Hi,
Sorry, I have not noticed that you had not used ideal command line parameters. To get the best result for your goals you should use
..\MidiPlayer5\mod2midi.exe ...\mod\agony\agonyintro.mod ...\mod\agony\ 0 1 1
That is split by (mod)track AND by instrument. This way you should not get more channels per track that is possible with only 'by instrument' splitting ('by instrument' collects one instrument to one midi track regardless of mod tracks/channels).

And I almost forget that FSMP also has already got an 'Optimize' feature in the Event Viewer/Debugger.
1. Use the '0 1 1' command line parameters to create your mid+sf2 pair.
2. Open the converted midi in FSMP.
3. Open the Event Viewer/Debugger (bottom shortcut icon on the right side of player)
4. Make sure 'Reflect selected value' is checked and select the first Program change event on track.
5. Now deselect 'Reflect selected value' (it's not necessary but prevents selecting an other event accidentally).
6. Right click on events pane and click 'Select to End' on context menu (*or better select the range that you know cannot conflict with other tracks that use the same channel).
7. Now click the 'Optimize' button.
8. Repeat above steps on all tracks

Finally save the edited file by clicking File menu -> Save File as.

*Note: You cannot do optimize blindly! There can be more tacks that use the same channels. Eliminating all redundant program changes on a given track can influence other tracks that use the same channels. Unfortunately the differences between tracker and midi philosophy cannot be eliminated fully. If you can make sure that in your DAW each track also represents an independent channel/independent virtual instrument (regardless if the channel reference is the same on 2 or more tracks) then this is not a problem.

HI.
Thank you for your explanation. I've done exactly as described but as you was writing this does not work.

I'm not understanding one thing: If each track created used a by an instrument... and in the mod i've 4 instrument for example why if i remove redundant program changes, even with the original soundfont don't work as you have predicted? I mean: if a track have to use only that instrument... why i'm getting different sound?

You told me about channel, is it possible for a track to not use a different channel and is not possible to have the channel amount as the mod instrument count and multiple note on a track if that's thew case? I don't know If i've explained what I'm trying to mean... I'm going to test other mod.

Editet: I think i've understood: this not work because i've 4 channel and 12 trrack... but what if I can edit channel too? Is this possible with FSMP. OR If i've created by modtomidi 12 track and 12 channel? I think in this way until 16/32 channel i'll have no problem. Or maybe... have multiple track but having the same mod instrument which will use the same channel? Or even trough your editor give the chance to change the channel for the whole track to let me do some test. I'll really appreciate. 😀

Thank you.

Aldo

RetroGC wrote:

HI.
Thank you for your explanation. I've done exactly as described but as you was writing this does not work.

I'm not understanding one thing: If each track created used a by an instrument... and in the mod i've 4 instrument for example why if i remove redundant program changes, even with the original soundfont don't work as you have predicted? I mean: if a track have to use only that instrument... why i'm getting different sound?

Unfortunately I cannot explain this much more clearly than above. Please, read about the difference between midi tracks and midi channels in specification. But the point is:
A 'Track' is only an organizing concept in SMF (Standard Midi File) files and totally independent from channels that are part of the Midi protocol. E.g. You can have 65536 tracks in a type 1 midi file that still can use only 16 channels. And type 0 midi files can only use 1 track but can use all 16 channels in the only 1 track.
Short midi messages have a channel context but long midi messages (SysEx) are channel independent. But neither short nor long midi messages has a 'track' context. It's the beauty of Midi 😀 So even 2 separate tracks are not fully independent from each other if they use midi messages that use the same channel.

RetroGC wrote:

You told me about channel, is it possible for a track to not use a different channel and is not possible to have the channel amount as the mod instrument count and multiple note on a track if that's thew case? I don't know If i've explained what I'm trying to mean... I'm going to test other mod.

