MML: What are people's actual complaints with the damn thing

[Arkhan]

All this rambling about HuC/MML/Mod players has me wanting to ask some questions, as a survey.

So

1) What actually scares you about MML?
2) How is MML any scarier than trackers, which are just columns of bullshit that look like a bad session of VIM that happens to make sounds?
3) Are you aware that you can use any fancy-as-piss DAW to craft PCE songs, and then simply MIDI-->MML---> Run it through Squirrel (for example) to get it to play something on PCE?

I am just trying to sort out why people run to the hills or completely give up on it.   It get's such a bad rap for no reason, I think.   It's not isolated to these scene.   There are people everywhere who go OMFG MML JESUS f*ckING CHRIST, NO.   GIVE ME TRACKER.

Also, perks to using MML for PC Engine:

1) If you did everything in a DAW, you accidentally made a CD soundtrack too probably
2) If you're using Squirrel, you get the magic of sound effects.  See: Atlantean.   It's got chiptunes and SFX.
3) It's super usable for creating a game.



This is my workflow for creating music for PCE and MSX:

1) I open up FruityLoops
2) I use the mouse cursor to pet the cute girl that is my wallpaper in FruityLoops
3) If PCE, I make a 6 channel song.  If MSX, I make a 9 channel song.
4) If PCE, I use Chip32 to approximate instruments.  If MSX, I use an MSX PSG VST, and an OPL2 FM VST.
5) I make the songs with MIDI controllers/piano roll editing.
6) Ok cool, it's done.  I hit the "midi" button and export a MIDI
7) import midi into 3MLE.  Every channel is conveniently converted to MML
If the songs are long, I use 3MLE's optimize button to shrink em up.
9) Paste into Squirrel file, or Musica file for PCE and MSX respectively.
10) Hit a button or two.  Done.

I made YouTubes about this before.   It's stupidly simple.

I really want to basically dispel all these myths about MML, and get you goons to use it and get somewhere with tunes.

[ccovell]

1) It looks like Brainf*ck.  You say it's meant to be created by DAWs, but it's still a textfile, which itself makes peering into it (and recoiling in horror) an inevitability.
2) Columns are the operative word, a very good thing.  Not every MML editor I've seen offers columns of notes that are editable in those columns.  Trackers' emphasis on hex values for effects is bullshit indeed.
3) I'm not a musician, so never used one.  No comment.

[touko]

More or less like ccovell said .
A deflemask user can make music for any systems supported by the tracker,and it's obviously easy to find some deflemask musicians than MML .
I'am not a musician too, and i don't care if musics are mml or tracked, i only need musicians .

For now my fav goes with MML because there is a functional driver.

[Arkhan]

1) It looks like Brainf*ck.  You say it's meant to be created by DAWs, but it's still a textfile, which itself makes peering into it (and recoiling in horror) an inevitability.

But, you shouldn't be creating the song in MML.   You can, but that would require you to be the kind of composer that is capable of sitting with a pencil and a blank sheet of staff paper to compose songs.

It was designed for people who at least understand sheet music.  It wasn't really designed for you to just start smacking at it until it sounds like music.

2) Columns are the operative word, a very good thing.  Not every MML editor I've seen offers columns of notes that are editable in those columns.  Trackers' emphasis on hex values for effects is bullshit indeed.

The MML file can be formatted however you want.   You can (and should) define patterns, and then stamp them down in each channel, similar to what you could do in a tracker.   The Atlantean music files are very tidy like this.

You don't have columns.  You have channels broken up by headers.  You can even put commentblocks to separate channels.   Every pattern that you define can be named whatever you want.

I'd go so far as to say if you do all of this, the file is clearer than looking at a song in a tracker.

Running through a converter produces some goobery looking stuff, but you can do a top-down sweep and reformat it pretty simply to make it readable. 

3) I'm not a musician, so never used one.  No comment.

I think this is the most important trait of people who are pretty averse to MML.   You're not musicians.   So, you probably can't read sheet music, which then implies MML is going to be gibberishy too.

To put it bluntly, the problem is you guys, not MML.   You could manage MML with a little bit of effort, though, I think.


http://3ml.jp/

^^^^ This program has tabs for each channel.  It shows a visual piano roll representation of the MML that you type into the editor.

It has a play button so you can hear it.

I honestly think this is more feature-full and intuitive than a tracker.    You get that visual representation instead of columns of nonsense that happen to make a eurowank demotune if you hit play, lol.

It seems one of the big scaries is that people want live-note-editing and immediate results.

