Polyphony SC-55 vs SC-55mkII

[Salient]

I came to a strange conclusion today. At least, I wouldn't have expected it this way.
Here's the story.
Since the SC-55 mk II has 28 voice polyphony and the original SC-55 has only 24 voice polyphony, I would expect note drop-off to be more of a problem with the original SC-55 than with the mark II.
However, probably either due to differences in tone generation (maybe some tones that were made of one voice in the SC-55 are now made of two in the SC-55mkII?) or a difference in fall-off priority I have noticed that it can work the other way around.

Listen to these two samples and be 'amazed'

Settlers II music on SC-55
Settlers II music on SC-55mkII

[Great Hierophant]

Or it could be that the music is relying on the capital tone fallback feature found only on the SC-55 and other first gen Sound Canvas devices. 

[Salient]

Or it could be that the music is relying on the capital tone fallback feature found only on the SC-55 and other first gen Sound Canvas devices. 

But then I would get occasional "NO TONE" messages, which I don't get. I think this particular track only uses standard tones anyway.

[Cloudschatze]

Looking through the manuals, the SC-55 and SC-55mkII appear to share the same note-sounding priorities, default voice reserve settings, and number of voices per instrument (for the instruments involved).

I don't have an SC-55, so can you verify that its default voice reserve settings match those of the SC-55mkII (either by SysEx data request, or via the front panel)?

Part 1: 6
Part 2: 2
Part 3: 2
Part 4: 2
Part 5: 2
Part 6: 2
Part 7: 2
Part 8: 2
Part 9: 2
Part 10: 2
Parts 11 - 16: 0

If I had to guess, I'd say the SC-55mkII output is probably the most functionally correct. That is, Part 9, where the dropouts are most noticeable, has the lowest priority (in the piece). With the SC-55, the dropouts are occurring elsewhere. The SC-55 output certainly sounds better, in any event.

If more composers prioritized their Parts properly, or took advantage of the voice reserve functions, this wouldn't be an issue at all. Bump the voice reserve for Part 9 up to 6-or-so (on the SC-55mkII), and voila ~ no problems.

Given the release date of this game, and the 60+ polyphony usage for this piece, I'd say the music was likely composed on/for an SC-88.

[Great Hierophant]

Given the release date of this game, and the 60+ polyphony usage for this piece, I'd say the music was likely composed on/for an SC-88.

I was tempted to say something cheeky like: "Get an SC-88 and these problems will disappear!"

[Salient]

I have an SC-88 equivalent so that is indeed no problem but I was just wonderering how things could be different in this strange manner on what are supposed to be virtually identical sound modules.

@ cloudschatze: The voice reserve mappings are the same as you stated in your post.

[Cloudschatze]

@ cloudschatze: The voice reserve mappings are the same as you stated in your post.

Hmm. Can you send/receive the following data-request block, and post the output?

F0 41 10 42 11 40 10 14 00 00 01 1B F7
F0 41 10 42 11 40 11 14 00 00 01 1A F7
F0 41 10 42 11 40 12 14 00 00 01 19 F7
F0 41 10 42 11 40 13 14 00 00 01 18 F7
F0 41 10 42 11 40 14 14 00 00 01 17 F7
F0 41 10 42 11 40 15 14 00 00 01 16 F7
F0 41 10 42 11 40 16 14 00 00 01 15 F7
F0 41 10 42 11 40 17 14 00 00 01 14 F7
F0 41 10 42 11 40 18 14 00 00 01 13 F7
F0 41 10 42 11 40 19 14 00 00 01 12 F7

[Salient]

Hmm. Can you send/receive the following data-request block, and post the output?

F0 41 10 42 11 40 10 14 00 00 01 1B F7
F0 41 10 42 11 40 11 14 00 00 01 1A F7
F0 41 10 42 11 40 12 14 00 00 01 19 F7
F0 41 10 42 11 40 13 14 00 00 01 18 F7
F0 41 10 42 11 40 14 14 00 00 01 17 F7
F0 41 10 42 11 40 15 14 00 00 01 16 F7
F0 41 10 42 11 40 16 14 00 00 01 15 F7
F0 41 10 42 11 40 17 14 00 00 01 14 F7
F0 41 10 42 11 40 18 14 00 00 01 13 F7
F0 41 10 42 11 40 19 14 00 00 01 12 F7

I did that and the sc-55 briefly displays: TRANSMITTING
and the output is (from MIDI OUT of course) is this:

F0 41 10 42 12 40 10 14 00 1C F7
F0 41 10 42 12 40 11 14 01 1A F7
F0 41 10 42 12 40 12 14 01 19 F7
F0 41 10 42 12 40 13 14 01 18 F7
F0 41 10 42 12 40 14 14 01 17 F7
F0 41 10 42 12 40 15 14 01 16 F7
F0 41 10 42 12 40 16 14 01 15 F7
F0 41 10 42 12 40 17 14 01 14 F7
F0 41 10 42 12 40 18 14 01 13 F7
F0 41 10 42 12 40 19 14 01 12 F7

This is all abracadabra for me but I assume you know what to do with it

[Cloudschatze]

...and the output is (from MIDI OUT of course) is this:

Cool, thanks.

So, apparently the manuals are correct, with regard to the default Assign Mode settings as well. Since these settings also affect polyphony, I wanted to make sure they were the same.

[Salient]

Cool, thanks.

So, apparently the manuals are correct, with regard to the default Assign Mode settings as well. Since these settings also affect polyphony, I wanted to make sure they were the same.

So then the mystery of the difference is not solved yet?

[Cloudschatze]

So then the mystery of the difference is not solved yet?

Well, since I get dropped notes with my CM-500 (GS Ver. 1.10) as well as with the SC-55mkII, I suspect that there is probably a later SC-55 ROM release that "fixes" its behavior similarly. I don't have the SC-55 Service Notes, so I suppose I could check with Roland...

[menkau_ra]

If I need it only for games, do I need those 4 extra notes on 55mkII?

[Salient]

If I need it only for games, do I need thise 4 extra notes on 55mkII?

For games, it seems that with an original SC-55 you are even better off than with a mark II (although this may only be REALLY true for a handful of games). Sound quality of the mark II is a bit better though.

[jharris01]

If I need it only for games, do I need those 4 extra notes on 55mkII?

Any SC55 model (original or mkII) will do just fine for gaming.