I don't know exactly what you mean but hopefully I can explain why it's working the way as it works now. In modules mod command's (volume, slide, etc.) context is a mod track. The analogy of this context in midi is channel (not track). So mod track commands can be converted to channel messages during conversion. That's why a 4 track mod will use 4 channels but can use more or less tracks depending on track organization. Also that's why the conversion actually only work up to 16 track mod files. Beyond 16 mod tracks the conversion is problematic due to 16 channel restriction of midi. Further problem is that mod note commands are self describing and do not have a 'state' like midi note messages. Every mod note command also contains what instrument should be used to play the note. So usually a mod track use many instruments. Sometimes every note of the same instrument in midi terminology (e.g. a piano) is a different instrument in mods! So instrument to instrument mapping of mods and midis is nearly impossible. And that's why you see so many program change messages in the converted midi.

Anyway I have experimented with your mod in Reaper. With some tinkering instruments are replaceable but there is the problem I have written about before:
In case of MODs sometimes instruments are not 'in tune' in respect of absolute pitches.
At first I had to set my soundfont's piano one octave down to get the same pitch. But later as you can hear you will get horrible out of tune notes. These notes can only work with the custom tuned samples of the converted SF2 file.
I have made a vide about it:
https://youtu.be/R9Xp4KDM00w

So overall I do no think general replacing methods that work with every mod exist (no matter how I would alter the conversion and track representation). And forgive me but I feel I have already sacrificed more of my free time than I wanted to this subject (since in my opinion the end result can not be satisfactory). The default conversion that preserves the most aspect of the original mod is the best as far as FSMP's mod playback ability is concerned and it works well.

Falcosoft wrote:

RetroGC wrote:

HI. Anyway I have experimented with your mod in Reaper. With some tinkering instruments are replaceable but there is the problem […]
Show full quote

HI.
Anyway I have experimented with your mod in Reaper. With some tinkering instruments are replaceable but there is the problem I have written about before:
In case of MODs sometimes instruments are not 'in tune' in respect of absolute pitches.
At first I had to set my soundfont's piano one octave down to get the same pitch. But later as you can hear you will get horrible out of tune notes. These notes can only work with the custom tuned samples of the converted SF2 file.
I have made a vide about it:
https://youtu.be/R9Xp4KDM00w

Just Listened... That's true. In this case there are different tunes, so piano would not be possible to repleace. I knew that it was different piano samples, as different instrument effectively i selected that song because i tought it was only a piano (for 3 tracks) and a string to replace...

So, it would not possible to have a conversion with each mod instrument, using a single track and a single channel dedicated (until 16) avoiding conversion for more than 16? I knew that you have dedicated a lot of time... and I relaly appreciate your effort.

Thank you

Aldo

New features added to test version:
http://falcosoft.hu/midiplayer_54_test.zip
1. Volume ladder indicates clipping.
2. Bassmidi mode supports 64 (4x16) channels (midi file has to use Midi port preference (0x21) meta event in tracks to designate ports 0..3) .
3. A special low latency mode in case of WASAPI output is available (Buffer has to be set to 10ms in Device settings dialog). In this mode Bassmidi buffering is fully bypassed so ASIO like performance can be achieved. It's especially useful for real time jamming with Midi keyboards.
4. When started with '/tray' command line parameter pressing the close button minimizes FSMP back to system tray. You can fully close FSMP with the help of right click context menu of system tray icon or selecting Exit in main menu.
5. More than 1 instance is not allowed to load to tray with '/tray' command line parameter. The second instance is loaded normally.

OK, I'm loving the last few improvements. These tiny tweaks are making a pretty big difference for me. Keep up the good work man 😄

So far I haven't run into any new issues with this test build, I haven't been able to get the low latency mode to work yet though. My audio gear probably just doesn't properly support wasapi.

I've found that the clipping issue I was running into before doesn't happen to me when I run it in WINE, still trying to figure that bug out. I'll let you know if I find anything more conclusive.

t9999clint wrote:

So far I haven't run into any new issues with this test build, I haven't been able to get the low latency mode to work yet though. My audio gear probably just doesn't properly support wasapi.

