Talk:Yamaha YM2612

Untitled
How is this chip compatible with the DX synthesizers? The most of the DX line had 6 ops, this chip only has 4 ops. I'm dubious of that statement 50.0.18.170 (talk) 18:27, 17 September 2012 (UTC)
 * The most of the DX/TX synthesizers is actually 4Op, not 6Op. You can count: DX9, DX21, SFG01 and SFG05 modules, DX27, DX100, TX81Z, DX11 (and let's not forget about FB-01 and V50!) contrariwise DX7, DX1, TF1 module, TX7, DX5 and TX802. OPN2 (and all OPN-family) is almost completely compatible with synthesizers based on the OPM/OPP family chips (DX21, SFG01, SFG05, DX27, DX100, FB-01): in the main FM part they are almost identical! All values are exactly the same, the only difference is in LFO and availability of DT2 in OPM/OPP. You can easily transfer instruments from OPN2 to OPM/OPP and vici versa. In turn, OPN2 is not compatible with DX9 (because it's just a cut off of DX7, which has very different ADSR and other settings of operators) and synthesizers based on the OPZ family (TX81Z, DX11, V50 and many other synths) (it has much more flexible settings of operators, OPN2 is can be transfered, but only manually). --Krik iddqd (talk) 07:35, 1 March 2018 (UTC)

I also sincerely doubted that, so I removed it. I later found the Japanese YM3438 application manual which leads me to believe it's more based on the YM2203C, not the YM2608, which makes more sense.108.73.143.1 (talk) 11:32, 15 February 2018 (PST)
 * No, it's exactly a cropped YM2608 (OPNA). This is supported by the presence of a completely analogous LFO, which was absent in YM2203 (OPN), and the layout, which is precisely a cropped OPNA, not OPN (compare block diagrams). --Krik iddqd (talk) 06:35, 1 March 2018 (UTC)
 * Yet the application manual for the YM3438 (CMOS version of the YM2612) _states outright_ that it is derived from the FM synthesis block of the YM2203C. Which the YM2608B's application manual also claims, which makes this matter a bit confusing from the perspective of keeping things 'encyclopedic'. Also, the YM2608B can't set FM channel 6 into direct DAC output mode, so there's at least a few low-level differences with the YM2608B, beyond the removal of the SSG and added integrated DAC. Eep386 (talk) 04:11, 29 May 2020 (UTC)

Sample audio isn't accurate
The sample audio for the YM2612 (second example) was taken from the Streets of Rage 2 soundtrack composed by Yuzo Koshiro playing on a Sega Megadrive/Genesis. Although the Genesis has a YM2612, it also has a 4 channel PSG for compatibility with the Sega Master System. Composers are allowed to use the PSG if they like when making music for Genesis titles, and that sample clearly includes sound from the PSG. That may confuse people because they may think the sound quirk referenced in the article is coming from the 2612 when it's not. Another sound sample should be used, or the same track should be recorded from an emulator with the PSG channels muted.


 * I don't even think we are supposed to have these non-free use samples uploaded here anyway. ~ Dissident93 (talk) 01:27, 24 July 2019 (UTC)
 * If somebody wants to replace the samples with free alternatives that suitably demonstrate the crossover distortion, I'm alright with that. Just that these were probably the more 'classic' examples of the DAC distortion. (The 'Evil Territory' song from Beyond Oasis uses only the FM synthesis, no PSG or PCM/DAC.) Eep386 (talk) 04:11, 29 May 2020 (UTC)

Error in technical details
YM2612's wiki page says : "The YM2612's FM synthesis block is an extended version of the FM block featured in the YM2203, adding two channels"

But the YM2203 has 3 FM channels and the YM2612 has 6 FM channel. If you make the maths, if you add 2 channels to 3 channels you got 5 channels not 6. However YM2612 has 6 FM channels and YM2203 has 3 FM channel.

It's a mistake or I missed something ?

Naming error
"A VST, AU and AAX and Reason Studios Rack Extension plugin called YM2612 was made by Inphonik." It's called RYM2612 and not YM2612. — Preceding unsigned comment added by 2A01:CB14:19F:F400:BC57:DCE3:D754:DF1E (talk) 09:06, 11 June 2022 (UTC)