Goodbye Tempest, Hello Freedom.

Well the Tempest turned-out to be a damp squib for me. Or rather DSI’s inability to implement features that were hotly discussed over a year ago. Maybe they will eventually pull the rabbit out.

Even so, there seem to be a lot of Tempesters who are perfectly happy with the current OS, and that has to be respected. As with everything else in life, people will have a different approach to, and expectations from, such an instrument. Everyone should come to their own conclusions.

[youtube http://www.youtube.com/watch?v=QereR0CViMY&w=400&h=255]

For me, the hardware interface, the sound and the overall concept were incredibly appealing. The forum was abuzz with excited discussions about the untapped potential, and we looked to be in line for a classic that would finally put the Roland box voodoo to bed. A powerhouse on paper, a legend in the making.

Now don’t get me wrong – in some respects it’s nearly there. I had fun designing sounds on this machine and I learned a lot along the way. The glass is definitely more than half-full, and I certainly do not regret the purchase.

But the sequencer is still too rudimentary for my needs. I didn’t see anything in the recent betas that give me confidence in user-requested improvements beyond the necessary bug fixes. DSI’s prolonged silences do not help. To me, most of the requests seem reasonable, and there appears to be broad agreement on the most important. The persistent bugs, though not serious, reveal a lack of attention.

Nobody is asking for a Cirklon-killer – only that the machine is given a sequencer worthy of it’s potential. And it seems so tantalisingly close.

At current pace, by the time the OS is ready, the second-hand market will already be mature, new competing products will have been released, and DSI will have lost sales. Brand damage has already occurred. I am amazed that the can’t see this! I would have said ‘bemused’, but it’s actually quite sad. I really want DSI and their products to succeed – the innovation is wonderful – but there seems to be no ambition to finish the job.

I can only speculate that the feasibility of new features is dependent on internal hardware limitations such as the the 4mb limit (honestly, what were they thinking!!), rather than the coder’s inability. He seems talented, if disorganized and overworked and, it would appear, fighting some serious battles to squeeze better performance out of this box.

Nevertheless, my biggest disappointed has been their habit of raising expectations and not following through. I honestly didn’t expect that, and I fell for it wholeheartedly.

They are also not giving the game away regarding what’s possible or likely to reach implementation. Engaging in further feature discussions seems pointless, as none have materialised. Beta testing has been the simple repetitive process of pointing-out the same few bugs in every release.

Admittedly my passion boiled-over after seeing the latest delay (of several months) was due to focus again shifted to the Prophet 12. OK guys, but at least tell us! It would be ridiculous to expect new product development to stop, but it is also reasonable to expect that existing products should receive the necessary attention.

In short, I am extremely disappointed with DSI’s behaviour. I would not have accepted this from another company, but I was rooting for DSI to succeed, and I trusted them. More fool me.

As a sweetener it was revealed that Tempest and P12 share a common codebase, and DSI have said that features will be rolled-back into Tempest. This sounds great, but makes me wonder, since the Oscillators are completely different, and P12 has no sequencer – the element that needs most attention. Time will tell.

I do earnestly wish DSI success with P12. It looks like it will be a winner and they have clearly worked very hard on it. They must be congratulated for the risks they are taking to innovate exciting new products. In the grand scheme of things, complaining about a few missing features seems churlish.

But when it comes to Tempest, I’ve simply lost faith. It had to go.

With the proceeds I have fostered a beautifully-built MIDIbox Seq4. Now THIS is a sequencer. It completely obliterates DSI’s effort. Open-Source with frequent updates still after 5 years of development. All that user input has been eagerly gathered and distilled into a great interface and a smooth, almost effortless workflow. And zero fuss from the unpaid developer.

IMG_2010

There is more than enough cash remaining to start a nice modular and have a big party. I’ve already sourced the PCBs to make my own drum modules, which will be triggered by Seq4’s 8 analog outs. I can simultaneously sequence FM percussion provided by FS1r. What I lose in compactness I more than compensate for in flexibility and power.

Not only that, but there is still change to cover the entire expense of my 4×4 build AND to a fund a new build using a new third-party ladder filter in a 2-octave keyboard enclosure.

My first Eurorack module is ready to be assembled. Little Dieter is due to be born any day now…..

IMG_2007

I’m sticking with open-source hardware from now on. No more ‘black boxes’. It’s AD 2013.

There will be no pining for Tempest.

Here are some Kits and sounds – the ones I managed to backup before the new owner ripped the Tempest from my arms. FWIW, he’s absolutely delighted with it.

Have fun!

I dedicate this one to DSI 😛
(One of my favourite basslines and easily possible on Tempest)

[youtube http://www.youtube.com/watch?v=e4OtIb6wDCQ&w=400&h=255]

Tempest Kits: Starship Stim v1

OS 1.3 is coming soon and it’s looking good. So I’ll be pushing out some ‘proper’ kits after it’s release.

This first kit is a taster – it has no drums. Rather, it is the soundtrack from an imagined life aboard a roving interstellar craft. No samples used – this is a nice workout for Tempest’s analogue side.

The top row of pads are drones – various ambiances, the bottom row are computer and ship noises. Bank B is fully stocked – 32 sounds in all.

To activate a drone, press one of the top 8 pads. The sound will continue until it is stopped or it’s voice is stolen. Several of the drones are designed to ‘choke’ each other so as to help prevent a cacophonous mess. But that’s largely up to you. It’s easy to go overboard.

To stop a drone, select a sound by pressing Shift + pad and press the Mute button twice. By pressing the Solo button twice, all sounds will be cut except the currently selected pad.

It takes a little practice but learning to coordinate it all is good for improvement of live tweaking chops. YMMV.