What happened exactly? No sound at all? Make sure you selected WASAPI as output mode instead of Directsound since it cannot cope with 10ms buffer (no special low latency mode can be implemented in case of Directsound) . I have tested it with Audigy 2 and also with integrated Realtek/VIA/Soundmax HD audio and all of them worked so I don't think special gear is required.

HI.
I'm trying 5.4 version too and I've just a question: i'm loading midi file with custom soundfont (named as midi file). I can see in the grid the soundfont loaded but during the playback the software is not using the soundfont custom for that midifile but the general soundfont.
Thank you for your help.
Aldo

Falcosoft wrote:

t9999clint wrote:

So far I haven't run into any new issues with this test build, I haven't been able to get the low latency mode to work yet though. My audio gear probably just doesn't properly support wasapi.

What happened exactly? No sound at all? Make sure you selected WASAPI as output mode instead of Directsound since it cannot cope with 10ms buffer (no special low latency mode can be implemented in case of Directsound) . I have tested it with Audigy 2 and also with integrated Realtek/VIA/Soundmax HD audio and all of them worked so I don't think special gear is required.

I just tried it on my gaming rig and it worked great there.
My laptop was the my only windows machine that I had on hand at the time of the last post. I was trying it with Intel's HDMI out and it didn't seem to like it at all, it just stuttered and was a broken mess.
Looking back at it, it's possible that I didn't update MidiPlayer with the latest test version. I will try this again and let you know.

RetroGC wrote:

HI. I'm trying 5.4 version too and I've just a question: i'm loading midi file with custom soundfont (named as midi file). I can […]
Show full quote

HI.
I'm trying 5.4 version too and I've just a question: i'm loading midi file with custom soundfont (named as midi file). I can see in the grid the soundfont loaded but during the playback the software is not using the soundfont custom for that midifile but the general soundfont.
Thank you for your help.
Aldo

Have you tried adding _000 to the end of the soundfont name?
Example:
2-12-Fithos_Lusec_Wecos_Vinosec.mid
2-12-Fithos_Lusec_Wecos_Vinosec_000.sf2

Here's what I was able to get with 5.4 and this method
https://downloads.kor.ninja/Music/2-12-Fithos … cos_Vinosec.mp3

t9999clint wrote:

HI. Have you tried adding _000 to the end of the soundfont name? Example: 2-12-Fithos_Lusec_Wecos_Vinosec.mid 2-12-Fithos_Luse […]
Show full quote

HI.
Have you tried adding _000 to the end of the soundfont name?
Example:
2-12-Fithos_Lusec_Wecos_Vinosec.mid
2-12-Fithos_Lusec_Wecos_Vinosec_000.sf2

In this way this works. I didn't remember it was necessary to set _000.
Thank you.
Aldo

RetroGC wrote:

In this way this works. I didn't remember it was necessary to set _000.
Aldo

It's not necessary, you can also set the default bank used with real-time loaded soundfont in Device Settings dialog 'Real Time Loaded Soundfont -> Default Bank'. In your case seemingly 0 is required to set.
I have already explained the whys many times before:

By default FSMP uses the convention of original Sound Blaster demos where variation banks with the same name as the midi file al […]
Show full quote

By default FSMP uses the convention of original Sound Blaster demos where variation banks with the same name as the midi file always used Bank 1. The reason could be in case of early SB hardware Bank 0 was in ROM and/or could not be re-loaded by programs because of sfman32 limitations.
And many original SB demos relied on a default GM bank at Bank 0 besides the custom variation bank.
Midi+SF2 files converted from Mod/Xm/S3m also uses the Bank 1 convention.
But newer soft synths has no such limit and newer same named midi+sf2 combos do not follow the old convention and requires SF2 to be loaded at Bank 0.
SB_SF2_DEMOS.zip:
download/file.php?id=39958
(the demos require that 'Instruments -> Autoload Sondfonts' to be enabled and reset type set to GS mode)

RetroGC wrote:

I can see in the grid the soundfont loaded but during the playback the software is not using the soundfont custom for that midifile but the general soundfont.

