iZotope Iris 2

[Spitfire31]

I like Simon's experimental and ambient stuff but wonder how complicated it is to create such material?

Well, the only real answer to that is to try it yourself, isn't it?

My guess is that not only does it take a fairly extensive knowledge of synth programming but also a trained ear and a deep background in music theory and practice.

/Joachim

[tq]

If you want to achieve the level of quality of Simon's work, then Joachim's answer is correct. However, I found Iris fairly easy to use. It's pretty straightforward to "paint" in the sounds and select the part of the sound/sample you like. As with all synths, some base knowledge of filters, LFOs etc. is helpful, but there are much more complicated synths out there.

[oneummon]

Thanks, and all the credits for Simon for his work, no question about that! I know what I like and have used to get the results working intuitively with samples. But that approach doesn't necessarily work with synthesis, right?

My question was more like, can you paint here and there, finding things intuitively or do you have to know the synthesis you're working with to get any results even near the same ball park where Simon is playing?

I'd like to use discount now and concentrate on Iris later, atm I don't have time to do that.

[Sampleconstruct]

Iris ain't no rocket science, a pair of trained ears help, a lot of patience does too. Painting dozens of little spectral dots over a time span of maybe 10 seconds in a sample can take up to an hour or more if you want to do it accurately and are looking for musical and not just random results, of course the sample itself has to offer enough interesting content for the start. So planning a patch for Iris has several stages, unless you just want to randomly reanimate your forgotten sample folders.
1.) Create one or several samples/soundscapes with other instruments, or from processed field recordings, or from just edited raw field recordings, or from instrument recordings, or from live-recordings of music/concerts/rehearsals or from instrument and drum/percussion loops. If you want to do split patches create similar samples for the high and low range, as the RADIUS-RT mode which preserves the original time of each sample is very CPU demanding/unplayable at high polyphonies and introduces artifacts which often just don't sound good, so living with the shortcomings of old school sampling often yields better sonic results (how cool would the combo of a high quality granulator like HALion with he spectral masking bee!). I often use 1 long sample for one patch (2+ minutes) and then use a different segment of it in each osc, as I know that before the Iris import I construct the sample accordingly.
2.) Start creating spectral selections, horizontal or vertical, dotted or rectangular, round or edgy - smooth transitions help by drawing shapes which fade in and out, so it's much more than on/off as claimed in this thread somewhere by someone. Try isolating only the odd or even harmonics from a musical sound, you can totally change an acoustic sound, e.g. a trombone this way and morph in into a bassoon.
3.) Introduce some FX assignments and LFO modulations in Iris, the options are few but can be used very effectively to add something musically useful. Bascially we can modulate volume/pitch in a sample, we can modulate the filter cutoff and resonance and with Iris 2 we can now modulate several useful FX parameters like delay time, chrous depth, distortion amount and tone and so forth.

Over time you learn to paint shapes with the few tools you have, you know when the magic brush will yield good results and you know when it won't.

Of course sometimes Iris magic can happen in more or less unprocessed field recordings, like in these sample from last spring, captured in the nearby deep woods:
http://soundcloud.com/sampleconstruct/i ... iris-set-1

[oneummon]

Sorry guys, I guess you answered to my question. I'm in a rush and didn't read properly. Just decided to have a day off and try myself. That's the way to do it anyway...

edit: Thank you Simon, let's see how far do I get within a one day

[tq]

It sounds more complicated than it really is. While you need patience to get excellent results, it's really fun to "draw" the sounds you like. And some things are pretty cool and intuitive, like getting 'sparkling' sounds with painting a cloud of dots in the spectral graph, or getting 'growing' sounds by painting triangular shapes (as Simon describes it above in more detail). Basic rule: Have fun!