You can get that with 3MLE, or this goofy thing:

http://benjaminsoule.fr/tools/vmml/

BUT BE WARNED: THIS TOOL HAS THE OCTAVE SHIFT COMMANDS BACKWARDS FOR WHATEVER REASON.

I grew up using trackers and crap like Octamed to make music.   Once stuff like FruityLoops came out though, I stopped, because the Amiga software was always a bit clunky and unintuitive.

but, here's the funny thing.

You can use a tracker, and still use MML to get music going in your games.

FOUR songs from Insanity are old MODs I had on a floppy disk that I made on an Amiga 500.   I transferred them to PC, opened them in MODPlug, exported MIDIs, and turned them into MML....

[DarkKobold]

I hope none of this is 'inspired' by Catastrophy. Our problem has been finding a musician willing to work with us at all. We found one of Gredler's friends, who seemed unreliable, and lazy. Gredler was just trying to give him something to do.

We found the musician for Haunted Halloween 85 and 86, but he wanted cash up front.

Cabbage has been doing a few sound effects/midi conversions on the side, but he has his own projects. Catastrophy isn't a priority to him.

So, MML is not our problem. Interested musicians are.

[Arkhan]

I hope none of this is 'inspired' by Catastrophy. Our problem has been finding a musician willing to work with us at all. We found one of Gredler's friends, who seemed unreliable, and lazy. Gredler was just trying to give him something to do.

We found the musician for Haunted Halloween 85 and 86, but he wanted cash up front.

Cabbage has been doing a few sound effects/midi conversions on the side, but he has his own projects. Catastrophy isn't a priority to him.

So, MML is not our problem. Interested musicians are.

Catastrophy didn't spark this.

People have been flailing and afraid of MML since before Insanity was finished, lol.

I am just curious what people's actual problems are, because some of them might be a tad unreasonable, or completely fixable with a bit of effort.


I'd like to help people get more into the swing of things, because it's pretty frigging useful.

What kind of music do you even need for Catastrophy.   I can do it lol

[Bonknuts]

I think things being "intuitive" or not, is kind of irrelevant once you have experience with whatever procedure or toolset. What matters, is efficiency or the perception of it. What matters is the learning curve. And what matters is the community behind the toolset. There's a large community behind trackers, so that's a valuable resource. If the community support is there, then things such as asking questions, figuring out techniques, and looking at other people's work. That's what matters. There is not one concept or idea that fits all.

 Also, this:

You claim this is stupidly simple, but look at all the conversions you have to do, and things you need to check and verify. You know what's simpler than that? Not having to convert a damn thing. If I were to use Deflemask right now - then that's it. There's no conversion processes. No avoid this, or fix or change this in during step X, etc. Not only is it one and done, it has ALL the advantages of a tracker interface - "what you hear is what you get".

 People get comfortable with what they know, and don't like to change something that already works for them. And add to the fact that PCE really isn't in demand as a development system, don't except people to go out of their comfort zone just to produce something on this target platform.

 Until you make a program interface that has NO conversion process, and has "what you hear is what you get" - one step process, then you're not providing a competing product on the same level. I don't know why that's so difficult to understand.

 The fact that your "player" and Squirrel app doesn't support "samples", is a pretty big disadvantage IMO. It's one thing not to use samples, it's an entirely different thing not to have support for it.

 And I'm with touko: I really don't care what the format is, as long as it can produce whatever PCE devs have done so far - then who cares. But knowing chiptuners, at least relative to the PCE era stuffs, you need to know your audience. Rant and rave all you want, and trust me I agree with most of the philosophy of your rants on this subject, it won't change the fact.

 You want people to make music on the PCE? Make something like Fruity Loops that directly outputs PCE music files. A program that is ONLY specifically PCE. A program where "what you hear is what you get" type interface. Because chiptune musicians, the good ones, are always tweaking the sound to get something interesting or specific, and down to even just one part of a song. Visual presentation of this is important. Dicking around with MML output, is not its equivalent. 

[Gredler]

I am not a musician, so it's hard to explain why I could not get a great squirrel song, as I generally can't get a great song anywhere else either, so my complaints should be chalked up to inability to compose music and lack of knowledge on the subject entirely.

The complaints I've gotten from the friends who tried to make music are probably less to do with squirrel specifically but making actual chip-driven-tunes, but I'll try to convey their trepidation's as best I am able.

1) The gap between in editor instrumentation/sound and in game.
  The two guys that I've asked to help have had issue finding a VST or instrument that is similar to the squirrel preset wave forms. The youtube video (

) is appreciated but ultimately not very useful for finding a consistent sound between authoring software and the rom playing on emulation/hardware, especially considering how different each envelope and octave shift affects the sound of waveforms.

