Ok here we go 🤣

Recorded your LA and SC MIDI files on my CM-32L and SC-55 (not MKII unit)...

XEEN recordings with CM-32L and SC-55

I prefer the SC version for sure...

HunterZ I will do that recording, I have my xifi elite on the pc next to my dos machine.... I assume you want to hear only the LAPC-I output right? I am gonna use a stereo cable from lapc-i to the front panel of my xifi... now only to finde where this blacksmith is in the town... I have both games in near mint condition but never actually played them , although I am an rpg fan they just don't appeal to me much I guess....

edit: it seems I spoke too soon, Mau1wurf1977 the recording king beat me to it 🤣

Mau1wurf1977 , do you use a freeware tool to record? do you have any recommendations before I search the web for one?

Audacity... It's excellent!

perfect! I connected the rca-output of the LAPC-I (I am 99% sure this is line-out level) to the line-in2 on the xifi drive and audacity works a treat! thanks for the tip!

Looking forward to the LAPC-I recording after comparing Mau's recordings to mine and noticing a lot of differences (would also really like to hear SC-55MkII if anyone has one):

CM-32L versus MT-32 (1.07):

• The sounds on the logo screen on the two recordings sound reversed from each other. This is because I swapped the audio cables that go from my MT-32 to my SC-88's line in because when hooked up correctly (1) the footsteps in SQ1VGA were coming from the wrong side and (2) Cakewalk's MIDI panning was backwards. I'm now not sure at all which way is "correct"... Anyone have any thoughts?
• The bug spray sound on the CM-32L sounds like a whoosh instead of the MT-32's bass guitar sound!
• Fireworks sound like punching instead of MT-32's dinging.
• Horns in main menu song clip badly on MT-32 but not on CM-32L.
• CM-32L walking sound in Darkside intro sounds like very quiet footsteps instead of MT-32's square wave "deet-doot" sounds.
• Can hear doors opening on CM-32L, but not at all on MT-32!
• Walking sounds much more distinct and realistic on CM-32L (MT-32 has the same "deet-doot" in-game as in the Darkside intro it seems)
• Blacksmith's hammer sound is much more realistic on CM-32L
• The one thing I don't like as much on CM-32L is that one set of horns almost drowns out the other in the blacksmith's shop music, while it sounds slightly more balanced on MT-32.

SC-55 (non-MkII) versus SC-88 (non-Pro, in SC-55MkII + Init GS mode):

• Organ sounds in Darkside intro have weird stabby/piercing peaks on SC-55.
• Hammer sound has a sort of echo or double-ringing on SC-55.

It's really looking to me like this game may be "optimized" for:

• CM-32L over MT-32
• SC-55MkII over SC-55
• GS over GM

Update on the "backwards" MT-32 issue:

Compared the MT-32 and SC-88 manuals and it looks like the MT-32 does implement the MIDI pan controller values backwards from SC-88, so it just depends on whether the game composers were paying enough attention to know better. I believe that Mau's CM-32L recording has it the correct way, as the bug spray is on the right side of the logo screen: http://www.youtube.com/watch?v=6vLSoU0E07M

For reference, here is the MT-32 manual's pan implementation info:

Panpot: Status: BnH, Second: 0AH, Third: vvH vv = Panpot Value 00H - 07FH ( 0 - 127 ) n = MIDI Channel 0H - FH ( 1 - 16 ) Orient […]
Show full quote

Panpot:
Status: BnH, Second: 0AH, Third: vvH
vv = Panpot Value 00H - 07FH ( 0 - 127 )
n = MIDI Channel 0H - FH ( 1 - 16 )
Orientation of sound is as follows.
127 = LEFT, 63 = CENTER, 0 = RIGHT

And SC-88:

Pan (Controller number 10) Status: BnH, 2nd byte: 0BH, 3rd byte: vvH n=MIDI channel number: 0H - FH (ch.1 - ch.16) vv=pan: 00H - […]
Show full quote

Pan (Controller number 10)
Status: BnH, 2nd byte: 0BH, 3rd byte: vvH
n=MIDI channel number: 0H - FH (ch.1 - ch.16)
vv=pan: 00H - 40H - 7FH (Left - Center - Right) Initial value = 40H (Center)

In other words, increasing values from 0-127 will pan right-to-left on MT-32 and left-to-right on SC-88.

