Future of tracker music

[Saga Musix]

But there's 38,149 XM files on Modland. 76,420 MOD files. HVSC has 52,884 SID files so far. You look at the KVR OSC (one synth challenge), and there's only 40 songs per synth that demonstrate a synth's capabilities. Despite the fact every single one is more capable than SID or XM.

There are thousands of songs using those plugins featured in KVR OSC as well. It's just that most people are not going to label their music as being made with those synths. I'm not sure what the point here is. It was fun to participate in KVR OSC but it also showed me very well that I would never want to rely on only having a single synthesizer available to write a piece of music, no matter how powerful it is. This is essentially what you suggested before, if I understood you correctly. A tracker with a built-in synthesizer will only be alive as long as people are happy to use that one synthesizer, and outside of an enthusiastic fan base, that won't be the case for very long. There's simply no one-size-fits-all synthesizer or even sampler engine.

MPTM and XRNS aren't successors because they no longer act as standalone modules, they're just project/session files.

Sorry but that's a gross misunderstanding of the features and capabilities of those formats. MPTM was not (just) made to support VST plugins. MPTM is a logical evolution of the IT format with features such as per-pattern time signatures, custom tunings and lately even OPL3 support - so there you have your built-in synthesizer, even if it's a very old and not exactly powerful one. If you don't use VST plugins, a MPTM file is just as independent as any IT file, but it can make use of features that IT can't.

[zikey]

There are thousands of songs using those plugins featured in KVR OSC as well. It's just that most people are not going to label their music as being made with those synths. I'm not sure what the point here is.

The point is, I cannot use OpenMPT or Renoise the way I want (nor is there anything nice that does what I want). I want a rich, dynamic sound, keeping the size small. Lossy / compressed samples and real-time synthesis/effects. I want the source file to play by themselves, without extra setup for external dependencies.

I like MPTM for what it is, a conservative IT derivative, like how HVL relates to AHX. I'm not trying to downplay your work on it. It contains some of the best module players in any library. I understand the apprehension to do drastic or exotic things like OPL3 or MED instruments. However, those would at least be supported by libopenmpt. VST is out of sync with those ideals and is really just desperation to appeal to a changing industry. MadTracker, Sk@le and Renoise are also guilty. OpenMPT's bleeding edge is really just a mere sequencer that just sends MIDI data to plugins rather than something that can compete or blossom with its own sound engine. And it's held back in a lot of ways - relying on ImpulseTracker's dated 90s sampler. it can only import samples rather than store the original files. So I can't even keep my Opus or GSM files small in OpenMPT.

Renoise has a far more advanced internal sampler and built in FX, but it still is pressured by the music industry, and plays second fiddle to external plugins, external sampler libraries etc. Renoise doesn't even have built in synths. It's sampler engine could easily form the basis of a wavetable/granular synth. And there was nothing stopping Renoise from doing their own plugin format, doing a better job than Buzz did, etc.

I'm not sure what more I can say to explain myself. I don't want to choose between highly restrained chiptune formats or squeaky clean DAW recordings. I want the middle ground that doesn't exist, I want fancy synths in music formats that play in the browser via WebAssembly and xmplay/foobar2k. Something like OpenMPT, but with a more fancy sampler (timestretching, granular) and extensible built-in synthesizer engines. An API/modular system for doing custom synths/effects that get bundled in the module like samples, stuff like that.

Is it really hard to imagine? Look at how good 4klang can sound. libopenmpt can't achieve that fidelity without tons of samples or external plugin installations. Look at PreTracker, AmigaKlang and Cinter. Synths that pre-generate samples, hardly real-time, but still saves a lot of space, and offers a lot of sonic possibilities. Combine that with highly optimized Opus drum samples and CD quality music can be represented under 60kb fingerprints easily.

Basically modern chiptune.  Speaking of chiptune, there's a lot of things you can't quite do, like combine different sound chips. 4mat's FMX combines SID with FM: https://www.youtube.com/watch?v=EbeoLqEpb9s. But it's a combination that no tracker can handle. I'd love a tracker that mixes and matches whatever chip you desire, and still produces small files. VGM format has incredible untapped potential in that regard. But actual soundchips aren't as good as real synths and samplers.

[Midori Mizuno]

I want the source file to play by themselves, without extra setup for external dependencies.

I don't understand your point here at all. All formats created with OpenMPT play without any external dependencies other than libopenmpt (or any other compatible playback engine/tracker. Save for MPTM of course, which is only supported by libopenmpt) as long as they're not using VST plugins, which is a feature exclusive to OpenMPT.

VST is out of sync with those ideals and is really just desperation to appeal to a changing industry. MadTracker, Sk@le and Renoise are also guilty. OpenMPT's bleeding edge is really just a mere sequencer that just sends MIDI data to plugins rather than something that can compete or blossom with its own sound engine.

