How to get into synth / executable music?
category: music [glöplog]
knl: not my place to spill the beans, sorry. I hear he's doing something with it quite soon though.
Here is a heart so you can remember how much I hate you since you just teased me badly.
Anyway, it all went downhill.
Anyway, it all went downhill.
Sesse: Trust me, i dislike V2 as much as any other 64k synth. All the stuff I said I felt missing today is also missing in anything I coded. Yes. But, point is: It's from 2000. Back then I had to make lots of compromises because I had to solve the impossible size<>speed<>quality triangle somehow (plus development time, there was a TP2000 to attend). There were lots of things I couldn't do because the thing had to run on a Pentium II 300 without SSE and still produce a whole song AND leave room and cycles to run an intro next to it. There was no kkrunchy either. For comparison: the current kkrunchy version gets the original fr-08 executable to <48k if I'm not mistaken.
But this was TEN years ago. Ten years in which the whole demoscene, computer games, music production software have changed a LOT. Ten years in which CPUs have become so fast and compressors so good now something like 4klang is possible (which, in best 4k tradition, is completely optimized for size and wastes CPU cycles like hell, but hey, we GOT them. We're only drawing one poly per frame anyway :). Ten years in which realtime graphics don't look the same anymore; ten years that made music software so potent that I now know lots of musicians who look at their old MIDI gear and wonder how they were _ever_ able to cope with that stuff.
And what did the demoscene do with those years, audio wise? Apart from a host of quite cool 4k synths (but those bet everything on size, which is very reasonable)? Squat. Ok, ACM compressed samples (which kinda fail under Win7), Windows Speech API + vocoder (which also fails under Win7), simple Strong-Karplus models, and of course some delay line based glitchery. But that's about it.
Trust me, if I wanted to say "haaw, haaw, I'm still the best" I'd phrased it differently. Because I'm not. Part of the fun in the demoscene is that constant game of surpassing each other, even when the things people compete with can't even really be compared :). But the state of demoscene audio is either playing MP3/Vorbis files or having synths that compromise quality for size or even take pride in doing everything with samples again. Yeah, 1990 called, they want their mod files back. No, you can keep the lossy-to-death compression, thank you.
That's what my rants were about. That nobody gives a fuck about pushing the envelope, even if it is very possible. So if there's a new, brain blasting 64k synth on the way I'm very looking forward to it. :)
But this was TEN years ago. Ten years in which the whole demoscene, computer games, music production software have changed a LOT. Ten years in which CPUs have become so fast and compressors so good now something like 4klang is possible (which, in best 4k tradition, is completely optimized for size and wastes CPU cycles like hell, but hey, we GOT them. We're only drawing one poly per frame anyway :). Ten years in which realtime graphics don't look the same anymore; ten years that made music software so potent that I now know lots of musicians who look at their old MIDI gear and wonder how they were _ever_ able to cope with that stuff.
And what did the demoscene do with those years, audio wise? Apart from a host of quite cool 4k synths (but those bet everything on size, which is very reasonable)? Squat. Ok, ACM compressed samples (which kinda fail under Win7), Windows Speech API + vocoder (which also fails under Win7), simple Strong-Karplus models, and of course some delay line based glitchery. But that's about it.
Trust me, if I wanted to say "haaw, haaw, I'm still the best" I'd phrased it differently. Because I'm not. Part of the fun in the demoscene is that constant game of surpassing each other, even when the things people compete with can't even really be compared :). But the state of demoscene audio is either playing MP3/Vorbis files or having synths that compromise quality for size or even take pride in doing everything with samples again. Yeah, 1990 called, they want their mod files back. No, you can keep the lossy-to-death compression, thank you.
That's what my rants were about. That nobody gives a fuck about pushing the envelope, even if it is very possible. So if there's a new, brain blasting 64k synth on the way I'm very looking forward to it. :)
surprised he hasn't popped up in this thread yet actually.
Isn't this gonna end up in one of those "why should we use DX10 when there is still new stuff possible with DX9" debates? I mean frankly even V2 is unexploited in terms of not-the-author-using-it. (With some exceptions.) Are the coders to blame for not pushing the envelope, or is it somewhere the artists fault too?
kb: I remember 10 years ago when it had already been 10 years since people were saying "Are people STILL using MIDI?" ... and in 2010 it's still around. It's crazy.
sesse: Yes yes, the non-veiled hint is well recieived. :)
sesse: Yes yes, the non-veiled hint is well recieived. :)
I think the main point is that this stuff is quite hard to do. That's it.
Anyone can take a NeHe and render a few polygons, but how many demogroups are there that push the envelope in *any* field, not just synths?
And looking at 64k intros, there really have been very few improvements in this area, graphics-wise or sound-wise.
Not because it's impossible to do, but simply because it's hard. Too hard for most, I reckon.
(Disclaimer: "too hard to do" may also mean "too hard to do in the tiny amount of spare time you have")
Anyone can take a NeHe and render a few polygons, but how many demogroups are there that push the envelope in *any* field, not just synths?
And looking at 64k intros, there really have been very few improvements in this area, graphics-wise or sound-wise.
Not because it's impossible to do, but simply because it's hard. Too hard for most, I reckon.
(Disclaimer: "too hard to do" may also mean "too hard to do in the tiny amount of spare time you have")
Consider than a plugin writer do not want to use more than a few percent of the host CPU.
But when you launch a demo you expect to give it 100% of your CPU.
There is an alley there for expensive physical synthesis that plugin writers can't afford.
btw, is there any unlimited executable music compo?
But when you launch a demo you expect to give it 100% of your CPU.
