Donald,

I have programmed a (hardware-based) sequencer in my college years. From that experience, I can tell you, that for sequencer (as for most devices) it is much harder to record than to play back. Without going into too much detail, the sequencer does not know when the next midi event is going to come in, so whenever it does, the processor has to interrupt whatever it is doing and record the midi event accordingly. Unless you have a very fast processor, capable of handling a lot of interrupts, the recording resolution will be limited, especially keeping in mind that that same processor has to do many other things, such as driving the user interface and generating sounds. (There are other ways to record, such as polling the MIDI input, but these too have to be occuring at a finite frequency).

On the other hand, playing back a sequence is much simpler - the computer knows when the next note will need to be played, and can schedule its activities accordingly. Hence, the recording resolution may well be set well below that of playback. In the XP-80 case, one tick of the recording resolution corresponds to 5 ticks of playback, hence using the internal sequencer to record causes quantization error.

I think that you are right in attrubuting the poor sequence quality to the quatization error - the sequencer does not record exactly what you are playing in the first place, but rather something more robotic sounding.

Regards,
Alex

Having said that,