2) Timing monitoring and beat/bar/note consistency
  The guys also had issues with their songs getting out of sync due to dropped notes and or note lengths. The songs always spiraled out of coherence and would "break". If there is anything that I was happy with about my song I made with squirrel it's that I was able to overcome this by using macros/bars that I could loop. It still gets out of sync, I think I am 1/4th a note off somewhere, but it was very methodically planned and created straight in mml. I have not been able to make a good song converting a midi to mml because of this, the only "decent" music I've been able to get out of squirrel is by hand coding the mml.

3) Getting a midi that was compatible with a clean conversion to mml.
  The midis we've found royalty free have been so outside of the 4-6 channel midi requirements that I was not able to get anything to sound even half decent (also due in part to the above two issues, the songs even when truncated to 4 channels would eventually sound way off in instrumentation, and in timing). Without creating a midi with converting it to mml in mind, in a specific application (which so far only seems like 3mle and fruity loops are confirmed as working tools for this), it seems very difficult to recreate midis in squirrel as mml. Perhaps if I worked with a musician who could create a specifically planned mid that we could pipe through 3mle and "spot check" like you mention in the tutorials, but without knowing the sheet music/what I am looking for, converting from a random midi to mml looks like converting Russian to Portuguese - I can't understand either, so spot checking is impossible because I don't know what I am looking for.

Like I said - I am about as far from a musician as you can get and still be attempting this (DK is probably further off, but refuses to even attempt it) - so many grains of salt have to be taken with my issues with squirrel. 

1) I open up FruityLoops

This might be the root of all evils. I'm a cheap (read: baf) guy and have not wanted to drop $100-200 on a music creation software sweet I would barely know how to use, but maybe if I could create midis with this all of our troubles would be over. Buying stuff makes you better at things right?

[Arkhan]

You want people to make music on the PCE? Make something like Fruity Loops that directly outputs PCE music files. A program that is ONLY specifically PCE. A program where "what you hear is what you get" type interface. Because chiptune musicians, the good ones, are always tweaking the sound to get something interesting or specific, and down to even just one part of a song. Visual presentation of this is important. Dicking around with MML output, is not its equivalent. 

I think the expectation that someone in the scene write some sort of full DAW-esque program specifically for PCE is both selfish, and borderline stupid, and misses the point of what I said completely.   

Writing something like that is a MASSIVE undertaking and realistically would require a small team, and to what end?     

It's not simple.  People have been banging away trying to do similar things on all kinds of machines for years.

The closest to successful was Prophet64 on the C64, but the player code for that is so resource intensive that it can't feasibly be used for anything but listening.

Also, you can already get a visual representation of a song with MML if you put it into 3MLE.  Once you know where you want effects to be at, placing them into MML isn't really any different than putting them into a tracker.  You do realize this, right?  Dicking around with MML is the same as dicking around with hex columns in a tracker.  Don't sit and tell me a tracker's effects column is somehow better.  It's not. 

On the topic of "chiptune musicians, the good ones", what do you think I do with my stuff?  The deep ass kick drum sound I got wasn't an immediate result.  Nor were the lead sounds in Reflectron, or that weird bass noise I made.

It's not really much different to compose a song, get it playing, and THEN go through and sprinkle fancy shit onto it like slides, vibrato, and such.    You often start talking about sound design as opposed to musical composition.   They are two different things, but you conflate them.   This is an incorrect thing to do. 

Example:  MIDI.   What is MIDI?  It's a bunch of digital data that jams into any listening device and THEY ALL SOUND DIFFERENT.    You compose and end up with a MIDI.   You can sit and fiddledick with sounds all you want once you send it into some kind of synthesizer or software program.    You're going on about step B, missing the point completely about step A.

The conversion process required is hardly that difficult. You make it sound like it's all super difficult and cumbersome.   It isn't.   Maybe it is for someone who doesn't understand music or the software, but it's all pretty commonplace stuff to worry about.   You will have the same kind of issues in a tracker, honestly.  Your drums are mapped to notes.  If you don't make the MML drums to the same notes, you'll have to go update them later. 

Whoop. Dee.  Doo.   

I don't think it's fair to knock something as bad because you may not get it.  It's probably better to try learning it, as opposed to dismissing it, possibly because it's not what all the cool kids are doing.

The thing of it is, you're talking about hobbyish/trackertune scene people and their wants/desires.  When you have a workflow that "has ALL the advantages of a tracker interface", you forgot to mention that it also includes all of the disadvantages of a tracker interface.   Trackers are not the be all end all, despite what apparently everyone seems to think.   I think writing a song in a much more sophisticated program, and then outputting midi, going to MML-->PCE is a much smoother process than trying to compose shit in a tracker.


