Hi folks,

I have a thread going in this forum, asking what you would like to see from Valhalla DSP in the future. This is sort of like that thread, but from a wider perspective: What are the sounds you hear in your head, that you haven't been able to get at yet?

For me, there are a few "holy grail" sounds I am still seeking:

- Spatial barberpole expander/contractor. I know how to make sounds spin around your head, but I want to make something that makes a sound continually expand outward, or continually implode.

- Orchestra in a box. The current ÜberMod modes sound like really nice chorus and ensemble effects, but still can't simulate the sound of a full orchestra. I want something that takes a simple sawtooth, and turns it into this:

http://www.youtube.com/watch?v=wiyoLa9z1ao

- BoC in a box. Similar idea, but instead of adding orchestral overtones, it creates the broken 16 millimeter film recording of multiple generators of warped tape of drifting analog synths and gritty samplers that the best BoC tracks get. I actually have an idea for this, but we'll see if it works...

- Ligetizer. Similar to Orchestra in a box, but with the ability to get tonal clusters. The Pendereckizer is similar, but more chromatic.

- Really distorted filter emulations. In general, I want to stay away from analog emulations, but this is an exception. Urs' Diva is getting pretty good (pretty amazingly good, as a matter of fact) with the Moog ladder emulation. But my Monotron still sounds insane compared to any digital filter. I wonder if the noise and low voltage of the Monotron is part of what adds to this. Quite frankly, I might be better off building a few Altoid's boxes with Korg 35 filters in there, as these only use a few transistors (like, 5 transistors + 2 opamps/FETs). A digital emulation would take a few percent of an Intel core, which translates to several million transistors.

- Distortion that emulates what REALLY LOUD sounds do to the ears. There are certain nonlinearities in the ear, that produce incredible spatial effects when listening to really really loud sounds. I love this sound, but it is really bad for me, and I can't pursue my love of this and be able to hear my own plugins. So signal processing based on nonlinearities in the ear would be cool.

So what are your dream sounds? Think big here. These may never be plugins, or may not be the sort of sounds that can be expressed in plugin form. The Om vibration of the universe probably can't be captured, even at 24 bits / 96 kHz, but we can sure try.

Sean Costello

I like your ideas here, particularly the distorted filter thing. Diva's filters are great, yep (big fan of OTA filters, Diva seems to have nailed them). In a different realm (more "digital" I guess), I really like the Chris Kerry filters for SynthEdit.

I also like the sound of the "ear distortion" thing... that pretty much sums up my personal holy grail quest: BIG and UGLY.

Cool ideas! I especially like the BoC thingy! The wow and flutter mode for ÜberMod we once talked about on Twitter might get you a little closer.

I don't have any ideas myself at the moment but I always loved fractal based stuff (fractal modulations sound insanely musical) so something in that direction (no idea what exactly) would be nice!

Cheers
Dennis

I had a literal sonic dream of singing holographic laser rainbows. I try every once in a while to re-create it, but haven't yet. Attempting to be technical, it was harmonic tone clusters of formants with mid-side processing of tempo/rhythm synced pitch, tempo and spacial positioning effects.

I feel like a lot of the really cool effects and sounds have come from technology that strived for something more than it could be under the present conditions. Stuff like lofi- samplers, "bad" time stretching, etc...Not sure where I'm going with this....

Mine are simple (to the point of being boring) and not too far off from work you've already done (and both are plugins I want, not just abstract sounds/ideas):

1. An Echo Chamber plugin that sounds like a real-world 60's echo chamber. Particularly, the one at Capital that you hear on all The Beach Boys recordings (the Abbey Road one would be great too). No digital reverb even approaches that sound, which is much more than a reverb (speakers playing into a chamber, recorded back into a mic, into a tape machine and console).

2. The other is very similar. I'd love a plate reverb that had the real sense of sound bouncing off a metallic physical object that you hear when you listen to recordings done on real EMT plates. The UAD EMT140 sounds ok, but misses that piece big time, so it just sounds like a digital emulation of a plate as opposed to a plate.

I had a dream in my early 20's, where I played a Micromoog, and the most beautiful complex orchestral tonal cluster came out. The ribbon control had a rainbow-esque shifting iridescence on its surface. I'm still trying to make that sound.

Having a dream with mid/side processing is a pretty complicated dream.



I don't know if there is any "good" time stretching. It is pretty much an impossible task, as you are defying physics. I remember that Barry Truax talked about exploring the "inner complexity" of a sound via time stretching. In my opinion, he was exploring the particular complexity of the artifacts associated with his time stretching techniques.

Plus, let's face it, perfect time stretching would sound like musicians playing really really slow. Every now and then this is a good thing, like with The Melvins or Sunn O))). For the most part, it would be really boring. The massive FFT-based time stretching of Justin Bieber that went around a while back has all sorts of phase randomization artifacts, that turn the piece into interesting ambient music. Perfect time stretching would be Justin Bieber singing the song at 16 bpm. Not exactly fun.

In my "youth" (I started programming in my late 20's, so not really young), I tried to find DSP techniques that were artifact free. I'm more interested nowadays in finding interesting artifacts.

