thawkins
-
Posts
2,014 -
Joined
-
Last visited
Content Type
Profiles
Forums
Blogs
Downloads
Store
Posts posted by thawkins
-
-
So now I've got my new laptop, which of course looks & works like a dream (for the price I paid it fucking better, though). However it's got no FW port for my aging Saffire LE and this means it's GAS for new audio interface time.
are you using a macbook? you can just get a firewire to thunderbolt adaptor. works fine with my Saffire Pro 14.
Yeah but it's one of the new ones, which means I have to get FW to TB to USB-C which seems like the dodgiest solution and it seems the adapters themselves will cost as much as a "new" used interface.
-
So now I've got my new laptop, which of course looks & works like a dream (for the price I paid it fucking better, though). However it's got no FW port for my aging Saffire LE and this means it's GAS for new audio interface time.
I've found a Tascam US-1800 on CL for an OK price. It seems it's got enough ins and outs for patching all my gear so far. Bonus for having enough ins to record drums as well, which I might even have an use for.
On the other hand, there's also an Apollo Twin (for 4x the price of the Tascam), which supposedly has amazing preamps and would force me to downsize my setup, which I kind of find appealing because it seems I am more productive with less stuff and I like it when I can put my stuff in my backpack.
And finally, the non-GAS option is to make some use of the shit I have lying around: a bass multieffects pedal with builtin USB interface, a Behringer UCA200-something and the Koma Field Kit.
-
this will always keep happening innit? I got the field kit, was gasing for the thyme but this seems like a similar but more viable fx solution for my setup... should I do it? iunno, I have barely made any music as of late anyway... this synth world is terrifying
I got the e-mail too since I have the field kit, looked at the thing for 1 minute, then closed everything before I could read enough for any GAS to materialize. Because it of course sounds really good, with delays and stuff and I distinctly remember the word "sequencer". I'm afraid to click the link now, I already have a huuuuge expense on the way since I finally decided that it's time to replace my 8 year old laptop and my phone (which I used for TouchOSC) is also shitting itself.
And I as well have made little music recently, instead spending the time messing around with trying to get my workflow in order. I have been told that I was much more productive when I didn't have all this gear around and they're probably right. So I'll get my new lappy, new audio interface (because new lappy no longer has firewire) and see where to go from there. Probably I'll get a Behringer XTouch Mini as well. But that'll be the LAST thing I buy, promise.
-
Awesome, thanks, I'll have a look. Yeah basically the sequencer generates long lists; I was thinking it may be easier to convert everything to arrays, because each sequence will eventually contain several parameters per step, plus sequence length data for each channel..
Yeah I think for one sequence I would save it as a multi-dimensional array where the number of rows corresponds to the sequence length, and each row contains all the data for the current step (use a placeholder value like -1 or "null" to signify lack of value). That would be good for one sequence per file.
For saving per channel or even the whole project, things will probably get more involved and I'd go see if there is a description for the .nan file format somewhere to get inspiration from. Maybe the dumbest and easiest solution is just to save everything in a fixed directory structure so that there's a main project folder and then subfolders for each channel containing sequence data and channel global data (9 files per channel).
-
I found that a good entry point into PureData was Automatonism - it's free and really easy to get into -
https://www.automatonism.com/the-software
It's basically a load of prebuilt modules you can patch together, modular-synthesis style with the bonus being you can look under the bonnet at anytime to start to understand the low-level stuff.
Yeah, in the three years since I posted this thread I did give puredata a go and had a lot of fun making a somewhat nanoloop-inspired sequencer.. but then ran into a dead end when it came time to figure out how to save sequences to files.. I'll get back to it at some point, but nobody I asked seemed to be able to give me a good answer on how to actually do it.
The thing with PD seems to be that you also need to implement low-level stuff a lot. So in order to save the sequences it depends a lot on how you store them within PD. If it's just a list of 16 numbers then I think you could probably use the qlist or textfile object for that. This should explain how to use it.
-
If you're not familiar with the Klee, check it out - it doesn't produce repeating sequences, it has three simultaneous playback positions and performs mathematical operations on each step depending on how it's configured and where each "head" is. So with 16 steps you can create a pattern that behaves in a predicable, controllable way but never actually repeats.