Time to swap my MT-32 back to the original setting, oops!

is there still a need for a LAPC-I recording? AFAIK the CM32-L is the same as LAPC-I ... say the word and I'll record!!!

Not having any experience with LAPC-I or CM-32L, I'll leave it up to you guys to decide whether it's worth it.

I have a SC-55 MKII as well. Can provide one tomorrow if you are really keen. What do you want me to initialize this unit as? Or just leave it default and listen for the bugs?

I take it you want to compare a SC-55MKII with your SC-55 setting on your unit? Can do, just let me know what I need to initialize it so it's the right way 😁

The clipping might be recording related. I take care when I record that it never clips. Audacity tells you if there was a clipping event and then I start over again with a less sensitive recording setting.

Regarding Stereo channels. I have 2 programs for testing. Wing Commander: Barracks screen, bucket water dripping from left speaker.

Doom first track at the beginning drums from the left, guitar on the right.

Otherwise just insert headphones and compare to your speakers.

AFAIK many Sierra games have Stereo reversed, so they aren't the norm but sometimes they need revering the cables for some of their games...

keropi wrote:

is there still a need for a LAPC-I recording? AFAIK the CM32-L is the same as LAPC-I ... say the word and I'll record!!!

Yup they behave the same, although the LAPC-I is apparently a little bit nosier due to being inside a PC...

either way, more like a test in recording I uploaded a 192kbps mp3 here: http://www.multiupload.com/M0YL5P4WG7

you hear the whole intro (some parts have lower volume since there is speech that you cannot hear there) , the blacksmith then the tavern and finally some outside music... the doos open/close sounds are lame at least... they SUCK. 🤣

keropi, that sounds quite a bit different than both the MT-32 and CM-32L recordings to me. The Darkside intro music sounds kind of hollow but clean, like the reverb is missing or greatly reduced? The title screen music sounds like one or more of the instruments are messed up somehow, like maybe the SysEx didn't all get through? The blacksmith hammer sounds good, more like CM-32L than MT-32.

well I didn't do anything special, just recorded... I played gods before though with roland support... maybe it messed it up? will make a fresh-boot intro recording this time I will look on the xifi panel, I think I have the crystalizer on... this time the track is a little different as there is a save present and you don't see the cutscene... it goes straight to the title screen , then kill the slime, enter blacksmith then the tavern and lastly the trainer...
some notes: it seems lapc-i was buggered from Gods. Also before I go to the blacksmith you hear a really nice creaking sound from the door and footsteps. After you hear the blacksmith song those fx vanish forever... the game does not behave that well, so I did a 2nd recording after starting the game again... 🙁

http://www.multiupload.com/XW6EFUFCFM

Pulled (another 😵 ) all nighter...

XEEN recording with SC-55MK2

I did a "Init GS, Sure?" before running your MIDI file. Sounds quite the same to me, but maybe you guys spot something as I'm not familiar with this game / music...

keropi: That second LAPC-I recording indeed sounds almost identical to Mau's CM-32L to me.

Mau: That sounds just like your SC-55 recording as far as I can tell. Still has something weird with the organs in the Darkside intro part (around 2m00s - 2m30s in the various recordings). What program are you using to play the MIDI file to your synths?

Trusty Media Player! Been using it for ages. I know it works 🤣

Cloudschatze wrote:

You may want to see this thread...

http://queststudios.com/smf/index.php/topic,2483.0.html

HunterZ wrote:

Thanks for the link. Interestingly, I don't get buffer overflow errors on my real MT-32 (1.07 ROM I think) with DOSBox 0.74 - although I do get a Exc checksum error.

I also haven't seen the NO INSTRUMENT error on my SC-88. Maybe the SC-88 implements the capital tone fallback feature that you mentioned? I haven't tried that thing with firing an arrow though; I'll try to remember to test that.

