DSP austerity in 2018

Elektron Gear Digitone/Digitone Keys
#298

Yeah you could tell it what sounds to cut out first (high/low priority), but other than that you have no control over it. Voices could cut out at random times in a single pattern, depending on fx used, complexity of the parameter sequencing, type of oscillator and filter used, etc. Overall it was a pretty lackluster implementation of probability trigs.

It’s nice to know exactly how many voices are available, no matter what you do. Nothing worse than voices starting to cut out when you’re halfway into programming your pattern. Instead of just programming around a hard voice-limit you now have to account for complexity of patches, p-locks, etc all while being completely in the dark with the amount of available voices at any given step.

1 Like
#299

…i’m in love with mistakes…flaws…and limitations…
and dedicated music instruments that sound great and do what they do, no matter how hard i try them…

and when i’m in love, i’m most creative…

and i really dig way deeper into creativity than into tech specs…

1 Like
#300

I guess the point I was trying to make on this thread was - just use the right tool for the job.

The pro keyboardist never did explain why he was looking at the Digitone in the first place, which puzzles me when I think of the scenarios that “pro” keyboardists typically work in - big pop productions, jazz ensembles, cover bands (wedding bands, bands that focus on a period like the 80s, Journey cover bands, Genesis cover band), live musicals, , etc.

#301

ok now let’s talk about novation peak and waldorf quantum coz i’m really ok with what i can do with the Digitone for now !!!

#302

I just used the microtiming trick to get two sounds playing on the same step+track last night for the first time. I read about it a while back, but your video got me to try it out. REALLY handy. :slight_smile:

3 Likes
#303

Fast? The Coldfire processors in the Digitone run at 250MHz and it requires two (2!!) of them to power the 8 voices. That isn’t fast by any means.

#304

And yet, they’re fast enough to do everything the DN does, and do it well. I’m obviously someone that loves knowing what’s inside… pretty much anything, but what’s more interesting is how it’s used than what it actually is. The fact that they’re getting the quality of audio and the functionality they are in a small box, for what is actually a pretty reasonable price is pretty cool. And yes, as far as general purpose MCUs go that IS pretty fast. It may not be the latest ARM based design, and it may not be Sharc+Black Fin, but they’re hardly slow ICs.

There’s another way to look at this. Since the SID Station, Elektron have been using Motorola (now Freescale / NXP) ICs. I believe the SID Station ran on an HC11 for the OS and additional functionality external to the SID chip. There is something to be said for using the ICs and architectures that you have experience with, and squeezing every last bit of performance out of it, than just jumping on something bigger, and starting over again.

It kind of reminds me of the era where PCs were just getting to be good at fast action type games. You’ve got coders that knew the 3.5MHz 65816 in the SNES coding in assembly, and creating amazing games with cool effects. Then you have coders coding in mostly C (maybe a little assembly) on a 40MHz 386, which should have been able to wipe the floor with the SNES. It wasn’t until id Software came along with Wolfenstein and Doom that the PC started pulling ahead of those 16 bit consoles.

Maybe not the perfect analogy, but it’s what springs to my mind. Who knows, maybe Elekton’s next devices will be jam packed with top end DSPs, multi-core ARM processors, and ESS DACs. That would require a lot of learning, maybe more hiring, higher costs, and ultimately higher prices. Maybe it would be worth it, maybe it wouldn’t.

All I know is that the Digitone is the most inspiring, fun to use, best sounding digital synth I’ve touched in years, and excels at what it does. It could be bigger. It could use more expensive processors. It could do things the likes of which none of us has ever seen, but that’s not what it is. It’s exactly what it is, what it was designed to be, and it doesn’t seem like they can even keep enough of them in stock for demand. They must have done something right.

4 Likes
#305

This may seem a bit cliche on an electronic music hardware based site, as people have been saying it for years, but I suppose I’d challenge anyone that doesn’t think this device is sufficient at what it does to come up with something better. I know I can’t make a better digital synth period, but then, I’m not complaining either. If we went analog, then maybe I’d have a chance :smiley:

1 Like
#306

Analog based high poly counts with a flexible synthesis? No way, that’s a nightmare to design and produce, because error rates go up exponentially (price needs to rise in the same way). Digitally all of this is much more relaxed. Just get enough calculation power and some high quality converters and you are good to go. BTW, have you to taken a look into the hardware of the MPC-Live? They use an absurdly cheap ARM-board running a modified Linux version. Take this a few steps further (for example: an Intel Mobile board) and there are no calculation restrictions anymore. Then it’s just about packaging and usability.