Yeah this little thing is great. Seems very limited as far as getting your hands on one, but I'm definitely going to keep my eye on it. It reminds me slightly of the Faderfox SC-4 I've got, being able to sorta randomize a pattern, but the Klee looks much more user friendly to my eyes. That SC-4 is a puzzle box every time I turn it on. Thanks for bringing the Klee up.
Yeah with those long sequences i'm not talking about long melodies as such but more 2 or 3 etc times the tempo so you can create more interesting gaps between notes and high speed note changes you wouldn't normally get on a 16 grid. Of course, easily done on Elektrons ha.
The first thing that came to mind with so many steps was to increase the sequence speed so much that messing with the individual steps actually changes the timbre of the sound, not only some parameter. I think if the number of steps approaches the level where you could treat the sequence as the amplitudes of a single cycle waveform, it opens up so many new and weird ways to mess with the sound. If my math is correct, a 2205 step sequencer should be enough to sculpt whatever single cycle waveform (because at 20Hz the waveform repeats 20 times a second, and 44100/20=2205).
Should just sit down and hack something together in Pure Data. This is usually the point where I get the GAS for Max4Live because it seems much easier to use as it's so well integrated with Live.
Personally if I had something like this I'd probably use it to control things other than pitch, like running it with some slew and using it as a 128-stage looping envelope, or running it really slowly and letting a single pass be the structure of an entire piece of music or something.
Yeah, these more 'extreme' use cases are exactly what I'm talking about when I said that length of sequencer inspired me. I don't (necessarily) need 128 steps to program a drum beat (though I've definitely done it before) that wouldn't be my go-to use for a sequencer like that.
thawkins, if you're already using Live then Max4Live is your best bet no doubt. I've heard a lot of good stuff about PD though....I just don't think it's necessarily one is better than the other. From what I've heard, at least.
Speaking of, Live 10 is probably about to drop any day/week/month now. Almost definitely before the end of the year. I have no money though :/
I've been playing around with PD a lot this year. It's interesting to build totally new and weird stuff with. The only issues I have is that it's not really easy to integrate with Live in a comfortable way (maybe I just haven't figured it out yet). Basically I'd just like to have a rack (or two) in Live which send MIDI CC to PD, but the only way to do that in a sane way is get Max4Live and use the ControlChange device, at which point it makes more sense to move over to Max altogether. :)
Just Googled my MIDI CC problem again and there seems to be a really promising solution in the form of ClyphX's Macrobat racks. Going to try this out definitely! Seems like a nice hassle free (and scripting free) solution.
The MIDI Rack allows you to send MIDI messages (Control Change, Program Change and SysEx) from the Macros.
-
I think if the number of steps approaches the level where you could treat the sequence as the amplitudes of a single cycle waveform, it opens up so many new and weird ways to mess with the sound. If my math is correct, a 2205 step sequencer should be enough to sculpt whatever single cycle waveform (because at 20Hz the waveform repeats 20 times a second, and 44100/20=2205).
Should just sit down and hack something together in Pure Data.
One of the earliest things I tried when writing my music software was giving myself the ability to "draw" an arbitrary waveform to use in an oscillator. The results were consistently bad, in uninteresting ways - most things I tried either sounded like a fuzzier sawtooth or a cheap electric organ. Things like subtractive synthesis and FM synthesis tend to alter the harmonic content of a waveform in specific ways that are pleasing to the ear, whereas IME one is unlikely to stumble upon anything good by manually modifying the waveform.
Yeah I'm pretty sure that messing around with amplitudes in the microsecond range will not give easy results. However I also kind of think that maybe when the right tactile user interface is provided (like faders or the NURBS thing RSP mentioned), it could be more easily learnable. It's just that the ability to give the user pretty much limitless sound shaping seems really enticing.
Hmm, a NURBS based hardware oscillator would be pretty cool. A detented push encoder to step through the control points (and enable/disable the current point by pushing), two non-detented encoders to set the X/Y coordinates of the selected control point, a scale control to set the base speed (from audio rate down to really slow LFOs), a little OLED display to see the waveform, and direct CV modulation o the X/Y coordinates of at least two or three control points.
