README
vgm-conv 
Chip-type and clock converter for VGM.
Clock Conversion
vgm-conv can change chip's clock while maintaining note's frequency. Supported chips are sn76489, ay8910, ym2151, ym2203, ym2608, ym3812, y8950, ym3526, ymf262 and ym2413.
Type Conversion
vgm-conv supports non-trivial conversion across various chip types.
|FROM|TO| |-|-| |ay8910|ym2203, ym2608, ym3812, y8950, ym3526, ymf262| |sn76489|ay8910, ym2203, ym2608| |ym2413|ym2608, ym3812, y8950, ym3526, ymf262| |ym2203, ym2203.ssg|ay8910, ym3812, y8950, ym3526, ymf262| |ym2608, ym2608.ssg|ay8910| |ym2203, ym2203.fm|ym2413, ym3812, y8950, ym3526, ymf262| |ym2608, ym2608.fm, ym2608.r|ym2413| |ym2612, ym2612.fm, ym2612.dac|ym2413| |ym3812|ym3526, y8950| |ym3526, y8950|ym3812|
Note: downgrade conversion (ex. YM2203 to YM2413 conversion) is highly limited thus don't expect much.
Limitation
- YM2413 to YM2608 supports FM1-6 and rhythm conversion. FM7,8,9 channels are ignored.
- YM2608 to YM2203 does not support the rhythm part conversion.
- SN76489 to AY8910
- Noise conversion is pertially supported; SN76489 has the independent noice channel and the periodic noise generator but AY8910 does not, so full conversion is not possible.
- As for YM2612 DAC, only register 2A stream can be converted. VGM's DAC stream commands is not supported.
- Dual chip is not supported.
- YM3812 to YM3526 or Y8950 conversion drops the waveform information.
- Y8959 ADPCM is not supported.
Install
$ npm install -g vgm-conv
Example
Convert YM3812 clock to 4.00MHz
$ vgm-conv -f ym3812 -c 4000000 -o output.vgm input.vgm
Convert YM2612 to YM2413(@3.58MHz)
$ vgm-conv -f ym2612 -t ym2413 -o output.vgm input.vgm
Convert YM2612 to YM2413(@4.00MHz)
$ vgm-conv -f ym2612 -t ym2413 -c 4000000 -o output.vgm input.vgm
Convert only YM2612 DAC part to YM2413 7.5bit DAC
$ vgm-conv -f ym2612.dac -t ym2413 -D useTestMode=true -o output.vgm input.vgm
Convert YM2203's FM part to YM2413 and SSG part to AY8910
$ vgm-conv -f ym2203.fm -t ym2413 input.vgm | vgm-conv -f ym2203 -t ay8910 -o output.vgm
Usage
$ vgm-conv --help
vgm-conv
Chip-type and clock converter for VGM.
SYNOPSIS
vgm-conv [<option>] <file>
OPTIONS
-i, --input file Input VGM file. Standard input will be used if not specified.
-f, --from chip Specify source chip type.
-t, --to chip Specify destination chip type.
-c, --clock clock Specify clock in Hz of destination chip. The typical clock value will be
applied if not specified.
-D, --define name=value Define converter option variable. See below.
-o, --output file Output VGM file. The standard output is used if not speicified. If the given
file is *.vgz, the output will be compressed.
--no-gd3 Remove GD3 tag from output.
--voice-table file Specify the voice table file.
-v, --version Show version.
-h, --help Show this help.
CLOCK CONVERSION
AVAILABLE CHIPS
sn76489, ym2612
ay8910, ym2151, ym2203, ym2608
ym3812, ym3526, y8950, ymf262
ym2413
CHIP CONVERSION
FROM TO
ay8910 ym2203, ym2608, ym3812, y8950, ym3526, ymf262
sn76489 ay8910, ym2203, ym2608, ym2612
ym2413 ym2608, ym3812, y8950, ym3526, ymf262
ym2203, ym2203.fm ym2413
ym2608, ym2608.fm, ym2608.r ym2413
ym2203, ym2203.ssg ay8910
ym2608, ym2608.ssg ay8910
ym2612, ym2612.fm, ym2612.dac ym2413
ym3812 ym3526, y8950
ym3526, y8950 ym3812
YM2203 to OPL (YM3812/Y8950/YM3526/YMF262) OPTIONS
-D ssgAttenuation=n Set SSG volume attenuation level to 0.75*n(dB). The effective range is -63<=n<=63.
YM2612 to YM2413 OPTIONS
-D decimation=n Decimate 1 of n PCM data. 2 to 4 is recommended if USB serial device (like SPFM) is used to play VGM. n=0
disables the feature and results the best playback quality. The default value is 4.