What's your actual issue here exactly? OpenMPT's aim had always been to implement a few popular legacy sample-based tracker formats and it succeded at it very much. VSTs are just a nice afterthought feature, many people enjoy having access to them in OpenMPT. You're not forced to use them if you don't like them and i don't think OpenMPT is or had ever been striving to "appeal to any kind of industry". It's an open source project made out of love and passion for tracker music, rather than something trying to "sell itself". I feel like your thinking is too fixated on mainstream music industry in this context.

I don't want to choose between highly restrained chiptune formats or squeaky clean DAW recordings. I want the middle ground that doesn't exist, I want fancy synths in music formats that play in the browser via WebAssembly and xmplay/foobar2k.

MPTM is too restrained for you? Well then you don't have other choice than use other software. And who said that music made with DAWs must sound "squeaky clean"? You can process it whatever way you want, even pass it through analogue gear. Or if you want something yet more different, you could buy an old MIDI synth like Yamaha's MU2000 which supports XG MIDI and has a built-in sampler if that's your thing. There's lots of possibilities and ways to create digital music and making it sound the way you want. What you're imagining, regarding modern tracked formats played in realtime and utilising fancy internal synths isn't very likely to happen, and even if there was some kind of project matching your description i don't think it would gain lots of traction. Be the change you want to see i guess, instead of complaining about what OpenMPT isn't.

EDIT: On the second thought. there already are things that could become that, like KlysTrack. except as of now it doesn't have any external playback engine implenentations. Or, as you mentioned, VGM format which supports emulating various old gaming systems, including arcade sound hardware incorporating PCM and FM synths for example - there's DefleMask which can make use of some of those systems, even though it's a generally very clunky tracker.

[Saga Musix]

Maybe this middle ground you want doesn't exist for a good reason. Because this middle ground will look completely different for everyone else and thus there is not a single solution: There simply is no one-size-fits-all synthesizer for everyone. The closest to a completely self-contained, high-quality music production system (until they added VST support) was probably software like Reason, and for that very reason it also wasn't exactly portable. Quite frankly, I'm tired of people who propose gigantic dream worlds like what you describe here and asking for attention for their ideas without ever trying to implement that idea themselves. Synthesizers come and go. 4klang can sound great but it's a lot of work and difficult to use, it's not a tool I would want to force people to use for all their music. And many people already grow tired of its "standard" sound and use different 4k synths instead. Similarly, any of the other synthesizers you mentioned are all there to fulfil a specific niche, and they will be superseded by the next, better tool. Synthesizers come and go, but trackers/DAWs that do not enforce a specific type of synthesis are here to stay, if OpenMPT's age of more than 23 years is any indication. Similarly, synthesizers that can be used in any tracker/DAW the user chooses are here to stay, because that allows them to be used by anyone, and not just those that are willing to study a specific music creation paradigm.

And this is my last word on this discussion, because there simply is not anything more to say.

[LPChip]

One addition, there currently is no DAW at all that is this feature rich. Not even amongst sequencers. Sure, if you look at big DAW's such as ableton, bitwig, fl studio, they do come with plugins out of the box, but even then they are limited, where people quickly want to add their own plugins, and then the same problem exists again.

But if we go back in time and look at trackers when they fluorished, you only had samples, and the sound quality was bad. We settled for that, because there simply wasn't anything better other than actual recordings, but that was far out of reach. People have created music tools with a very fixed soundset, that was heavily used, and support for addition was possible too. For example Rebirth was a very popular software package emulating various hardware synths and string it together to create music with it. It got to the point that if you wanted to play a song, you almost always had to also install a specific module pack/skin pack too, which basically is the same as using a module with free VST instruments, and ask people to install those before your song works.

Its possible, but in 2020, people don't want to go through the hassle anymore to install something so they can maybe once listen to a tune. So instead, people want the mp3 version to listen to the tune, which is why people stopped sharing the source and just go for the mp3 version only.

One last issue that is also resolved here is the fact of song theft. In the past, with the source being available. More often than not, people downloaded the source, made a few tweaks then claimed it their own. Some people got famous using this opposed to the original creators.

One famous example is Timbaland who stole part of a module made by Tempest.

[Louigi Verona]

Hey everyone!

Interesting discussion. I remember wanting to make a YouTube video about trackers as music tools and also talking about the future of tracked music. Should I do this eventually?

In my personal opinion, trackers are unique tools in many ways and writing music with them is fun. Just like any tool that offers a new angle at writing music, it makes certain things easier, thus shaping the kind of music that is being done with it. I talked about this when I was interviewed for a podcast. If you are interested, give it a spin: The challenges and the rewards of writing electronic music.

However, at the same time, I would challenge the notion that trackers offer the same set of features as regular DAWs. Unfortunately, there are very serious issues that make certain relatively basic things pretty much impossible.