I remember a long time ago reading (I think in the first edition of The Computer Music Tutorial, which is really out of date but a treasure trove of information abut 80s and 90s digital synthesis techniques) about a method of synthesis that generated waveforms by scanning through a path across a three dimensional curved surface and outputting the Y value of the scan location at the sampling rate. You could create cyclic waveforms by scanning in a circular pattern, and change the timbre by moving the position of the entire path on the surface of the 3d curve. I can't remember exactly what it was called. It would be really interesting to implement something like this as a eurorack module, using either a grayscale bitmap as a height map to generate the 3d curve, or maybe being able to accept a USB webcam and use the luminace to generate the curve, with the scan path fixed, so that you could change the timbre by pointing the webcam at different things and the pitch by changing the scan speed. Either one would be really interesting, and could probably be implemented around an Arduino board.
I should have studied programming or EE in college so I could build this kind of crap.
Yeah, NURBS or even a reverse FFT (i.e. basically a massive sum of different sinewaves) could be interesting to mess with if you've got 100+ faders.
-
Feeling some serious GAS for the Make Noise 0-COAST, but I have to concentrate on the projects I already have. No reason to build a modular at the moment (even if this thing is a standalone).
This is the correct thing to think.
Yeah, I wouldn't personally want a sequencer that big, although I do appreciate control-per-parameter stuff in general.
I definitely would want a sequencer that big. Not ONLY a sequencer that big, but I'd definitely love to use it sometimes.
Also, once you have that many steps it starts to really push up against the limits of the brain's ability to recognize a repeating pattern anyhow
Good.
so it starts getting less and less practical from a purely compositional perspective.
Wrong.
Short term/working memory only persists for about 30 seconds, so any pattern longer than that will be perceived by the brain as continually new, rather than repeating (there are ways around this, obviously, but as a rule of thumb repeating patterns in music that are meant to be perceived as repetition should be kept well under 30 seconds). That means that a 128 step sequence running slower than about 240bpm usually won't be heard as a repeating pattern.
Good.
That doesn't mean it isn't useful, but it does mean that there might be more manageable ways to create a comparable experience for the listener (like, say, a 32 or 64 step sequence with some kind of generative elements that keep it from repeating exactly the same pattern. Something like the mighty Klee sequencer, maybe.
I do stuff like that all the time, but I also like longer, less repetitive sequences. This is very easy to do when dealing with traditional instruments performed by musicians, but difficult to do easily in software/hardware/with electronic music. I've gotten relatively good at it with software, but the more options the better. And using it with hardware is also something I do a lot (especially since much of my hardware is often controlled by software) but the more options, the more tactile feel and visual feedback from real instruments that I can get, generally, the better. The Rytm (and all the Elektron stuff with conditional trigs) is really useful in that sense but it's still not quite as quick as I might like for writing. Maybe in actual functionality it's simpler on a smaller device like that, but I can't help but want to try something larger and with more options on a sequencer. I'm watching the video on the Klee right now and it's certainly a very interesting machine that I'd never heard of, I'm definitely intrigued by it.
Btw RSP I'm definitely not being snippy in my response, I just wanted to cut it up to show really where my mind/way of approaching the same facts and ideas differs. :)
Oh, I come out of drone and drone-adjacent stuff and pretty much agree with you (although I've mostly avoided sequencing altogether until the last couple years), I was just pointing out that a semirandomized pattern of far fewer steps can achieve a similar effect and is more practical in some ways, if only because it's easier to physically manage.
Personally if I had something like this I'd probably use it to control things other than pitch, like running it with some slew and using it as a 128-stage looping envelope, or running it really slowly and letting a single pass be the structure of an entire piece of music or something.
If you're not familiar with the Klee, check it out - it doesn't produce repeating sequences, it has three simultaneous playback positions and performs mathematical operations on each step depending on how it's configured and where each "head" is. So with 16 steps you can create a pattern that behaves in a predicable, controllable way but never actually repeats.
The first thing that came to mind with so many steps was to increase the sequence speed so much that messing with the individual steps actually changes the timbre of the sound, not only some parameter. I think if the number of steps approaches the level where you could treat the sequence as the amplitudes of a single cycle waveform, it opens up so many new and weird ways to mess with the sound. If my math is correct, a 2205 step sequencer should be enough to sculpt whatever single cycle waveform (because at 20Hz the waveform repeats 20 times a second, and 44100/20=2205).
