Long-considered personal feature request list

First, a very brief hello: owned a TI for several years, but as a general rule, simply don't tend to look at support sites (save to DL updates), forums etc.: I had no idea even that the Feature Request forum existed until a few days ago. I simply haven't had the time. My apologies also for the need to split this into three parts: some twit has decided individual posts should be limited to ten-thousand characters (probably some predefined default in the software). Fair enough, but it might just possibly be an idea should that fact be indicated on the submission page: i.e. *before* post submission. Ah well...

Anyway, on to my own personal wish-list. I've read few a few pages of the forum in an attempt to avoid going over old ground, but a few old favourites do get a mention. A warning BTW; this will be quite extensive, simply because it struck me as more sensible to put everything in one place. Also, a few things at the end of this post relate to Accessibility for those unable (for whatever reason) easily to read the TI's visual display, but may still be of general interest. Having myself something of a sight problem, I am aware at least of two people who are totally blind and so need to rely on memory and the fact that the TI (whether by design or astonishing good fortune) is one of the most friendly pieces of equipment to operate without being able to see the display.

Right; here goes.

-------------------------------------------------------------------------

Classic Oscillators:

Oscillator Contour:

Modelled after the LFO Contour, this would add a staggering number of new sonic possibilities for the classic oscillators. Depending on how the waveforms are generated, this could be extremely simple to implement and should add absolutely no additional DSP overhead. It would allow also (e.g., should sine be selected) such things as morphing continuously between triangle, sine and square, with superb modulation possibilities.

Oscillator & LFO duty-cycle bias:

This would shift the waveform centre position within each duty-cycle with a negative/positive bias (a visual representation would show the negative and positive half of the cycle as stretched or compressed in relation to one another depending on the bias). Of course this has existed always for square wave (the simple `pulse width') and can in some fashion be done now for the wave-tables with PW (although of course actually the effect is quite different), but allowing it for all classic waveforms would open up yet another tremendous array of unique possibilities without adding any additional DSP overhead. As an example, for sine, certain settings would create an effect somewhat analogous to classic SY/TG77 (*not* DX7) two-operator PM, but vastly more versatile for creating unique sounds with a single oscillator and without the need for multi-operator FM/PM. For square, one could set one oscillator to a positive bias, the other to a different negative with differing amplitude, and generate some fascinating complex pulse-width combinations.

Oscillator & LFO Waveform compression/Expansion:

Not to be confused with any kind of audio compression/expansion, this would `squeeze' or `stretch' the waveform in relation to it's duty-cycle. For compression, a visual representation would look as though the wave had been `crushed' from each end towards its centre position, leaving a zero-amplitude gap at each end so the over-all amplitude and cycle length itself would remain the same. Expansion would do the opposite: increasingly expanding the centre zero-point whilst squeezing each half of the waveform towards its own respective centre-point (halfway between the over-all cycle centre position and each half-wave's respective end), again whilst not changing each half-wave's amplitude or the over-all cycle length. For sine, the sonic result of compression would be _somewhat_ similar to some grain-table effects (in reality the AM of the new frequency with the cycle fundamental), but would sound very different indeed for other classic waveforms. For expansion, sine would begin generating two distinct `half-sines', resulting in new complex harmonics somewhat analogous to self-ring-modulation. Again, depending on the waveform generation implementation, this should add no extra DSP overhead and open up yet another vast array of tonal possibilities. For LFO and given the correct waveform, a combination of this and duty-cycle bias would implement the `swing' effect some have requested.

Hyper-everything:

Although at present limited to saw, I don't see why this couldn't be extended to include some other classic wave-shapes (even if only sine, triangle and square) without adding any additional overhead.

Additive sine:

Although a full additive engine implementation would I imagine *vastly* be beyond the scope of any plans for the Virus, building e.g. a simple 9-voice additive sine engine based on the hyper-saw concept should be possible without further taxing the DSP. Each sub-wave could be tunable to harmonics derived mathematically from the fundamental, and the volume of each adjustable theoretically without adding any additional overhead, in the same way as hyper-saw; indeed I would imagine an optimised implementation might use even less DSP than the hyper-saw, since detuning wouldn't be necessary and the calculations would be fixed to harmonics. This would give the additive proponents a good deal of what they want without the need for any fundamental change in the TI's design philosophy. Theoretically, a sub-oscillator for each sub-wave also would tax the DSP no more than hyper-sub, since it would be derived from a divide on each sub-wave.

