ChanServ changed the topic of ##yamahasynths to: Channel dedicated to questions and discussion of Yamaha FM Synthesizer internals and corresponding REing. Discussion of synthesis methods similar to the Yamaha line of chips, Sound Blasters + clones, PCM chips like RF5C68, and CD theory of operation are also on-topic. Channel logs: https://freenode.irclog.whitequark.org/~h~yamahasynths
l_oliveira has quit [Quit: ChatZilla 0.9.92-rdmsoft [XULRunner 35.0.1/20150122214805]]
futarisIRCcloud has joined ##yamahasynths
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
futarisIRCcloud has joined ##yamahasynths
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
_whitelogger has joined ##yamahasynths
futarisIRCcloud has joined ##yamahasynths
_whitelogger has joined ##yamahasynths
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
andlabs has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
andlabs has joined ##yamahasynths
<cr1901_modern>
Lord_Nightmare, Sarayan: Remind me again when you have the time... how many versions of the sin/exp tables exist if you know?
<cr1901_modern>
(by versions I mean "the IP that Yamaha used when creating the table on the die". There might've been multiple versions of the same ROM for die shrink etc)
<Sarayan>
Not sure you can talk about IP about something that can be reproduced 1:1 with math
<Sarayan>
in any case, to answer your question, one
<Sarayan>
even when going from nmos to cmos, the actualy contents are identical afaict
<Sarayan>
but that's for the "consumer" chips. The fm-in-multiple-chips synths have different (bigger) tables.
<cr1901_modern>
Ahhh
<cr1901_modern>
And I suppose s/IP/"building block"/
<cr1901_modern>
the idea is "they designed it once, used it repeatedly"
<Sarayan>
in particular, the ym2203 and the opl3 tables are identical
<superctr>
i don't think yamaha ever felt the need to recalculate the sin/exp tables
<Sarayan>
and these are pretty much the two extremes in the consumer fm area
<superctr>
and they're just mask roms
<Sarayan>
ctr: they could have added bits, in input or in output
<Sarayan>
bigger resolution, more filling
<superctr>
it would be interesting to see what the PCM chips do
<superctr>
YMW258, SCSP and AICA
<superctr>
they have similar envelope generators to the FM chips
<superctr>
so i would expect at least the exp table to be there
<superctr>
but the values could be different
<cr1901_modern>
They probably didn't spend much time calculating the tables anyway... prob some fortran script to generate evenly spaced sin/exp intervals in the early 80s
<superctr>
i think that perhaps a modular approach could work in emulating those chips, since all yamaha chips seems to share IP in some way
<superctr>
maybe it could also fix the chips that still don't sound quite right in MAME, like the OPX (YMF271)
<Sarayan>
according to david, a modular approach doesn't work because of a pile of niggly details
<Sarayan>
all the subtle differences at any point in the pipeline makes generic code unreadable
<cr1901_modern>
YM2151 and 2612 have different pipeline lengths according to nukeykt
<cr1901_modern>
actually number of pipelines* even
<superctr>
well, YM2151 has more operators
<cr1901_modern>
(apparently 2612 has two pipelines)
<superctr>
32 vs 24 operators
<superctr>
and 2612 is basically two YM2203s slapped together
<superctr>
with SSG and IO peripherals taken out
<Sarayan>
the sin and exp roms are part of the pipeline, so if you only have one copy of them you only have one pipeline
<superctr>
access to them can be multiplexed
<Sarayan>
the multiplexer would take more surface than the rom itself
<superctr>
having two pipelines properly means having to duplicate everything on the die
<cr1901_modern>
I could misunderstand what was meant by "two pipelines"
<cr1901_modern>
Hardly the first time
<Sarayan>
do we have a 2612 shot ?
<superctr>
we have 3438 at least
<cr1901_modern>
Idk if it's on sipr0n, but I know _of_ a 2612 die shot
<cr1901_modern>
someone was using it for ladder effect stuff
<superctr>
the fullsize might be somewhere else in the thead
<superctr>
top thread on spritesmind sound section
<andlabs>
sik said something about the spritesminds files being gone?
<andlabs>
idk
<andlabs>
oh btw
<andlabs>
we lost a Wonder MIDI cart :<
<andlabs>
it sold for roughly $3000 on YJA
<cr1901_modern>
Great week for preservation /s
<Sarayan>
you see old1, oldold1 and old2? That's where the number of channels happens
<Sarayan>
you have one (serial) register per channel
<Sarayan>
2203 has 2, 2612 has 6, etc
<Sarayan>
2203 has 3 of course
<superctr>
hmmm
<superctr>
what about the 2610 then
<superctr>
since there's the B variant with 6 channels
<superctr>
if they did any major changes to the die or just cut off the two channels but kept the logic
<superctr>
similar to the VRC7
<Sarayan>
2610 seems to have 6, looking at the die
<Lord_Nightmare>
cr1901_modern: afaik two, MAYBE three: almost all chips use one common pair of tables, and ym2414B uses a different higher-resolution set
<Lord_Nightmare>
ops and egs might use a different set too, i'm not sure
<Lord_Nightmare>
the dx7 chips
<Lord_Nightmare>
and the pre-dx7 proto-fm chips yamaha did for their electric piano stuff (which also had built in patches), that stuff i have no clue about