BUT don’t missunderstand me: my latest new device is an A4MKII. So, I’m for sure no one who would complain about 8 voices on the Digitone :wink:

1 Like
#307

Hmm… Analog polygon based gaming system eh? :smiley:

Well, Mouser still sells ADI analog multipliers (though some of them are upwards of $100 a chip). :rofl:

I’m all for whoever is making a piece of hardware to build it exactly the way they want and are comfortable with. There may be other equally good methods, maybe some better, but I think picking the best thing for the way you work (whether that’s a synth for a musician or an IC for a synth maker) is probably the ideal solution.

I think it’s cool that some people build around fast chips and linux variations too. I say whatever you’re good at, use it, run with it.

I’m pretty sure there are plenty of people out there that think the things I build or the way I design them is absolutely absurd. Whole PCBs that only do one small function that possibly even a slow MCU could do 6-fold. hehehe

#308

This conversation is turning very interesting… (I am currently reading books on DSP, but still in Introduction chapters, hehe)

Let’s not forget that one of the most amazing cheapest synths ever made has almost no power, but just kicks hard and his a pleasure to play with. UX and sound, all there is.

2 Likes
#309

That was a cool interview. I’ve owned uTonic for years, and never gave much thought to its development. Kinda makes me want a PO-32. It doesn’t look like they have a “pro cover” or whatever they’re called specifically for the 32 though unless I missed it. For some reason the bare POs put me off a bit. (psychological issue on my part) I’m even someone that’s very careful and takes very good care of my gear. I’d definitely need a case for it if I got one. :smiley:

For some reason no matter how many times I’ve started trying to learn how to code, my brain just rejects it. I’ll understand it, and even point out errors that I see when my experienced friends are coding, but when I sit down and try to write my own program it all just flies out of my head like a school kid on test day. :smiley:

I have started dabbling with CPLDs though, bought a dev kit, and laid out a few designs for various logic modules, sequencer type things, shift-register based things, etc. I’ve yet to order any chips though and try them out. I did lay out a PCB for it though that converts it to through-hole for a daughter-board type scenario, and have about 100 of them. :smiley: The reason I can work with this though is that the tools allow for a schematic-like approach to building the “circuits”. Logic makes sense to me when it’s presented that way.

#310

at least the samplerate is not limited to the usual 44 or 48khz.
quality against quantity

#311

The CA-X is a generic cover for the “metal” series; it’s generic due to the fact that the sounds on all three of them can be anything. I’m definitely considering getting the covers for my 33 and 35, but maybe I’ll go for the Dichstudios ones instead…

2 Likes
#312

Nice! Thanks for the tip!

#313

Before I got my first PO (the Arcade) I also thought I’d need to get covers for each of the POs. However, the naked look has grown on me, although I supposed I’d hesitate to play with it live in a dusty outdoor environment. Then again, I’d be reluctant to take any of my electronic gear out into a dusty environment.

I did however get a card case from Target for carrying my POs and their batteries. They don’t always shut off by themselves so I end up removing the batteries to prevent premature burnout.

#314

Straight from the horse’s mouth:

6 Likes
#315

Oddly most of the thoughts I had about why Elektron does what, Simon confirmed…
People had me questioning myself because there was so much speculation on seemingly shady/negative reasons Elektron would do things, that I started putting myself in check to make sure I wasn’t a mind controlled “fanboy”…
Most of the stuff had valid good reasons and Simon clarified all that, if anyone’s wondering check out the 3 threads in the new AMA section…

Something tells me though there’s still going to be a group that refuses to believe Simon and thinks he’s lying about things to cover stuff up… :joy::joy::joy:

2 Likes
#316

To be fair: he need not be lying - he may be just not part of the ongoing conspiracy :wink:

(just joking)

1 Like
#317

I am not one to expect every synth to have unlimited voices but I have to say 4 more voices would be lovely. Even just 2 if it is a matter of horse power. Perhaps it can be a per project setting then. Trade effects quality for more voices? How high definition does the chorus need to be?

And yeah another bank of tracks. Drums and chromatics. Boom.