FM/PM:

In the same way as allowed for wavetables, a choice of classic analogue-style FM or DX-style PM for the classic oscillators would be an excellent addition. Self-FM would also be nice, although that might be getting into increased overheads and a lot of fiddling.

-------------------------------------------------------------------------

Wavetable Oscillators:

Harmonic Sine table:

A table allowing sweeping through fundamental-derived harmonics as sines would be a nice addition. I realise this can be done with formants, but that chews up a lot of power for something that could just be another table.

-------------------------------------------------------------------------

Envelopes:

Envelope Contour:

This would behave exactly as the LFO Contour, and allow a stunningly versatile envelope-shaping tool without any need to redesign the envelope algorithm. One could shift from linear to curves to a square-type straight-line envelope, allowing such things as delayed sharp attack, and probably a very close approximation of almost any analogue synth envelope shape one would care to name. Assuming the envelope generation algorithm is analogous to that used for LFO, this should not be more simple to implement with a minimum of time and trouble.

-------------------------------------------------------------------------

Filters:

Analogue emulations:

As so many others have requested, analogue HP/BP/Notch to complement the excellent LP. Other analogue emulations (classics, and also partialclassics such as e.g. the Roland MKS80 which has a unique sound) would be magnificent additions.

Digital: 6/12/18/24DB for each filter:

It would be superb also should one be able to set each digital filter between 6 and 24-db. This would allow up to 48db/Octave, but more importantly, allow far greater versatility when using the filters in parallel.

-------------------------------------------------------------------------

[Continued in 2/3]

[Continued from 1/3]

-------------------------------------------------------------------------

Modulation Matrix:

Envelope Trigger:

I'm not going to get into the Envelope Loop discussion. It's a superb idea, but I imagine will never be viable on the TI. Envelope Trigger however definitely shouldn't be too hard to implement, and would go a good deal towards satisfying many loop requirements if triggered with an LFO.

Oscillators and Noise as sources:

I wasn't going to mention this since so many have done it already, but it could be so fundamental to creating new sounds that I might as well do so in the faint hope that something can be done. A while ago, somebody asked why you would want e.g. to modulate filter cutoff with an oscillator. My only answer is to suggest trying it if possible on an analogue (or conceivably the Applied Acoustics Tassman, although I've not tried this and it might just die a horrible death). Failing that, have a listen to Isao Tomita's interpretation of Pictures at an Exhibition, and the pseudo-speech at the beginning of is version of Holst's Planets. Both the voice effects in Pictures and the pseudo-speech in Planets were done in part with very subtle filter-cutoff FM, or at least the same effect can be achieved. If properly done, the results can be astounding in a way that simply can never be achieved with static filter frequencies. I believe it's also much of the reason synthetic formant speech *never* sounds real; everybody seems to forget that the larynx is frequency-modulated subtly by the vocal cords, and never include this in the emulation. For a completely different example of filter FM, have a listen to the TB303 Devilfish mods (http://www.firstpr.com.au/rwi/dfish/) As a last example, try the following on a synth that allows filter FM from it's oscillators: feed white noise into the input of a key-following band-pass filter with resonance high enough to give clean, well-defined notes without moving into self-oscillation, if possible with it's centre-frequency synced to an oscillator (results will be very much better if you can do this). Then bipolar-modulate the filter cutoff subtly with a second harmonic sine (i.e. twice the fundamental) also synced to the same oscillator. With modulation/resonance adjustment, You should be able to achieve a set of odd-harmonic filters analogous to blowing gently through a clarinet or wooden flute with a single filter. Add this to a clarinet or flute emulation, and if you're lucky, you'll have something very much like a real acoustic instrument and impossible otherwise to emulate without complex physical modelling. Unfortunately, as somebody else pointed out, it may very well be impossible accurately to modulate sources in the TI at audio frequencies without ending up simply with a complete disaster. A pity should this be the case, since again, this is a unique effect otherwise impossible to achieve.

-------------------------------------------------------------------------

Portamento:

Portamento Contour:

Yep, I know; Contour again. But once more, this should be very simple to implement (assuming of course the portamento curve is based on a similar algorithm to that use to generate LFO shapes). If carefully designed, it should be able to satisfy _all_ the requirements of linear/exponential portamento whilst offering some vary unique variations available in very few (if any?) other synths.