There is an alley there for expensive physical synthesis that plugin writers can't afford.
btw, is there any unlimited executable music compo?
next to that, even IF you code a nifty synth yourself and produce some technically decent electrofunk or folkmetal with it, most sceners still vote for cliché goa made with said 10yo synth. so then why bother!
ah. ah. decent electrofunk is already 10yo :)
oh shit, wait...
oh shit, wait...
I think an absolutely kickass approach towards making a better softsynth for demoscene productions would be something that is actually fun to use. User experience and stuff like that. You know, an ui that doesn't alienate you in 2 simple steps.
I would have made so many more V2 tunes IF it wasn't such a pain in the ass to use. So, there's 4klang which i tried to get into. Man, what a nightmare. Same with the Werkkzeug family. This scene lacks a lot of people who have a sense for creating something that's actually easy and fun to use.
I would have made so many more V2 tunes IF it wasn't such a pain in the ass to use. So, there's 4klang which i tried to get into. Man, what a nightmare. Same with the Werkkzeug family. This scene lacks a lot of people who have a sense for creating something that's actually easy and fun to use.
ahaha ronny, actually, 4klang makes a lot of sense...
kb: Since you keep iterating the point about speech: Nemesis (TG04) did speech compression in the synth, without using any OS dependencies (would sort of be hard to do in a crossplatform intro, no?). Granted, the output quality is not stellar, but the input quality wasn't all that great either. :-)
Oh, and I have to agree with plainoldsagacity: Most intros, just like demos, don't push limits. :-)
Oh, and I have to agree with plainoldsagacity: Most intros, just like demos, don't push limits. :-)
Quote:
ahaha ronny, actually, 4klang makes a lot of sense...
and so do v2 and werkkzeug, etc ...
all do present basically what the underlying data model looks like.
and i can really understand ronnys and other peoples complaints about that.
but then again ... most of the time you simply cannot completely seperate the ui/presentation from the internal data models. especially not in a size restricted environment (as all of the above mentioned tools are used in)
"this vst is made with a musician-centered design paradigm!"
maybe apple should do demotools.
then finally musicians and graphicians dont have to complain anymore because they need to use shitty codertools ;)
then finally musicians and graphicians dont have to complain anymore because they need to use shitty codertools ;)
understand me gopher, I really think that once you've been using a bit 4klang, it really makes sense... But you have to know what you're going to do (which mean, from A to Z) with it and once you know, well, it's about "I'm gonna do this so I know I have to send this through this and add that" so basically, 4klang works well at doing that.
gopher: quartz composer? It's made by apple, and it's pretty much a demotool (I've made demos with it myself!) :)
Not sure it's all that much better than say werkkzeug (which it's pretty similar too) for usability though.. perhaps it's a bit easier but I'd say there are tradeoffs.
Not sure it's all that much better than say werkkzeug (which it's pretty similar too) for usability though.. perhaps it's a bit easier but I'd say there are tradeoffs.
I never said the tools make no sense or suck. I said, they're tough to use and you need to dig through shit in order to get something out of them. And hell, maybe i'm just too old or too lazy but since i don't have any time to waste i don't see myself solving a miracle anymore just for the sake of it.
Gopher,
there is no excuse for an unusuable or unintuitive ui. These are just bad solutions for problem that could have been solved in a different way.
Gopher,
there is no excuse for an unusuable or unintuitive ui. These are just bad solutions for problem that could have been solved in a different way.
Oh come on, from someone who know about the tracker era :)
(hug rp)
(hug rp)
Quote:
I think an absolutely kickass approach towards making a better softsynth for demoscene productions would be something that is actually fun to use. User experience and stuff like that. You know, an ui that doesn't alienate you in 2 simple steps.
I would have made so many more V2 tunes IF it wasn't such a pain in the ass to use.
I find V2 quite easy to use. What do you find wrong with it?
as if commercial guis are sooo userfriendly and non-alienating.. e.g. just randomly:
that thing has more knobs than a czernobyl control board!
that thing has more knobs than a czernobyl control board!
These kind of VSTi are used for PRESETR MAYEHM.
(or whatever "I loaded a patch and modified it, look, I make electronic music)
(or whatever "I loaded a patch and modified it, look, I make electronic music)
Typical scene-softsynths are all easy to use ... If you have a deeper understanding of what goes on "under the hood" ;-) Some have parameters arranged nicer than others, but in general, they'll all make absolute no (or only littel) sence to the typical non-scene desktop-musician! Especially nowadays, where a typical hobby-musician has never even laid hands on an old hardware synth.
When I look at my JP-8080, my Supernova or my Virus C, they'll probably be just as confusing as the softsynths that go to great lengths to simulate them to people starting with music today. They're used to nice Apple-stuff with nifty userinterfaces, understandable lables for the few settings required to operate them and so on.
"Usability" of softsynths should be a configuration that was controlled by an old analogue slider, allowing more and more advanced features and flexibility to be set individually, the higher the slider was set. Set to 0, there should be three knobs allowing you to play with some combination of filters and osciallators. Then you could gradually expand the knowledge by raising the detail-setting ... or something.
When I look at my JP-8080, my Supernova or my Virus C, they'll probably be just as confusing as the softsynths that go to great lengths to simulate them to people starting with music today. They're used to nice Apple-stuff with nifty userinterfaces, understandable lables for the few settings required to operate them and so on.
"Usability" of softsynths should be a configuration that was controlled by an old analogue slider, allowing more and more advanced features and flexibility to be set individually, the higher the slider was set. Set to 0, there should be three knobs allowing you to play with some combination of filters and osciallators. Then you could gradually expand the knowledge by raising the detail-setting ... or something.
Referring to a synth here is a complete failure. I haven't got any big trouble with the simple voice part of V2, that's fine. But recording and mixing in it is a different case.