Why did you put player and samples in quotes, anyway? 

You have to realize, a lot of commercial software had songs written much differently than how you seem to think.

Hubbard wrote songs for C64 games on real instruments, and then converted things over to the C64.   

ALOT of musicians did this.  Mega Man's tunes were made this way, even.  Do you think they had some visual representation of all of this crap?  no.  They didn't.  I would go so far as to say people like Galway, Hubbard, and Follin are "the good ones" you speak of. 

Some of Hubbard's noises were accidents due to bugs in the SID chip.  It wasn't him sitting with some fancy ass program where he jiggled stuff around until he went OH COOL. 

1) The gap between in editor instrumentation/sound and in game.
  The two guys that I've asked to help have had issue finding a VST or instrument that is similar to the squirrel preset wave forms. The youtube video (

) is appreciated but ultimately not very useful for finding a consistent sound between authoring software and the rom playing on emulation/hardware, especially considering how different each envelope and octave shift affects the sound of waveforms.

Yeah.  Chip32 is free, and you can get approximate sounds out of it.  You can also take those waves from Chip32 and make them as custom waves in Squirrel, to get the same sounds out of them...

Once you have the song playing though, it's really a matter of just fiddling with wave parameters and listening to the song again.  Building the ROM and relaunching is slower than getting an immediate playback, but, it's... not that bad.   Maybe an extra second or two. 
 
 

2) Timing monitoring and beat/bar/note consistency
  The guys also had issues with their songs getting out of sync due to dropped notes and or note lengths. The songs always spiraled out of coherence and would "break". If there is anything that I was happy with about my song I made with squirrel it's that I was able to overcome this by using macros/bars that I could loop. It still gets out of sync, I think I am 1/4th a note off somewhere, but it was very methodically planned and created straight in mml. I have not been able to make a good song converting a midi to mml because of this, the only "decent" music I've been able to get out of squirrel is by hand coding the mml.

Did they use 3MLE to verify that the timings were all correct?  Sometimes, all that is needed is adding rests to finish off a measure so things stay in sync.  If someone has an example, I can show you. 

Othertimes, it's just a matter of importing with a different quantization setting in 3MLE when you import the midi.
 

 
3) Getting a midi that was compatible with a clean conversion to mml.
  The midis we've found royalty free have been so outside of the 4-6 channel midi requirements that I was not able to get anything to sound even half decent (also due in part to the above two issues, the songs even when truncated to 4 channels would eventually sound way off in instrumentation, and in timing). Without creating a midi with converting it to mml in mind, in a specific application (which so far only seems like 3mle and fruity loops are confirmed as working tools for this), it seems very difficult to recreate midis in squirrel as mml. Perhaps if I worked with a musician who could create a specifically planned mid that we could pipe through 3mle and "spot check" like you mention in the tutorials, but without knowing the sheet music/what I am looking for, converting from a random midi to mml looks like converting Russian to Portuguese - I can't understand either, so spot checking is impossible because I don't know what I am looking for.

  lol, yeah, lots of MIDIs have polyphony on channels, so converting them is going to suck.  It's better to write stuff with 6 channels, no polyphony in mind. 

[Gredler]

Yeah.  Chip32 is free, and you can get approximate sounds out of it.  You can also take those waves from Chip32 and make them as custom waves in Squirrel, to get the same sounds out of them...

I have not used any music making tools (only coded directly in mml) but I have suggested Chip32 to the musicians I've talked to. These guys don't seem to understand the correlation between squirrel youtube video's wave #'s, and what they're making in Chip32. One of the three guys I've talked to was using "renoise" and apparently was not able to get Chip32 working with that. Like I abridged above, I will need to bite the bullet and purchase some music creation software if we can't land a musician who can be shown chip 32 and make usable midis

Once you have the song playing though, it's really a matter of just fiddling with wave parameters and listening to the song again.  Building the ROM and relaunching is slower than getting an immediate playback, but, it's... not that bad.   Maybe an extra second or two. 

This is what I suggested to them as well, and how I came upon the music we have now except I was trial and error build and testing with MML programming by hand, since I don't have any music generation tools outside of free 3mle

Did they use 3MLE to verify that the timings were all correct?  Sometimes, all that is needed is adding rests to finish off a measure so things stay in sync.  If someone has an example, I can show you. 

Othertimes, it's just a matter of importing with a different quantization setting in 3MLE when you import the midi.

