MIDICraft's MIDINET CD-ROM/MIDICraft's MIDINET CD-ROM.iso/DOCS/YAMFM4P.ZIP/YAMFM4P.DOC

Text File | 1991-06-05 | 5.6 KB | 121 lines

                            Yamaha FM 4-OP Doc

I was rather surprised by  all  the  postings  on  how  difficult it was to
convert 4-OP patches to a 6-OP synth. Since I  do it all the time by ear, I
wondered if I had some magic ability that I was using unconsciously. Then I
read the postings more carefully, and  discovered  that most of the posters
had never tried the conversion, didn't even  have access to both a 4-OP and
a 6-OP synth, and  were  apparently  speculating  based  on spec sheets and
overheard conversations.

What a wonder is USENET; such  wholesale production of conjecture from such
a trifling investment in fact.

Fact is, the conversion is  pretty  simple,  if  you know the basic tricks.
Here then, for  the  net's  edification:  a  primer  on  converting 4-OP FM
patches to a 6-OP synth.  Since  you  probably  want something more precise
than the back-of-the-envelope  technique  I  usually  use,  I  sat down and
worked out formulas for all the tough-to-guess parameters.

Caveats: All comparisons were done by ear, between a DX-100 (a DX-27 with a
mini keyboard) and a DX-7s.  So  if  you  have  a  TX-81Z, FB-01, or DX-21,
please think before you flame.
___________________________________________________________________________

ALGORITHM: Use the  obvious  mappings.  Two  cautions:  pay  attention when
copying parameters, since you will probably be mapping operators 1/2/3/4 to
1/2/5/6. More than once I copied  half  the parameters into operators 3 and
4, and the others into 5 and 6. :-( Also, watch out where the feedback loop
is.

FEEDBACK LEVEL: Same, although the  4-OP  has  slightly  more "bite" at the
same level. That is, at FBL=7 the  4-OP  has more "bite" than the 6-OP does
at 7, but the 4-OP at FBL=6 has  less  of  a "bite" than the 6-OP at 7. You
may want to try adding 1 to the original FBL, and see how it sounds.

FREQ RATIO: Same. All of the 4-OP  ratios  can be achieved by a 6-OP synth.
Of course, the 6-OP synth should be set for Frequency(Ratio).

DETUNE: Same. Well, almost. 1,2,3 on the  4-OP really come out as 1.5, 2.5,
3.5 on the 6-OP, but it doesn't have  those  so you'll have to make do with
1,2,3.

Incidentally, I don't know about the old DX-7, but my DX7s is never in tune
on all keys with the DX-100. The  variations  are all less than 0.2 Hz, but
quite audible when you beat sinewaves  together.  Wasn't this what made the
Minimoog sound so good? :-)

OSC. WAVE: If you have a TX-81Z and  this  isn't set to 1, you get to lose.
I'd guess that as long as no more  than two operators use a Wave other than
1, you could come pretty close by  using additional modulators. But I don't
have a TX-81Z, so I don't know. I'd want an oscilloscope for this.

ENVELOPE GENERATOR: The 4-OP synth gives you control over four rates (range
0 to 31 on the first three, 0 to 15  on the last) and one level (range 0 to
15). The 6-OP gives you control over  four rates and four levels, all range
0 to 99.

    4-OP          6-OP
    ----                ----
     AR  corresponds to  R1
    D1R  corresponds to  R2
    D2R  corresponds to  R3, which defaults to 0 on 4-OP and 99 on 6-OP
     RR  corresponds to  R4
                   L1  must be set to 99
    D1L  corresponds to  L2
                   L3  must be set to 0
                   L4  must be set to 0

To convert AR, D1R, and D2R to R1, R2, and R3, multiply by 3 and add 10. To
convert RR to R4, multiply by 6  and  add  10. Yes, this doesn't quite work
out as you approach maximum; you'll have to fudge it a bit. A value of zero
should be set to zero, since this means an infinite time.

The relationship between D1L an L2 is audibly non-linear.
Refer to the table:

    D1L  15  14  13  12  11  10   9   8   7   6   5   4   3   2   1   0
    L2   99  96  92  87  82  78  74  70  66  63  60  56  53  50  47   0

Any mapping except 15->99 may be  as  much  as  two  steps too large; I was
doing the best my ears could manage.

OUTPUT LEVEL: Range 0 to 99 on both, but slightly offset. The 6-OP level is
equal to the 4-OP level plus  8.  Yes,  this means that 4-OP levels greater
than 91 cannot be reproduced by the 6-OP  synth. On a carrier, this is easy
to deal with; on a modulator, it is not. Some day I'll figure out how to do
this with another modulator. For now, you get to lose. Fortunately, none of
the patches I wanted to convert used  a modulator Output Level greater than
90.

Of course, the Output Level of the unused operators should be set to 0.

KEYBOARD RATE SCALING: The 4-OP has a  bit of rate scaling even when KRS=0.
And on the DX-100 (at least), rate scaling has discontinuities in it, which
makes it a tough call to get *exactly* the same. But this is close:

    4-OP KRS   0  1  2  3
    6-OP KRS   1  2  5  7

KEYBOARD LEVEL SCALING: Start by setting  the 6-OP level scaling parameters
as follows: Break Point = A1; L-Curve =  -LIN; R-Curve = -EXP; L-Depth = 0.
Then set the R-Depth to 90% of the  4-OP KLS. 99 becomes 90, 50 becomes 45,
etc.

This is almost perfect, but not quite; I'd like to hear from anyone who has
refined it. I tried  moving  the  break  point  and  fiddling with the left
curve; it didn't sound any better, and was harder to program.
___________________________________________________________________________

There you have it. Following  these  rules,  I  converted  a variety of the
factory presets  from  my  DX-100  to  my  DX-7s;  the  two  were virtually
indistinguishable. Except that the DX-7s sounded much cleaner.

The LFO, AMS, PWS, AMD, PMD, velocity curve, and pitch envelope are left as
an exercise for the  reader.  :-)  Seriously,  I  find  most of the factory
settings for these pretty wretched, and they are easy to set by ear.

<csg>