[tq]

P.S.: Working with it makes it really obvious why they should have added more brushes/painting tools, a good time-stretching algo and a rhythmic grid / snap to tempo option (maybe also a "pattern" painting tool that adds repetitive elements on the basis of a rhythmic/spectral grid). It's such an obvious limitation, and it would make Iris 2 about 200 % more useful... As it is, rhyhtmic elements are a lot of (guess) work. It's still a fascinating tool, but I can see why so many people are annoyed by the missed opportunities from v.1 to v.2 ...

[Harry_HH]

P.S.: Working with it makes it really obvious why they should have added more brushes/painting tools, a good time-stretching algo and a rhythmic grid / snap to tempo option (maybe also a "pattern" painting tool that adds repetitive elements on the basis of a rhythmic/spectral grid). It's such an obvious limitation, and it would make Iris 2 about 200 % more useful... As it is, rhyhtmic elements are a lot of (guess) work. It's still a fascinating tool, but I can see why so many people are annoyed by the missed opportunities from v.1 to v.2 ...

+1
In addition to these, when Iris 1 was launched, I was expecting in the next ver. even more innovative ways to develop the idea of combining visual/graphical editing tools/interface and the sound editing. I think we are just in the beginning of opening these possibilities. Lack of risk-taking, more futuristic approach was the biggest dissapointment to me in the ver. 2. H.

[Caine123]

Dude it is a holiday. I make money off people too, but I get a day off, and my irate customers can take a hike on my day off, no matter WHAT it is, even if I am in the wrong! And I tell them, take a day off, you will feel better.

So Caine123, take a day off. Feel better. On Monday, bash em.

Jon

yes, i know santa claus exists !

[Sampleconstruct]

One route Iris 2 could have taken is to add modulation options for individual spectral bands and user-defined shapes. One tool I frequently use is GRM Tools Fusion which let's you set up up to 8 bands, each one has 8 parameters like length/delay/frequency/voiume/stereo width/pan position/feedback which can be modulated by the random modulator which has a speed control (controllable LFOs would be better of course):

So if we could apply these sort of spectral manipulations in each of the 4 oscillators, then have granular synthesis/time-stretching and all of the other features already mentioned in this thread inside a "sample player", I would definitely apply for a second life as a musician and sound designer (which I will probably do anyway).

http://www.youtube.com/watch?v=hqxA3Sxl-qQ

[MillerSam]

Wow great Fantastic

[Gribs]

I think that I might have just been sold on a GRM Tools purchase some time in the future. It looks like it is on sale at DontCrack right now, but I am not going to exceed self-imposed holiday limit and I have new instruments, sound sets, and effects to learn.

The computations required for Iris to do something with, say geometric primitives like ellipses and polygons, even just convex quadrilaterals, that is similar to the morphing, moving line segments for Fusion might be a bit much to do in real time at audio rate; I don't know really. In terms of computational geometry and intersection of the spectrogram at any slice in time to extract a set of frequencies to be superimposed to create the sound though, I think the math problem is or can be solved.

If were to get a second life and I was allowed to choose, I would probably direct myself away from electromagnetics and optics and more towards neural networks and signal processing algorithms and/or audio and higher rate DSP. Either that or MD/Ph.D. at Mayo in neurology and signal processing possibly related to human brain processing of sensory inputs. I would still have one foot planted firmly in "math land". There is an invariance of crap in the research world, though... corporate crap, university crap, medical center crap, self-owned company crap...

[lxegs]

Ableton Live users, caution, there's a bug :

if you drag a sample from the ableton live browser, it will crash your DAW if you try to assign an enveloppe/LFO

if you use the windows explorer, there's no problem at all !

[wasi]

Amaranthaudio's Cycle does some similar stuff too.

There's a lot Izotope could have done with the spectral side of Iris, but from where they've taken it with Iris 2 that would be almost a complete 180. Not that I wouldn't welcome it, I just don't see it ever happening after they've taken such a big step in the other direction. Maybe when they run out of filters, distortions and modulation modules to add.

[BBFG#]

Does GRM tools require PACE?
(A simple yes or no is all I'm looking for here.)

I see it 'no longer requires iLok', but that often still implies the 'no dongle/PACE' is still there.