

     • Tija-nee during calamities
     • an actual bear in a hazmat suit
     @guizzy @tija yeah but think about it this way: if they were perfectly fine with ~177 kbps per line, they could've also upped the sampling frequency. That would reduce the glitches. With a comparable bitrate, you can store an audio clip with 44.1kHz sampling frequency.

     Sample reencoded with ADPCM compression applied – it won't sound better than the original for obvious reasons but…

     z0020_08-adpcm-44100Hz.wav
   • an actual bear in a hazmat suit (guizzy@pleroma.guizzyordi.info)'s status on Sunday, 18-Oct-2020 12:11:31 JST an actual bear in a hazmat suit

     in reply to

     • Tija-nee during calamities
     @roka @tija Both sound perfectly serviceable, maybe a bit less bass on the ADPCM one?
   • roka (roka@pl.smuglo.li)'s status on Sunday, 18-Oct-2020 12:23:00 JST roka

     in reply to

     • Tija-nee during calamities
     • an actual bear in a hazmat suit
     @guizzy @tija I guess one reason could've been them pushing this data as-is to the SCSP for bandwidth reasons (the LSI didn't give a flying fuck about source data sampling rate as long as it was PCM and of frequency rate less than 44.1kHz)… but it's an FMV game.
   • roka (roka@pl.smuglo.li)'s status on Sunday, 18-Oct-2020 12:24:02 JST roka

     in reply to

     • Tija-nee during calamities
     • an actual bear in a hazmat suit
     @guizzy @tija I guess there was Cinepak decoding happening on one of the SuperH's at the same time as playback but still.