I don't think any of the guys except for cabbage can do this. I have been able to do it with the royalty free midis I've downloaded, but like you say below here, it's basically giberish because the midis weren't made with 6 channel polyphony in mind . I just need to make some midis and go through the process knowing how it works so I can learn this timing verification process

lol, yeah, lots of MIDIs have polyphony on channels, so converting them is going to suck.  It's better to write stuff with 6 channels, no polyphony in mind. 

Thanks for the help and tips, maybe we can get a musician to chime in and join the conversation

[Arkhan]

If you want help making Catastrophy music, I could help.   How many songs do you need?  What kind of songs?

The timing verification process is pretty easy.  You just have to look at the LAST note for a channel, and make sure it ends on the end of a measure.  Otherwise, it will slowly fall out of sync.  If you find that there's a gap, just add rests until the little white line indicates that you're at the end of the measure.

[Gredler]

If you want help making Catastrophy music, I could help.   How many songs do you need?  What kind of songs?

The timing verification process is pretty easy.  You just have to look at the LAST note for a channel, and make sure it ends on the end of a measure.  Otherwise, it will slowly fall out of sync.  If you find that there's a gap, just add rests until the little white line indicates that you're at the end of the measure.

Thanks for the offer, we're not worthy! We are just a couple guys "curious if we can make a hucard game" for education and fun, no rush or urgency to get it done quickly. We can continue the discussion in a more appropriate thread, but yeah we will want tunes and sfx in the game eventually

I honestly think the issue with MML isn't MML, it's the lack of interested parties - where are all the musicians at? You can't be the only one who wants to make rad games and has elite keyboarding skills

[Arkhan]

I don't know.  They're all using trackers, and waiting for someone to support them, I guess.

Once I finish Inferno I could probably goober something together pretty easily if you tell me about what you want out of it.

lol.  use some of the demo tunes that came with Squirrel for now.   I feel like the Shadow of the Beast example is a good choice.

[Bonknuts]

You bitch and moan about why people are or aren't doing this or that, and when given answers when you ask why - you're response is complain/attack/insult.

 I'm not sure what want. Do you honestly want to know people's opinions and perspectives, or are you just interested in knocking people down or insulting them? It's like there's no compromise with you, unless it involves agreeing 100% with you.

Usually, when you posts stuff like this (one man crusade to rid the world of trackers), I try my best to ignore it and let your rant play itself out. I figured this time you wanted some honest feedback. My mistake. Whatever. Screw everybody else, right? By all means, continue on with whatever this post is supposed to be.

[elmer]

You have to realize, a lot of commercial software had songs written much differently than how you seem to think.

Hubbard wrote songs for C64 games on real instruments, and then converted things over to the C64.   

ALOT of musicians did this.  Mega Man's tunes were made this way, even.  Do you think they had some visual representation of all of this crap?  no.  They didn't.  I would go so far as to say people like Galway, Hubbard, and Follin are "the good ones" you speak of.

Yes, it was all done very differently ... and Martin, Rob and Tim were all assembly-language programmers that wrote their own music drivers.

They knew exactly how the drivers worked, and they could tweak them whenever they wanted any changes.

In those days, the game programmer just received a binary blob with a few documented entry points to call. The musicians kept their code and techniques to themselves.

Oh ... and don't forget ... they were also being paid a living-wage for making those tunes. 

That's just not how the computer-music world, and especially the homebrew game-world operates anymore.

What matters, is efficiency or the perception of it. What matters is the learning curve. And what matters is the community behind the toolset. There's a large community behind trackers, so that's a valuable resource. If the community support is there, then things such as asking questions, figuring out techniques, and looking at other people's work. That's what matters. There is not one concept or idea that fits all.

That is the world of today ... musicians don't write their own drivers. They don't need to. They just want to jump in, learn some reasonable-to-use tools, and get on with making music.

Writing something like that is a MASSIVE undertaking and realistically would require a small team, and to what end

Yep ... to what end?

It's already been done, and it's called deflemask!

Sure, it's far from perfect, but it seems to be meeting most people's needs, and it keeps on slowly getting better.

Just take a look on YouTube for chiptunes made (in the Western world) with MML rather than a tracker of some kind. I found a one, a video by you ... and that was pretty much it.

That was a search for "mml chiptune".

Then I searched for "deflemask chiptune", and pages of hits came up, lots of them for PC Engine tunes.

It really doesn't matter if your workflow is better, or if your results are better ...

... if a homebrew developer wants a chiptune from anyone other than you, then it's probably going to get done in deflemask.

Personally ... until there's a working deflemask player for the PCE (if ever), I'd just recommend that homebrew developers use the CDROM format and either ADPCM or CD-AUDIO for music.