You can also see in the grid to what bank the soundfont is loaded. If you see in FSMP's 'Bank MSB' control that the instruments do not use the same bank you can know there are some problems (either with the midi files or the designated bank)

HI. while i was trying to recording System Shock Soundtrack, i've noticed that there's a midi file that won't work if I override loop settings to 1. It works fine if i set loop but if i remove override loop and it play correctly until the end.
Aldo

Falcosoft wrote:

in the grid to what bank the soundfont is loaded. If you see in FSMP's 'Bank MSB' control that the instruments do not use the same bank you can know there are some problems (either with the midi files or the designated bank)

Thank you.

RetroGC wrote:

HI. while i was trying to recording System Shock Soundtrack, i've noticed that there's a midi file that won't work if I override loop settings to 1. It works fine if i set loop but if i remove override loop and it play correctly until the end.
Aldo

You use a converted version from XMI and it seems the conversion resulted in overlapping loop points. Multiple overlapping loop points are not supported currently. Other players does not support multiple loop points at all so maybe that's why it remained unnoticed. The original XMI version does not have this problem. I have attached the fixed version of yours and also the original XMI version so you can inspect the differences/problems of your converted version.

Thank you. I've not converted myself, just downloaded from mirsoft.
Aldo

RetroGC wrote:

Thank you. I've not converted myself, just downloaded from mirsoft.
Aldo

I have added workarounds to avoid problems with overlapping and orphan loop start/end points. I could only test it with your uploaded problematic file, so please test it with others and report back.
http://falcosoft.hu/midiplayer_54_test.zip

I know Midi Player is mainly for playing MIDI, but given that you can already easily edit a few things with the Event Debugger, I think it would be great to get more functions there.

Here are a few suggestions:
1. A replace funtion. (e.g. replace "pan" with "volume" for a certain channel)
2. Certain MIDI Types (I think 0) often only has Bank MSB and LSB entries for channel 1. Could the program automatically add those if they are edited and don't exist?
3. Add events to any position you want them (instead of only being able to replace existing ones)
4. An automatic incremental filler
What I mean with the latter is: The user can let's say put in volume 0 for a channel at 1-01-000 and volume 100 at 2-01-000. You can then select those 2 and let the program add those values automatically adjusting the time so it will increase evenly over time until it reaches the given positions and values.

I know I could maybe use other software to do that, but I only found software that doesn't "natively" use MIDI anymore. They often use their own format and the user has to import and export MIDI, instead of being able to save it directly as MIDIs. And they often extremely complicated in my opinion. Unless you have any sequencer you can recommend?

Anyhow, keep up the good work.

Hi,
Of course more functions are always better, but I have limited time and my focus is elsewhere since not many people use the Event Debugger for serious editing. The main purpose behind the debugger is to help troubleshooting bugs/problems. A full midi editor would require much more work. And as you said there are many professional editors/sequencers for these jobs. Anyway you can easily achieve what you want with e.g. MidiPlex. It's a full featured open source low level midi editor:
https://github.com/stascorp/MIDIPLEX

hello

i have a small problem with falcosoft midi player 5.3 and roland sound canvas va 1.13

with this midi file here
https://www50.zippyshare.com/v/eQL1Ypnq/file.html

the instruments are wrong.
sound canvas va shows the correct names but only piano is played on all tracks except drum track.

(i use the custom patch names for sc va and capital tone fall back)

after playback starts i have to press the reset button in fs midi player to get correct instruments
(i am not 100% sure but previous sound canvas va 1.11 didn't have this problem)

but cakewalk music creator 7 plays the midi file with correct instruments

hope this can be fixed, or is there a way to send a gs reset before playback starts ?

thx

edit: i found the solution 😀

in "sysex options" i have to uncheck : "enable sysex in files"

other normal gs files worked fine with the option enabled
and other sc88 files from roland have no problems as well
guess its this sc88 file....

disable the option completely seems also not the best solution for other gs midi files.....
because reverb and chorus settings are not loaded properly.