Portamento Glissando:

Others have mentioned this one, so I won't say any more other than to urge it be done if possible. An envelope-trigger option for each glissando note-change also would be superb.

Portamento polyphonic Chord follower:

Not really a high priority, but worth mentioning. I'm not sure on how many synths this feature exists (certainly it's present on the old Roland MKS80 Super-Jupiter). Basically, it ensures polyphonic portamento tracks note positions: i.e., assuming you play a four-note chord, release it, then play another, portamento will track the changes such that the lowest played note in chord #1 will track to the new lowest note in #2, the second to the new second, etc.. In many situations, this can ensure a very musical glide between the two chords, rather than a random shift depending on which notes you happen to have pressed first in the new chord.

-------------------------------------------------------------------------

Effects:

Distortion:

All distortion types as filter-saturations:

Again, I won't say more, since others have requested this already.

Frequency Shifter:

Frequency Shifter feedback:

I imagine some people are saying `why the hell do you want that?', but the answer is simple. Frequency Shifter feedback would allow the creation of continuous phase-shift effects the resonance of which is defined by the feedback amount, something that can sound absolutely superb at very slow rates. Imagine a phaser that forever `climbs' or `descends', with no identifiable start or endpoint to the cycle. A unique and brilliant sound for lush, swishing strings or synth or guitar leads, *very* rarely heard nowadays, since frequency-shifter feedback effects cannot be created in any other way. At higher feedback moving into self-oscillation, you can create fascinating climbing or descending oscillators that have no start or end, and appear to climb or descend eternally. At higher rates, you move into very bizarre and unique 60s-style BBC Radiophonic Workshop Sci-Fi effects, when frequency-shifters were far more common. The only problem might be that there could be too much latency in the Virus frequency-shifter algorithm to emulate analogue FS feedback, in which case the sound would end up very far from musical (a sort of mangled combination of comb-filter and mess).

Delay:

Delay Feedback phase inversion:

Can this be done already? I'm not near my TI at the moment, and believe it or not, I can't recall off-hand. I'm fairly certain however that it cannot.

Reverb:

Classic Spring Reverb emulation:

This is something I've *never* seen work or come close, _ever_ in an emulation, save as a convolution effect. But I suppose there's always a first time, and I imagine it could go down very well should it be possible. With Access's attention to detail, they might just pull it off. To be honest, I've never understood why it seems to have proven so elusive or difficult for *anybody* to emulate.

Phaser:

LFO shapes:

Something that has always bewildered me a little is the question as to why the phaser's LFO is limited to one shape. It's completely out of keeping with the other effects, and seems to make no sense other than the fact that possibly the internal patch format never allowed for a phaser LFO wave-shape value component. Of course I know it can be done from the Matrix, but why not simply include all LFO possibilities as part of the phaser modulation settings in the same way as for other effects?

vocoder:

Classic filter emulation:

About the only thing that has disappointed me consistently with the TI is it's vocoder. Although reasonably useful as a formant filter effect, it simply does not come close for true vocal vocoding. The problem is very definitely in the filters, which sound (and look) nothing like the spectral characteristics of a true analogue vocoder. I've tried emulating several, including the Roland sVC350 (one of which by some *miracle* I have still in almost perfect condition after nearly thirty years), but the results were extremely disappointing. I'm not asking for prosoniq-style analogue emulation; simply an alternative digital representation that more closely approaches vocoder band-pass filter characteristics.

Effects insertion points:

Again, I know others have mentioned this, but I think it's worth it once more, just in case. A simple implementation could be to have a value displayed beside each effect in a multi-effect chain, indicating it's default insertion point. Changing this value would simply alter that effect's position and increment/decrement the other effects in the chain accordingly. This would avoid a load of appalling GUI garbage, and make it very simple to understand where everything was without giving everybody a headache or changing the UI design philosophy. Of course, I suppose if people *really* wanted to drag and drop things around in VC... Yuck!!! Yep; definitely not a GUI proponent; hate the damn things!