-D dacEmulation=fmpcm|test|none fmpcm: use pseudo 6-bit DAC is used (default).
test: use YM2413 test mode is used 7.5bit DAC but disables all YM2413 FM channels.
none: disable DAC emulation.
SN76489 to AY8910 OPTIONS
-D mixChannel=value Specify the AY8910 channel used for noise output. The value must be one of 0, 1, 2 or none. The default value is 2.
If 'none' is specified, all noise part will be silent.
Since AY8910 has no independent noise channel, SN76489's noise channel will be mixed with a tone channel into the single AY8910's channel specified by this option.
-D mixResolver=tone|noise|mix This option determines the behavior when tone and noise are requested to be key-on simultaneously on the same AY8910 channel.
'tone': tone will be output. noise will be silent.
'noise': noise will be output. tone will be silent.
'mix': both tone and noise will be output (default).
-D periodicNoisePitchShift=n The pitch shift amount of the periodic noise conversion. pow(2, -n) will be multiplied to the noise frequency. The default value is 4.
-D channelAttenuationMap=n1,n2,n3,n4 Volume attenuation mapping for SN76489 channels. n1, n2, ... n4 correspond to SN76489's ch1, ch2, ... ch4 respectively. The default value is 0,0,0,0.
-D whiteNoiseAttenuation=n Additional volume attenuation for the white noise. This value will be added to the n4 specified on volumeAttenuationMap. The default value is 0.
-D periodicNoiseAttenuation=n Additional volume attenuation for periodic noise. This value will be added to the n4 specified on volumeAttenuationMap. The default value is 0.
-D noisePitchMap=n1,n2,n3 The noise frequency of AY8910. n1, n2 and n3 correspond to SN76489's noise frequency 0, 1 and 2 respectively. The default value is 7,15,31.
EXAMPLES
YM2612 to YM2413 $ vgm-conv -f ym2612 -t ym2413 input.vgm
Both YM2413 and AY8910 to YM2608 $ vgm-conv -f ay8910 -t ym2608 input.vgm | vgm-conv -f ym2413 -t ym2608
YM2203's FM part to YM2413 and SSG part to AY8910 $ vgm-conv -f ym2203.fm -t ym2413 input.vgm | vgm-conv -f ym2203 -t ay8910
Only DAC part of YM2612 to YM2413 $ vgm-conv -f ym2612.dac -t ym2413 input.vgm
YM2612 to YM2413@4.00MHz $ vgm-conv -f ym2612 -t ym2413 -c 4000000 input.vgm
Voice Table (Beta)
The voice table configuration can be defined in JavaScript. Only OPN/OPNA/OPN2 to OPLL conversions are supported. To load the voice table configuration, use --voice-table <file> option.
const [HH, CYM, TOM, SD, BD] = [1 << 10, 1 << 11, 1 << 12, 1 << 13, 1 << 14];
const [Violin, Piano, Guitar, Flute, Clarinet, Oboe, Trumpet, Organ, Horn, Synthsizer, Harpsichord, Vibraphone, SynthBass, WoodBass, ElectricBass] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15];
module.exports = {
voiceTable: {
opn2opll: {
// Override `-D dacEmulation` option. This must be set "none" to enable rhythm channels.
dacEmulation: "none",
// User defined voice table
voices: {
// Original Tones can be defined from program number 16 to 1023.
16: [0x01, 0x01, 0x1c, 0x07, 0xf0, 0xd7, 0x00, 0x11],
},
mapping: {
// i: Voice Number (required)
// - 1...15: ROM Voice
// - 16...1023: User defined voice
// - 1024: HiHat, 2048: Cymbal, 4096: Tom, 8192: Snare, 16384: Bassdrum
// v: Volume offset: -15<=v<=15 (default: 0)
// o: Octave offset: -7<=o<=7 (default: 0)
"520030001c1f257fdf1fdf1f0709068607060608251515f5000000000100": { i: Harpsichord, v: 1, o: -1 },
"0f30005000120f121f1f1f1d01000106000f010b113117f1000000003e00": { i: SD, v: -1, o: -1 },
"3e5051501f171c10df1bdf1f07070e040701010154f65572000000002b00": { i: WoodBass },
"0f300050001a171a1f1f1f1d01000106000f010b113117f1000000003e00": { i: 16, v: 1, o: -1 },
},
// autoMap controls the fallback voice if no match is found in the mapping table.
// - An appropriate ROM voice will be selected if autoMap is true.
// - Slient if autoMap is false.
autoMap: true,
}
}
};
The original instrument hash to the program number mapping can be seen on the console after running vgm-conv without specifiying --voice-table. You can copy them into the template above.