For instance, it is generally impossible to zoom in. In a DAW I can get to any level of granularity. This is very difficult to do in a tracker. You can try to do workarounds, like blowing up a pattern, so that it is twice as long or four times as long, while increasing the speed, but this has it's own problems and makes it very difficult to edit. For someone like me who makes minimal house, where you want various forms of swing (like, swing on only the third note) - this is a real problem.

There is a limit as to the amount of commands you can apply to a given note, which makes it difficult to do several basic things at once. If I want to apply a note offset, a sample offset, panning and volume to a note, I should either hope that there exists a special command - or it's simply impossible. In a DAW there is generally no limit of commands one can apply to a note.

There are also many UI complications that, on the one hand, allow to do some things in a tracker easier, but for the most part make very basic things harder. Looking at what is basically a spreadsheet is geeky and fun and I personally love it, but I also recognize that a piano roll allows me to understand what's going on much faster and makes it trivial to move notes around.

Moving things around in a tracker is a chore. If I want to change a note in a piano roll, all I do is move it - it takes me a second and virtually no mental cycles. In a tracker I would need to make sure that I first find the note I need to move, make sure that the number of the instrument is the one I need, then make sure that when I'm moving it, I am not changing the position of the notes after it, then find a track to copy it into, making sure I'm not blocking any other long note. It's fine and doable, but complicated.

The fact that modern trackers treat a track as a mixer track is kind of problematic, because it doesn't actually limit which kinds of instruments you can play in a track, while frequently requiring you to use up more than one track for some instruments. So, it might require me to duplicate my plugins as opposed to simply assigning several instruments to a single set of effects.

You can use NNAs to go around this, but NNAs don't give you the amount of control over notes that you would have if you could just set them up individually. I don't want to have a single fade out curve for each note, I want to be able to have an instrument per track. So, the whole plugin addition thing to tracker's tracks is awkward: you can see that this was an afterthought, original trackers were not designed for this.

In general, modern DAWs have workflow versatility that trackers lack: there are many ways you can make a track in a DAW, whereas a tracker generally forces a single workflow.

It's been a while since I thought about that, so I am definitely missing other important things, but that's what I remember at the moment.

Now, obviously, various concepts of trackers exist, various quality of life solutions exist, even those with automation lines. But at that point I would say that those are no longer trackers, but DAWs with a weird vertical layout.

So, trackers are probably not going to go away completely, just like Amiga games are not going away completely. There are always geeks out there or people to whom the tracker interface simply speaks. But in terms of actual versatility, I think trackers are definitely very limited tools. Super fun - and I love them - but I don't see them being more than a very niche geeky thing.

[zikey]

I would challenge the notion that trackers offer the same set of features as regular DAWs

But at that point I would say that those are no longer trackers, but DAWs with a weird vertical layout.

It kinda sounds like you are confused with what DAWs are supposed to be. DAWs are for audio work, period. They're not just sequencers, not just MIDI editors, etc. They have to serve multiple purposes. A non-musician may use a DAW for editing sound effects for games, or editing audiobooks, or they might be purely a mastering engineer. Are trackers DAWs? At what point does a sequencer become a DAW? Asking the real philosophical questions here 

Renoise calls itself a DAW, but I wouldn't call it one, because by design it's not as versatile as Pro Tools, Cubase, Reaper or Ardour. It's important to understand their capabilities. ProTools and Cubase are optimized for working with audio stems/clips. Ableton is optimized for live performances, FL Studio is optimized for sequencing/automations. Renoise is optimized for precise editing and sample work, but not ideal for mastering or downmix-heavy work.

DAW sequencing does allows for very fine granularity, but music is typically highly structured and based around specific note lengths anyway. Even music where the tempo is not constant can be done by controlling the speed primarily, not the density of notes. Fine granularity is mostly good for capturing a performance at a constant speed, or to facilitate humanization, random errors in timing that make the note placements seem less perfect. And you do run into inelegance in trackers, but clever use of effects such as speed and note delay etc can mitigate it. Some tracker music changse tempo in specific parts of patterns JUST to increase granularity and make room for extra notes, while ultimately keeping the overall tempo in tact. It's confusing but magical and ingenious at the same time.
 
Of course, not all trackers have the same limitations. Some trackers allow you to add additional command columns as needed. Some trackers allow for polyphonic 'note release tails' per channel, some trackers allow polyphonic chords instead of arpeggios. OpenMPT doesn't have any of this. Basically there are a lot of innovations that can be applied to the typical tracker workflow that we aren't seeing because it's a niche area.

In a DAW there is generally no limit of commands one can apply to a note.

Yes in the sense of being able to use unlimited automation tracks, but there isn't any standard list of 'note commands' in MIDI aside from velocity, panning, pitch etc. And if you're using a VST plugin (and I imagine other plugin specs too), you are at the mercy of whatever controls that plugin exposes. A lot of tracker commands have no equivalent in most plugins. The DAW typically gives most control to its native built-in instruments.