-------------------------------------------------------------------------

[Continued in 3/3]

[Continued from 2/3]

-------------------------------------------------------------------------

New Effects:

Binaural circular Doppler:

A decent binaural Doppler emulation is almost impossible to find (I've heard only one that I found convincing; a private addition to CSound from many years ago that was never submitted), but can create truly astonishing effects: sounds swirl around in eternal circles or vanish into an infinite distance. Obviously, a simple `single pass' would be pointless, but a binaural Doppler that could be modulated in the same way as chorus would be brilliant, particularly with vary careful attention to detail. BTW, don't anybody bother looking to implementations such as the Doppler effect in Adobe Audition for examples of a good Doppler: compared to the CSound version I heard, that joke doesn't even begin to come close.

Octivider:

This is something seen *very* rarely nowadays (the only emulation I've found is the Audacity plugin, with which I've had no success at all), but a pre-distortion classic two-stage Octivider emulation such as the old Roland pedals would be tremendously useful, especially for monophonic leads and in situations in which the TI is being used as an effects processor.

-------------------------------------------------------------------------

VC:

Accessibility: Button, control and Graphics text labels: Arpeggiator:

I can't confirm any of this, as I've not tried it with Screen-Reading software intended for the blind, however it appears there are some accessibility issues with some of VC's controls showing up as un-labelled graphics (or not showing up at all?). Arpeggiator user-pattern creation appears also to be impossible to use without sight(?). Perhaps if there is somebody (either blind or otherwise who requires assistive technology software) who is reading this, they could comment on VC Accessibility?

-------------------------------------------------------------------------

Manual:

Corrected and fully-documented manual and parameter reference:

This is the one area where unfortunately the TI reference manual in particular leaves a *lot* to be desired. The TI having grown and improved so much over the years, the manual bares now very little resemblance to the TI as currently it stands. The only way to reference its functionality is to wade through both the original manual and the addendum, trying to cobble together a coherent picture from the various changes along the way. Whilst this is tolerable (if far from ideal) for most, it is an absolute nightmare for somebody unable easily to read the display or who relies on their memory and on Screen-Reading software to use the TI. With no continuously updated reference, they very quickly become lost as parameters are added with no reliable point of reference to the increasingly outdated original material, and the various addenda which often are less than clear on where a parameter can be found, particularly in relation to parameters that existed before it was implemented. Worse, with no complete parameter reference (such as a complete table of menus and modulation sources/destinations), somebody without sight has no way easily to know what is available. Of course a complete parameter reference would help everybody; it would be good to know at a glance exactly what can be routed where and where in the list a particular source/destination could be found, without having to rely on scrolling through the sources/destinations. As a competitive example, the Yamaha TG77 manual in general was absolutely appalling, but included one thing that negated all its many faults and made the TG/SY completely and immediately usable; *every* parameter was clearly and succinctly documented, with every menu laid out clearly and their position in that menu exactly and succinctly described. So long as one understood the atrociously-explained concepts, ambiguous wording and the general fact that the rest of the manual was a mess, the synth itself was usable within minutes, even to somebody with no sight at all. I know; I saw it done. By contrast, the TI manual is in general superb, but lacks the one thing that would make it immediately fully usable for somebody blind or otherwise unable to read the display (and probably many others as well); it lacks a simple, clear, concise, fully-documented parameter reference.

-------------------------------------------------------------------------

Well, that's it. I did say it would be extensive, but hopefully not _too_ verbose, and a clean, concise list of those features personally I should like one day to see.

Thanks for the thread link; I'll take a look and let them know if it's of use also to them.

As for Additive: like every feature I've requested, I've tried to be as careful as possible to strike a balance between what might be possible and what I believe is likely to happen given Access's time and resources. Also, I've tried to ensure every request builds on existing functions or algorithms (so far as I can speculate as to how the TI Engine might be coded) so they might possibly be implemented with a minimum of time and trouble. Asking for a full Additive engine from scratch would require a great deal of investment in time and R&D. While Access could very well think it a good idea, the cost could seem simply too prohibitive. but something simpler initially modelled on the hyper-saw just might be feasible. And should *that* part be done, then perhaps it may not be too much of a leap later to introduce other things such as individual partial phase-shifts etc.. In a sense it's trying to encourage a company gently towards a feature you'd like to see and which you hope others also might want, without overwhelming them on the first try.

I imagine this would be the case for a good many ideas we may never see, or that might take years to mature. Just think of how fascinating it might be to get hold of some in-house, experimental OS versions sitting in a TI used for testing new ideas? Who knows what one might find?