Should just sit down and hack something together in Pure Data. This is usually the point where I get the GAS for Max4Live because it seems much easier to use as it's so well integrated with Live.
-
Yeah but then I've gotta wear Google Glasses or a VR helmet like some sort of nerd or something. We make abstract electronic music alone in our rooms, we're not nerds.
Sure, I suppose there's also a way to stick a small projector on some stand and do stuff like this: (skip to 1 minute mark):
-
You know, with some google glass (or whatever the state of the art in augmented reality right now is), it is probably possible right now to build some minority report style overlay visuals to give you a visual representation of that 1000+ step sequence. And then it's probably possible to dynamically map things so that if you have a hardware control surface like the BCR2000, you can select the subsequence that you want to fiddle with, so you don't just have to twiddle your fingers mid-air. I'm like 90% sure you could set this up with Max and some AR toolkit.
-
Reorganized my studio desk this weekend.
Bad: no room for a full sized 88-key hammer-action thing in the foreseeable future.
Good: everything besides the Launchpad now neatly fits on a 50x90 cm space, the only thing I'm missing is a control surface for some live mixing in Live. I decided to give TouchOSC a shot and ditch my frankensurface for now to have some visual feedback, less hardware and less software hacks. I basically just replaced the hardware surface with a TouchOSC template sending/receiving the same MIDI. The first attempt worked surprisingly well (even visual feedback worked out of the box!) except the Device mode weirdly gets stuck and locked on devices. So twiddling the EQ on track A, then going to track B EQ, the faders still change values on track A's EQ.
I wish there was some documentation from Ableton for this scripting stuff. I'm sure they have something, but they only give it out to Real Device Manufacturers, and DIY folks are basically stuck reverse engineering the scripts that come with Live.
Edit: I know that there's several apps on Android for controlling Live, but at this point they all seem to be aiming for mirroring more or less the full GUI to the tablet screen, while I'm looking for something that's more simpler, like a mixer screen for controlling volume/pan/sends/eq and then a track screen for messing with the device parameters on a given track.
-
I kind of believe that if you have a good keyboard and a player who is skilled in dynamic and expressive playing, then the actual sound quality of a piano or whatever analog instrument emulation has a much lower bar to pass to sound great and believable. And since digital synths have had MIDI capability for ages already, any old semi-pro rack module should be good enough if you're not a professional pianist.
-
yeah I miss having access to a piano, even though my only real experience has been digital pianos. I should actually look out for a real upright piano, especially since my girlfriend said we can get one for the living room.. would be lush
Even the beat-up old Soviet-era Krasny Oktyabr upright (which go for less than €100 because of the hassle of transportation) that I fiddled on had crazy nice sustain compared to my Micro-X, but also the key action was way more expressive than any of my synth keyboards at home and this got me thinking. Of course I think I might be biased though because it's the same piano I used to practice on when I was learning music in school. What makes me think that I need a hammer-action is that I constantly find myself trying to play dynamically using different key velocities and it's more difficult to do on a keyboard that doesn't offer almost any physical resistance. Or maybe it's just because I have spent so much more time playing on a piano that it's just what I am used to and I probably also should stop comparing the cheapest Microkey keyboard to a €15k piano.
I was messing around with the demo of the gorgeous Pianoteq plug-in lately; I'm not normally a VST guy but it's comfortably one of the most beautiful plug-ins I've ever heard, it really made me want to buy it and pair it with something like an SL88. Fortunately I was saved by my own greed by the fact that my computer can't really handle the processing, but still...
I am thinking that there's certainly some Yamaha or Kawai box made in the late 90s/early 00s that has all the sounds I ever want. There was a Motif rack (with piano expansion module) on local Craigslist that I kept my eyes on for a while, but back then I thought don't need it as I already have the Micro-X. In retrospect I should have just bought it and ditched the Korg because I don't use it at all as a keyboard and I'm pretty sure the rack Motif has better sounds.
-
I got to mess around on a couple of real actual pianos (an upright Steinway and an Estonia concert piano) during vacay time and I am feeling that my Microkey has been blown out of the water badly and I need to get myself some hammer action. But I barely have room for a keyboard with 61 reduced size keys, I don't know where to fit 88 full sized keys even if I did decide to splurge.
