From Brian Willoughby Sent Sun, Sep 30th 2018, 18:42
One thing I noticed is that each group of algorithms has a different = amplitude, depending upon the total number of final oscillators. For example, that DX7 =E2=80=9Cstraight line=E2=80=9D algorithm has each = of the 6 operators at almost 16 dB down. A 4 operator DX would only be = 12 dB down, at most. Of course, for the algorithms that have the same = number of final oscillators on both 6-op and 4-op variants, the = amplitude reduction is identical. This is just a technique to avoid = digital clipping when the final oscillators are mixed. Of course, there are a lot more differences - between algorithms or = models - than just the amplitude of individual output oscillators. As for bandwidth, is there a chart of the sample rate for each DX model? = I expect that we have no information on the processing power available = inside. It seems doubtful that the processing was actually changed = significantly between 6-op and 4-op, other than perhaps run at a = slightly higher sample rate. Keep in mind that the main motivation for = 4-op was to keep costs down, not to allocate more bandwidth to each = operator. It seems unlikely that the design would have done anything to = increase the cost - at least not without making the same improvements to = the 6-op variants. Brian On Sep 28, 2018, at 1:04 PM, Kenny Balys <xxxxx@xxxxxxxx.xxx> wrote: > The 4-op DX9 has a little more crunch than the 6-op DX7. >=20 > Each operator gets more of the bandwidth.