DSP Robotics Support

Martin...

In case you're restless on a very rainy gloomy day... perhaps... maybe

I don't want to share any of my MacGyver.FSM:s on the forum since they are kinda junk  ...so i've attached one of them to an email instead (see email). So called "junk mail" FSM edition

Have made quite a few of these junk .fsm at this point Only way for me to quickly test some concepts since I stink big time at DSP and ASM.

Really have no idea if this particular idea would be of any use to others but would be nice to have ASM:ed oscillators similar to those two (see email) But without all the aliasing and preferably with sync (not non aliasing sync, just a reset at current phase) if possible.
Triangle and Sine that basically are ran through some multipliers. 

In my particular synth they would be placed in the LFO... and the Aux Oscillator, since it's a Sub Oscillator of sorts. Meaning the use of basic waveforms but with a bit of variation with two parameters.

Only bother if you feel like it... otherwise ignore this post No commitment nor stress iow since I can't offer much in compensation other than tiny occasional donations. 

Post your ASM on the forum if you choose to and/or manage to whip up an ASM at some point.
Don't share any of my junk .fsm though plz...

Weather forecast is quite good for the weekend (TGIF).
Just kidding. You should post here to get more responses.

martinvicanek wrote:Weather forecast is quite good for the weekend (TGIF).

Aaahw bummer!
I'll have to do the rain dance then...

martinvicanek wrote:You should post here to get more responses.

Yeah I know, I know...

It's not sharing itself thats the problem I just don't enjoy sharing unfinished junk, that sometimes is taken out of it's usage context.
I mean, folks will think i'm an idiot, instead of a moron morons are the lesser idiots you know

Fine. Here is this particular junk...

R&R wrote:...
I don't want to share any of my MacGyver.FSM:s on the forum since they are kinda junk
...

The thing is not to be embarrassed because we all went through the same learning process, so we understand.

Spogg wrote:The thing is not to be embarrassed because we all went through the same learning process, so we understand.

I'll never stop being embarrassed even if I learn.
Unfortunately that's how alot of us end up when being told we're idiots growing up

Sad but true...

Upside is it keeps one humble

Btw...
Simple usage case here:

Hi that's interesting, and not so junk, but i might look more closely.
I don't see aliasing for now.
Consider that even if it alias, when we have -40db on aliasing, it's normally not audible.
(or maybe more or less db, i'm not 100% sure but i could be happy sometime if i just shut aliasing or artefact
at -20db)
Seeing more closely, you use bandlimited triangle and a sin from Martin, and multiply it, this might not cause so much aliasing, the problem occur more when the first control is at 0 and the second at max.

But please, deactivate the asio out before sharing or use the DS out primary sound.
It cause bug when other's open your schematic, the asio is not configured in the same way but flowstone
keep the last setup in memory and try to apply it when we reopen the schematic.
Sometime i'm forced to replace some file in user data, or i loose most of the prim module..

Tepeix wrote:But please, deactivate the asio out before sharing or use the DS out primary sound.
It cause bug when other's open your schematic, the asio is not configured in the same way but flowstone
keep the last setup in memory and try to apply it when we reopen the schematic.
Sometime i'm forced to replace some file in user data, or i loose most of the prim module..

Aah good to know!
Fixed now. For these files atleast.

DS sounds terrible on my laptop for some reason, that's why I only use external ASIO interface or ASIO4ALL.

Tepeix wrote:Hi that's interesting, and not so junk, but i might look more closely.
I don't see aliasing for now.

There's quite alot of aliasing at HF. I have also crudely ducked one input by frequency...

Million dollar question is if it can be translated into simpler functions (rather more complex math, but using less cpu) without that many excess multipliers and additions, and also made somewhat alias free...

Uhm, yes, a lot of arithmethics in there. I would have to "decompile" it from spaghetti fsm (no offense intended) into plain math to see if it is possible to simplify the computations. But the sound is interesting, and a million $$ are a reasonable incentive So yeah, why not. Some aliasing at particular settings and high notes there is, may not be easy to remove.

I am embarrassed by the buggy scope I posted some time ago. You have to scrub the horizontal slider else it won't sync. While I struggle with R&R's intricate osc, here is a fixed version.

martinvicanek wrote:I would have to "decompile" it from spaghetti fsm (no offense intended)

The chef is utterly offended. Actually it's Fettuccini!

martinvicanek wrote:But the sound is interesting, and a million $$ are a reasonable incentive So yeah

If I could I would
Wish I had that amount of funny money to throw around

martinvicanek wrote:I am intimidated by the buggy scope I posted some time ago. You have to scrub the horizontal slider else it won't sync.

Nice! An unexpected bonus... I don't suffer from that problem that much, but then I'm feeding sample-lengths to the scope so. Might be why it doesn't show. Gonna take a look at that updated scope, thanks!

martinvicanek wrote:While I struggle with R&R's intricate osc

I appreciate it!

The struggle is always very real when dealing with R&R...
And... now... I feel guilty instead

Imagine how my newbie synths schematic looks inside

Yes sorry, it alias, doesn't test it with sufficient setting and high note to see it.
And complex to avoid this.. Different way but i'm not pro on them...

But i look at the first module Split Exp and try to see if i could make it more optimized, and clear.

First i try to decript what it does.. But finally i go blindly following some intuition and i think i'm pretty
close but it's not exactly the same.

Seams it consist when the knob go left to simply multiply the signal by the abs of this one,
and for right it's close to x+x-(x*abs(x)) two time in series. Then mixing with the original.


So if it could help here's an optimized but slightly different version.
(it could probably be even more optimize)

(Edit, now i see another difference, but it's when the volume of the osc is amplified, normally does'nt change anything when connected just after a classic osc)