But hey look at that Studiologic SL88...
-
The trouble with Reaktor is that it isn't multithreaded, so no matter how many cores you have, any given instance of Reaktor will be limited to a single core. You can get around this a bit by using the VST version, since different VST instances can be run on different cores by your DAW, but you're still limited in terms of what any one patch can do. Plus the GUI of the blocks thing in Reaktor 6 is REALLY resource hungry, when I was trying it out I could have patches that used 30%-40% of the core they were running on if they were minimized, but would completely pin it and stay way over 100% (i.e. bring your whole DAW session down) as soon as you opened the UI.
If this thing is properly multithreaded and gets enough good modules written it would be better than Reaktor Blocks as far as I'm concerned (regular Reaktor is kind of a different thing, though, and also seems to perform way better).
It seems like Reaktor has gotten into a similar trap to Pro Tools, where actually taking advantage of modern hardware would mean completely overhauling the audio engine. I guess they did finally do that with Pro Tools but no way am I giving Digidesign any more money, I had to get PT10 for work and it was like traveling back in time to the late 90s, in a bad way.
I'm not so fazed by the lack of support for multiple cores (after all my home lappy has only two), but stuff like the Blocks GUI pushing CPU usage % makes me a bit mad about the state of software today. I'd understand if this were some out-of-support freeware project, but it's something actively developed that people pay money for. Something's gone way wrong when in the year 2017 things are crashing when trying to render a 2-dimensional picture. I don't even mind if the audio engine isn't running on the bleeding edge, I'd just like the resources going towards processing sound.
Going to try the VCV again today, maybe they have updated and fixed the GUI glitches. If not, I'll probably have to remember to try again next year or whenever I get a new PC.
-
Thanks for the reminder, I installed nanoloop on my droid phone and instantly enjoyed it. :)
However there seems to be no way to tempo sync with Live or anything on the PC, so I am having a rough time trying to get the loops recorded. Does anyone have tips how to do that? I guess I could just work with the internal pattern sequencing thing too of course.
-
After getting the MS2000 I now feel I'm at the Gear Limit. My audio interface ins and outs are mostly full and I've run out of cables to hook everything up. I even have enough gear to have simultaneous "laptop" and "hardware only" setups, so at this point I should really stop buying more stuff.
But my treacherous mind has begun to focus on the most shoddy part on what I already have - the control surface. Right now it's a frankenhack made out of a first-gen Novation Launchpad (running the Launchpad95 remote script, which I don't really extensively make use of), a shoddily built Doepfer PE (which randomly sends some knobs' MIDI values without being touched) and a cheapo USB numpad which I run through ControllerMate to convert keypresses into MIDI. There's a custom Ableton Live remote script and a Pure Data routing/paging engine also involved.
I guess the good thing is that what I have now is Working and there's probably no one shot solution to replacing it all that would actually take less space (I think right now my frankensurface is smaller in terms of area than a Push).
What I'd like though is something with visual feedback for parameter names/values so I could just chuck the laptop to the side and not have to look at it. Maybe a frankenhack solution is to use Osculator and send values to TouchOSC?
-
I think the very first release was - If you didn't select an output device on the final mixer it apparently maxed out your CPU - But the version I used seemed pretty lean: Had a simple VCO->VCA->VCF chain going into a spring reverb and still had 97% CPU free (the meter is 'backwards' on the output mixer and shows how much CPU you have free rather than how much you're using)Ohh nice! I wonder if it's CPU hungry or not. Only one way to find out :)
I tried it out, sort of. I think my version of OS X is too old, and the UI did not render properly. It did not seem too CPU hungry with 1-2 modules but then I was basically not using it at all either. :)
The UI renders properly on my work lappy, so maybe I should just look into hacking my old one to upgrade the OS X.
As for Reaktor, I did try version 6 last year and some blocks patches straight up told me there's not enough resources, so that kind of put me off Reaktor back then. But I guess the Blocks thing is especially hungry, being a modular simulator? And the regular patching thing (that sort of looks like Pure Data) is probably much more efficient and nice?