Sean Costello

I'm looking into "fractal" modulation for the next rev of ÜberMod. In particular, using a noise waveform that has similar levels of complexity at different frequency areas. Fortunately for me, there is a way to "cheat" to get such a waveform without eating a ton of CPU. It may sound horrible, but I'll give it a try.

Sean Costello

Sean, what I am looking for is an effect that can literally make a sound come alive. I want to be able to take sound and turn it into something that sounds organic and living. I am not talking about just a morphing formant filter to simulate vocal phrases, although that is cool. Rather it should be something more akin to physical modeling, but not constrained to organic models which exist in real life. It should bring in random elements, but it should also be able to restrain those elements to exist only within given parameters. It should add complexity to the sound to simulate non-linearities of living, breathing creatures. And of course, it should be extremely flexible, editable and responsive to modulation.

What do you think?

I'll throw 2 things into the mix...

Iterative processors. I love tape delays and FreqEcho, and I think the reason they make pleasing sounds is that they continually transform by applying multiple iterations of the same change. Tape echo iteratively applies bandwidth reduction, speed instability, and occasionally playback speed. Freq Echo iteratively applies a frequency shift - and I guess some reverbs work by iteratively applying a multi-tap delay... but what about iterative subtle wave shaping, iterative slew-rate limiting, iteratively mixing in a bit-crushed signal, iteratively mixing in a tiny amount of double-speed playback with high roll-off to compensate for pitch gain... and so on. I'm sure that in the same way iterative fractal maths makes beautiful visuals, there are beautiful undiscovered sounds out there in the land of iterative audio processing.

Secondly, U&I software's MetaSynth was a revelation to me, really making me think about audio in a different way, and harnessing 2000+ oscillators with images acting as thousands of stacked multi-stage envelopes is amazing... but like the DX-7, where everything descended into nasty noise unless you really knew what you were doing when you edited sounds, you have to work at the program to stop it producing a rather unpleasant clangy obviously resynthesised sound. An artifactless Metasynth would be a dream come true!

I really like this question but still can't come up with an answer in the right form

I feel like I might be interested in ways to sort of cut away from proper physics and acoustics down to really minimal (even to the point of artefact) expressions of musical or tonal qualities.

Weird mic placement would be an example - I'm totally ITB but I often feel like, how would I simulate overhead mics here? Or a weird mic placement when recording at some interesting, non-studio location? Or simulate my ear close against the resonant chamber of an acoustic guitar?

Another example would be getting the energy levels to match an interesting natural sound, like a waterfall or a bunch of llamas, or Tibetan Lamas ... I think some verbs do pretty well here but would it make sense to do this with a formant filter? A phaser? Like, using something with resonant feedback but doing it in a way that was always organic (or at least with very good control over that sort of thing).

... Not even sure this post makes sense I feel like it's also pretty relevant to say that there's so much magic in having good, dynamic tech and the right kind of UI to be able to translate some of these vague, dreamy ideas into sounds, and I probably dream about that more than anything. Very happy with Valhalla stuff along these lines.

Valhalla Deverb

Examines audio files looking for repeated impulse patterns, identifies those parts of the signal that are the product of a reverb, chorus or delay algorithm, and subtracts them from the source. Repeats this until it's identified all individually treated instruments. Outputs a file that can be opened in the DAW of my choice, containing the original un-effected multi-track, complete with Room/Uber/Shimmer presets on each channel to reproduce the original.

Version 2.0 analyses spatial positioning of source sounds, and can isolate different sound sources, such as the individual drums in a kit recorded with a stereo mic.

Version 3.0 introduces the ability to generate presets for ValhallaGuitarAmp and VallhallaMicAndPreamp

Vocal / formanty stuff. Although I admit I'd prefer this in synthesizer form rather than effect form and I the vibe that Valhalla DSP is more about effects.

I'm not into sample libraries and none of my favourite software synths do particularly good vocal sounds. They don't even need to be realistic, but that just need that somewhat vocal character. For example, my ancient Roland S-10 sampler used to have a really cheesy choir sample that I just don't hear in synths these days. Kind of like that choir sample in Blue Monday (which I read somewhere was a sample of Kraftwerk using a Roland Vocoder Plus or something).

Maybe I shouldn't fight against sampling and just use the appropriate tools for the job, but artificial voice is what I would describe as my dream sound and I don't know how to get it (at reasonably cheap prices at least) in software.

This could let you map segments of the MIDI keyboard to different types of tone clusters, for super-quick performance. But I can't bring myself to say what such a feature would be called.

This sounds like a job for a future ÜberMod mode...

Sean Costello

That would be ideal for me, given that I've already bought a ÜberMod license.

I can't help but think you are cheating yourself out of extra sales here by including it in an existing product, but that would be great customer service if it did end up there ÜberMod.

Thinking about it a bit more, what would be really interesting from my point of view is if you could make a tameable formant filter (or, I guess in the context of a modulation effect that does chorus amongst other things), tameable formant parameters within the overall chorus / modulation effect. I've played with quite a few formant filters (freebies, admittedly) and I find them very prone to harshness and distortion. It's like there is a very small sweet spot over which you can modulate them. Now if you could put the R&D in modulating over only smaller but "right" ranges I'd be a very happy camper.

Oh, and FWIW, some of the more interesting standalone formant filters have more subtle vowel combinations than the standard 'a', 'e', 'i', 'o', 'u'.