Cytomic "The Scream" stomp box distortion plugin

[andy-cytomic]

Hey, getting an SD-1 preset for free is really a great bonus feature and I sure as hell will not complain because you couldn't directly compare it to the real unit. Thanks for the extra work you put in this!
Were there many parameters that needed to be changed to match the SD-1 schematic?

There are quite a few component value changes, and I needed to add an extra resistor in the feedback of the tone control op-amp, but they are pretty similar all things considered. It's an interesting back and forth between Boss/Roland and Maxon/Ibanez - first Boss used the negative feedback diode clipping in the OD-1, then Maxon copied this for the TS-808 and added a 1 pole shelving tone control, then Boss copied that for the SD-1 while also fine tuning it a bit. Lots of people like modding their TS-808 by adding an extra diode to one side of the main clipping diodes, I wonder if they realise they should have just bought an SD-1 instead!

[Beatworld]

it is possible to optimize the CPU usage as the glue?

Yes, I could lower the cpu considerably by removing most of the modelling - this is exactly what everyone else does that makes Tube Screamer emulations!

I would have to remove the variable bias and voltage, remove the battery modelling, cheat on the buffer modelling, remove the op-amp input stage modelling (so no slew rate limiting), remove the op-amp voltage rail clipping, remove the non-linear capacitor modelling in the main clipping diodes and also remove a few other things - all this would lower the CPU considerably. You would basically have something like the TSC in which, as far as I can tell, the only non-linear part modelled is the two main clipping diodes. Would this sort of thing be of interest? I could still allow customisation of the main clipping diodes, but changing them would require re-computation of a table so there would be a small delay between changing the diode model and you hearing the result - so this would still be more functionality than all other Tube Screamer models on the market. I was thinking of adding an HD button, which when disabled gives drops the detail of the model to save CPU - this is what I do in The Drop. How does that sound?

hhmmm, well The Scream sounds killer now.
So I reckon I will learn to live with the CPU hit

That said, maybe the HD on/off option would be cool.
You could audition how much of a fidelity drop in sound happens.
Maybe some sources would still sound ok with the HD option off.

Looking forward to the manual so I can read all about the various little control options.

[GrabtharsHammer]

Yes, a lower resolution version for less CPU usage and latency would be a great option for tracking guitars. Then mixing with the HD option enabled.

[Liero]

Anyone know if it's possible to get Fulltone OCD -type distortion from The Scream?

[farfadetfarfelu]

+1 for a low cpu option the idea is to be able to play guitar live with low latency and a cytomic effect chain!

[andy-cytomic]

Anyone know if it's possible to get Fulltone OCD -type distortion from The Scream?

I've looked at the schematic and the Fulltone OCD is not very similar to the TS-808, so in a purest sense I would say no as the circuits are quite different.

[andy-cytomic]

Yes, a lower resolution version for less CPU usage and latency would be a great option for tracking guitars. Then mixing with the HD option enabled.

If you switch off oversampling there is no latency, and lower cpu! If you want you can also switch the oversampling to Minimum Phase for both up and down and you get 17 samples of latency with any amount of oversampling - is that low enough?

The best way to lower latency is to set your project's sample rate to 88.2 khz/ 96 khz, and switch oversampling in The Scream to x2 realtime / x2 render - this will give you 0 mS latency and have decent anti-aliasing. Remember to double your buffer sizes if you do this as this gives you the same buffer / soundcard latency as you would have had working at 44.1 / 48 khz.

[contrast]

You guitar guys just need to set both oversampling options to 1x which will eliminate the latency associated with oversampling (it reports 0 to the host in this configuration). The CPU usage is also quite reasonable in this case.

crumbs, I was beaten.

[andy-cytomic]

Anyone know if it's possible to get Fulltone OCD -type distortion from The Scream?

I've looked at the schematic and the Fulltone OCD is not very similar to the TS-808, so in a purest sense I would say no as the circuits are quite different.

I did notice the Fulltone Fulldrive 1 and 2 are pretty much a TS-808 with a few component variations, so The Scream covers that in spades!

[farfadetfarfelu]

Thank you for the tips.
I'm using lots of cpu hungry effects with my guitar setup so all the cpu i can save just for a dirty sound is a plus.

I spent more time with the Scream, i have to say i'm amazed at the sound quality and versatility with the mods.
Fantastic VST, really looking forward to more.
It sounds like the end of hardware.

Not sure if it's possible to do a Zvex fuzz factory with the feedback loop type thing in software without introducing latency ?

[andy-cytomic]

Thank you for the tips.
I'm using lots of cpu hungry effects with my guitar setup so all the cpu i can save just for a dirty sound is a plus.

I spent more time with the Scream, i have to say i'm amazed at the sound quality and versatility with the mods.
Fantastic VST, really looking forward to more.
It sounds like the end of hardware.

Not sure if it's possible to do a Zvex fuzz factory with the feedback loop type thing in software without introducing latency ?

The hardware will always have a place, but it is handy to have a completely noise free version in stereo on multiple tracks that load and save with your projects!

I think you're getting confused with latency. Solving the circuits without latency is easy enough, but you will have aliasing as the harmonics generated by the non-linear elements don't have enough headroom to be contained. The only reason for oversampling and introducing latency is to contain these harmonics, and to de-cramp EQ shapes, it's not required to solve the equations, just to get better sounding results.

The Fuzz Face is actually a simpler schematic than the TS-808, so that can be solved in full with lower CPU and do all the same crazy self oscillation type sounds as the hardware. I can breadboard all sorts of different fuzz faces and choose the best structure that can be modified the most with component changes to produce the widest range of customisations and release it as a plugin. Doing a low cpu version may be trickier since there isn't much that can be left out to lower CPU.

[farfadetfarfelu]

Okay thanks i understand, that's why the feedback stuff i've tried in modular software often sounds strange like it has phase issues.
Feedback/circuit bending style sounds is one thing i've been missing in software.
The "bias" pot is a nice addition in this regard.

Another small thing with the scream : i've tried to set R107 value to zero (to remove it from the circuit - like a solid connection) but when set to zero the sound stops entirely.

Fuzz is a world in itself, you've got people talking about types of NOS germanium transistors like they would about vintage wine...

[resynthesis]

Another small thing with the scream : i've tried to set R107 value to zero (to remove it from the circuit) but when set to zero the sound stops entirely.

What happens when you set it to 1ohm?

[farfadetfarfelu]

It works fine and probably sounds similar to the circuit with the resistor removed.
The point was more about the slightly counter-intuitive no sound result when entering zero values.

[resynthesis]

Ah, 0 -> open circuit rather than short circuit then. I will have to get a bit of time to demo this properly