-
Ohh nice! I wonder if it's CPU hungry or not. Only one way to find out :)
-
Thanks modey & RSP! I'll check out Cluster and Harmonia; I'm not too familiar with the stuff but it smells stronly of krautrock which is very much my thing. :)
With regard to the SP404 (and samplers in general) I saw a guy play a live set using two SP404s and a Korg MS20 mini which kind of tickled me in a good way because I love people wizard sounds out of minimal gear. I'm pretty sure he used the Korg only as a filter and he had a lot of stuff prepared in the sampler so looking at these things it's always pretty hard to understand where the samples end and the live manipulation begins, especially if you're like me and don't really know how any hardware works. :)
-
So I got that MS2000R and I am loving it. Yesterday I was working on mapping Live's arpeggiator to my frankenpush and it turned into a jam instead. I even haven't got around to really reading through the manual and vocoding and processing everything and anything yet. I only wish it was a little more portable for lugging to practice sessions or whatever else.
-
Ah yes, the world of crossovers, forgot about it, thanks for the info/links, will investigate. Yep, probably better to get speakers that can go lower by itself, though I mostly mix on headphones anyway, always irks me listening back on these I miss the stomach churners.
ould do the trick too: https://www.thomann.de/intl/jp/the_box_hpf1.htm?ref=search_rslt_crossover_170808
If it's for listening and you don't need accuracy much then I say go for it if you want to. Use your ears and all the tunes you like and know to set up the relative levels and you should be golden.
I know what you mean about missing the bass - my first monitors were Edirol MA-10D and when I finally got my "real" Yamaha HS80Ms everything clicked into place.
-
quick side question;
the alesis mk2's I have here don't go below 50hz, what would be a good subwoofer speaker to go with it and how would you connect it to a mixer, such set-up / should the mixer have a special output for the sub? subN00b
This is probably a shit answer but I've read that trying to match subwoofers with your monitors properly is a hell of a task if you want that the result will sound accurate. So all in all it's probably cheaper just to get new monitors that have the bass response you're looking for.
But if you want to go through with it, it seems you need a crossover mixer like this (https://www.thomann.de/intl/jp/dbx_223_xs.htm?ref=search_rslt_crossover_264753_1) or some kind of hack to split the audio into 50Hz+ stereo and 50Hz- mono inside your computer or whatever is generating sound. And a subwoofer (https://www.thomann.de/intl/jp/fostex_pm_submini_2.htm). But after all that you'd need to really carefully mix and match with the setup so that your low frequencies will be more or less accurate. And looking at the prices here it seems you'd still be better off if you just got a new pair of monitors.
Edit: I guess this box would do the trick too: https://www.thomann.de/intl/jp/the_box_hpf1.htm?ref=search_rslt_crossover_170808
-
i wouldn't rate the ms2000r that highly tbh, i owned one for years and years and made several albums with it's help but it sounds cheap as fuck imo. also the 'sequencer' is barely a sequencer, just a way to create automation. the vocoder is out of control tho. really good.
they are cheap, mind, so what the hell get 2.
I think I have a really low bar when it comes to sound. As long as I can play it expressively with my keyboard and make it sound good, I am already in 7th heaven. And since it has an audio in and vocoder, all bets are already off since I can pipe all my Pure Data shit or AKFW single cycle waveforms through it.
Yeah, I have pretty much decided I'll get it. :)
Aleatoric/generative/algorithmic composition apps/plugins/etc
in EKT General Discussion
Posted
Yeah, that looks pretty intense and complex, but I would certainly like to try it out sometime (or see a video).
In general algorithmic and generative composition feels kind of off to me, it's like I have some deep revulsion against things generated by a computer. I think it's because I work with computers each day and I feel that they possibly cannot be a source of unique and creative stuff - stories for humans need to be made by humans. But on the other hand, all the music made will still be curated by a human, so the end result will necessarily have to pass that filter. And after all, my last album has one song which heavily features randomly generated notes, making me a total hypocrite here. :)
I have been thinking about implementing some generative/algorithmic things as well, though I have no idea how to actually implement my idea. Basically I'd like to have a system where I could somehow detect the playing "intensity" of one track and make the others react as well. So for instance if I am messing with the cutoff in one track, I would like the drums to also react to that somehow (either note repeat or some filter fuckery). Like sidechaining sort of.