There are also many UI complications that, on the one hand, allow to do some things in a tracker easier, but for the most part make very basic things harder. Looking at what is basically a spreadsheet is geeky and fun and I personally love it, but I also recognize that a piano roll allows me to understand what's going on much faster and makes it trivial to move notes around.

This is a legitimate concern, the more powerful and involved a tracker is, the harder it is to learn. Trackers are really optimized for efficiency and not convenience. For example, the actual sequence data for iconic SMB overworld tune is a mere 339 bytes, and that's for almost 2 minutes of music. Some trackers adopt the mindset of extreme efficiency - which is why older trackers only had 15 or 31 instruments, or why you type in raw hex (00-FF or 0-255) instead of 64-bit decimal points (0.69793734).

Personally, I believe a tracker interface can be married with a piano roll, and certain adjustments in the piano roll can be interpreted as certain effect commands. The software Buzz tried this, but kept the piano roll vertical for some reason.

The fact that modern trackers treat a track as a mixer track is kind of problematic

How many trackers can you call modern? I wouldn't call OpenMPT modern, as it's firmly based on Impulse Tracker, which is circa 1995 technology. Yes it has some forward-thinking extensions but nothing truly radical. Renoise does have a dedicated mixer and is flexible in how tracks are routed. And that's the only one I can think of.

So, trackers are probably not going to go away completely, just like Amiga games are not going away completely. There are always geeks out there or people to whom the tracker interface simply speaks. But in terms of actual versatility, I think trackers are definitely very limited tools. Super fun - and I love them - but I don't see them being more than a very niche geeky thing.

Trackers are tightly related to 'chiptune' music. Originally music editing was done in a text editor in hex and had to be pretty compact. For example, the 1.5 minute Overworld theme in SMB for NES is a just 339 bytes. Amiga music wasn't as tiny, but still needed to be quite small. Trackers came about as a natural way to quickly preview music as you write it. Horizontal editing still existed back then, Aegis Sonix used musical notation, and Cubase on Atari ST had a piano roll.  But now Renoise and OpenMPT and even the original FT2/IT trackers allow for more serious/modern sounding music. So it was an obvious transition. Many composers have outright abandoned trackers for fully fleshed out DAWs like Cubase where they have no limitations.

But I feel like every new innovation 'moves the goal post further' so to speak. A game composer moving from the NES to SNES would feel quite liberated to be able to use 'realistic instruments' despite the low memory available. Even the Megadrive's FM synth would've seemed exciting if all you ever knew was a pulse and triangle wave. Now we have a limitless supply of VST synths, so being limited to just one would feel oppressive, despite the fact that any synth like Surge today, is incredibly more capable than older systems.

As for me, I started with a DAW, and prefer to go backwards; I basically yearn for a tailor made tracker with only the features I'd ever use. I like trying new obscure trackers and synths because they often bring something new to the table, and help me brainstorm what essential features would be good for a future tracker.  I find personal satisfaction with working with weird hardware/software, or making my own custom tools. The technicality of trackers allow for expression beyond that of just music. Many creative tracker users have used them in unexpected ways, animating things, or going in reverse in patterns, navigating patterns in odd and creative ways. So it can facilitate what could be characterized as Music Engineering, where the final song is merely the cherry on top to the intricate crafting of the song and instruments, which can be only be experienced fully by studying the source file directly.

Generally speaking, trackers don't have to be limited in the ways you describe - there's just very little pressure to explore solutions and innovations.

Happy Holidays!

[Louigi Verona]

DAW is an all-encompassing term these days. But, not to get too bogged down in semantics, you can see that in my previous message I was comparing tracker features to the corresponding sequencing features of modern DAWs. I wouldn't focus on things that trackers definitely don't have or weren't designed to do, like audio.

And all of my analysis is precisely about that - sequencing. At the end of the day, all of us would have wanted a proper sequencer with a piano roll, even in the 80s. We just couldn't afford it and trackers became the first tools to democratize the production of music. And because they were a very particular tool with very particular limitations, this shaped the kind of music you could do with a tracker. Which is why trackers are an important example of how instruments shape the music.

But apart from that outcome and an important step in the development of computer music, which is ubiquitous today, I don't think that trackers are capable of securing more than a niche following.

One thing I wanted to comment on:

the more powerful and involved a tracker is, the harder it is to learn.

I am saying that trackers are not harder to learn per se, but that they are harder to operate. The difference might seem subtle, but it's of paramount importance: learning something is a singular event. It might be hard, but it's a cost exerted once.

But if something is difficult to operate, even once you have learned it, the cost is exerted every single time you are using a tool. And I claim that trackers are harder to operate that a typical DAW - while the result is at best the same.