Hi,

The instrument variation concept of the SC-55 and SCC-1 is clearly explained in the SC-55 manual (page 42) : http://media.rolandus.com/manuals/SC-55_OM.pdf

• On the SCC-1 and SC-55 : when an instrument variation which doesn't exist, is selected, it will fallback to an existing instrument variation, or else the capital instrument will be substituted.
• On the SC-55mkII and SC-88 : Roland has removed (or had to remove ?) the "variation fallback" feature, starting from the SC-55mkII, and when an instrument variation which doesn't exist, is selected, a "no instrument" error message will appear on the LCD display, and there won't be any playback performed on the affected channel(s).
• On the SC-8850 and Virtual Sound Canvas : when an instrument variation which doesn't exist, is selected, it will be even worse, because it will keep the current instrument and perform the playback with it, resulting in a wrong playback, since it may be using a different instrument than the one intended.

Now, being a fan of this great "Might and Magic : World of Xeen" game, especially thanks to its lovely musics, I've always been annoyed by this MIDI sound effect problem, happening on my SC-55mkII, when firing an arrow.

Then, since I've got knowledge in MIDI coding, a few years ago, I've modified the DOSBox's MIDI code, to finally be able to play this game without any problem, using my SC-55mkII.

Also, I intended to submit this fix to the DOSBox author, but since it's a fix useful for one game only, and also only if using either a SC-55mkII, SC-88(Pro), SC-8850 or Virtual Sound Canvas, I thought it wouldn't be included in the official DOSBox build anyway, so I didn't bother to contact the DOSBox author.

However, now I've realised that it would be selfish to not share this fix with other people who may be interested in it, and then, I'm finally sharing what I've done, on this forum.

So, to understand how my fix works, let me first remind you, that to use an instrument variation, you first send a "Bank Select" message, and then a "Program Change" (instrument change) message : the "Bank Select" processing will be suspended until a "Program Change" message is received.

Now, what my fix is doing, is scanning absolutely all of the MIDI messages being sent to the MIDI device, and when a "Program Change" message has been sent, the fix will simply immediately send to the same channel, a "Bank Select" message with a value of 0 (to select the capital instrument).

Since the "Bank Select" message sent by my fix will be suspended until a "Program Change" message is received, it won't change the current instrument selection being used, whether it's a capital instrument or an instrument variation.

Then, when the game will change the instrument, if it's first sending a "Bank Select" message, before sending a "Program Change" message, then the "Bank Select" message sent by the game will replace the suspended "Bank Select" message sent by my fix, but if the game is only sending a "Program Change" message, then the suspended "Bank Select" message sent by my fix will finally be performed to select the capital instrument, avoiding to select an instrument variation which doesn't exist on the MIDI device.

I've attached to this post, a Windows executable, the modified "midi.cpp" source file, and also a ".diff" file.

Now, if you would like to see this fix included in the official DOSBox build, you can contact the DOSBox author : if he receives many requests from many different persons, he may consider including it for the next official DOSBox release.

Also, if this fix is included in the official DOSBox build, it could be enabled with a new option in the DOSBox configuration file.

Finally, is anybody aware of other games having a similar problem, and then for which this fix is also working ?

TURRICAN0:

Thanks for sharing that info and DOSBox patch! I put down Xeen not long after last posting in this thread because it was a bit too cheesy to keep me interested, but it's good to know what's going on and that there's a workaround available thanks to your efforts.

I would also guess that the DOSBox developers would not be interested in incorporating the patch, but there is a healthy community of DOSBox users who like to incorporate unofficial patches into unofficial builds. I'm sure your efforts will be appreciated by people one way or another.

TURRICAN0, Munt emulates cm32-l/lapc-i perfectly so no need for your patch 😀
(you can test that in latest ykhwong's dosbox build)

So for the sake of argument, will any of these external Roland devices play World of Xeen correctly at default with all correct sounds present? I mean with no modifications?

1. MT-32 1st gen
2. MT-32 2nd gen
3. SC-55
4. SC-55MKII
5. CM-32L
6. CM-64
7. CM-300
8. CM-500