Use the volume and filter controls to get a nice balance between the drones and twist the bottom eight pads to your preferences. There’s lots of LFO action going on here, and slight tweaks can lead to nice surprises.

All sounds are velocity sensitive. Some are pressure sensitive. It is recommended to use seperate outputs and lots of FX 🙂

Play around with them (they don’t bite) and have fun!

Here’s a quick demo direct from Tempest’s main output. One take, no sequencer. A fraction sent to reverb.

I bet you can do better, so:

Here is the Kit:
StarshipStimV1_Beat

..and the individual sounds:
StarshipStimV1_Sounds
These will likely import to your ‘User3’ Folder.
Exported from Beta 1.2.6.18. I will update these as new betas are released.

If you make something nice with this, leave a link for da Massive.

I am open to Kit suggestions, but these must be backed by DRY audio examples.

Lemme know 😛

Tempest Recipes: Claps

I received a stern e-mail from the ether demanding to know when I’d provide a clap recipe. Rather than disappoint, and to avoid any risk of painful retribution, I’ll channel some ideas.

In the tradition of previous recipes we’ll nail-down the basic concept, leaving you with abundant noodling fodder. Let me preface by saying that Tempest makes fantastic claps!

The essence of the clap is that it is several sounds in one, each offset by a tiny amount, differing in volume, pitch and other nuances. Think of several people clapping at the same time – they will never hit at exactly the same moment. Multiple individual sounds separated by a few milliseconds produce a chorus effect that responds beautifully to reverb. Clap sounds are as varied and important as snares, and can scream individuality. So let’s make a basic clap template that you can twist to your own desires.

Initial explorations looked at using Tempest’s MIDI delay, but the available delay times are not short enough. Then I tried using square and sawtooth LFOs to quickly turn Oscs on and off. There’s a Guiro in there somewhere. Not great for claps though. If only there was something like a ‘crack module’ as found on some software synths….

The good news is that we do have a way to do this in Tempest: we create custom ‘cracks’ by using the envelopes and their (blink and you’ll miss it) delay parameter. Not to be confused with the MIDI delay, this parameter is accessed in the envelopes screen by scrolling 2 pages to the right and ‘delays’ or offsets the triggering of the envelope. With 5 envelopes we have bags of flexibility and control, more than enough to rival the crack of dawn.

The question is how exactly to use the envelopes? We could use one envelope per oscillator and have 4 sounds play in rapid succession, or we use 1 or all 4 oscs mixed and modulate the VCA so that it opens and closes very quickly. Or a mixture of both approaches.

  • To keep everything simple for now, just load the ‘Resonant 4k noise’ sample into Osc3 and pitch it down to -12 semitones.
  • Turn Amount values to zero for all envelopes, including the Amp envelope. We want these under full velocity control.
  • We start with the main Amp envelope. Leave it’s delay at zero because it will sound first. I set it’s velocity amount to 127 so that it is always the loudest portion of the clap. If you don’t already know – to set the velocity amount for an envelope, it’s on the second page of the envelope screen. Alternatively, when on the first page in the envelope screen, press shift to reveal the Velocity Amount (‘VEL AMT’) control. Use a short decay of around 20.

We can set different decays for each envelope. They can overlap without cutting each other. This brought to mind some interesting layering applications that I have noted to explore in a future recipe….
Back to our clap:

  • Keep in mind that our goal is a quick succession of hits that trail-off. To begin our ‘crack’ go to the pitch envelope, scroll 2 screens to the right and set the destination as VCA level. Yes, the pitch envelope is freely-assignable like Aux1 and Aux2 😯

The Amp and Filter envelopes are hard-wired, however don’t forget that, if we wish, we can still assign them to VCA level in the ModPaths…

  • For now, give the pitch envelope the same decay (20) and a delay of 30. Scroll back to the first screen and start turning-up the velocity amount. You should gradually hear the first ‘crack’ being introduced. Leave it at around 30. Essentially, it’s the Amp envelope re-triggered with new settings. To make it more ‘clappy’, reduce the delay to a value of 5 and let’s move-on to the Aux-1 envelope.
  • Here we repeat the same procedure but using different decay(25), Velocity Amount (40) and delay (8) amounts.
  • Then onto Aux 2 envelope, rinse and repeat. As this is the last part of the ‘crack’ give it a slightly longer decay (30), lower velocity amount (20) and, of course, a longer delay (11).
  • From here’s it’s mostly about balancing the envelopes. The key parameters being delay, decay and Velocity amount, but don’t be afraid to experiment. Notice how, when switching between envelopes, the screen stays on the delay page. This makes it easy and fast to tweak the crack to your liking across all Oscs. Combined with the tonal power of 4 oscillators, 11 on-board clap samples, not to mention a healthy injection of chorus and reverb, I would be very surprised indeed if you don’t find the basis of your perfect clap in there.

With 8 ModPath slots there’s absolutely no excuse to go wild – using each envelope to modulate various other parameters (pitch, pan, etc) at each stage in the crack module. That is an enormous amount of clappage, snappage and bappage. More than I could hope to continue writing about. But then, you are probably way ahead of me by now…

Note: I did make a crude attempt to measure the range of the Envelope Delay. At an amount setting of 127 there is almost 6 seconds between initial sound and it’s sequel. With amount set to 64 the time interval is approximately 3/4 second, and at zero there is no repeat. An amount of 74 equates to around 1 second. This indicates a log-scaled control. However, for claps we shouldn’t need to become bogged-down in spurious precision when we can just trust our ears. The delay does get short enough and you already know what a clap should sound like, no?

Tip: Try this with your snares/sticks etc. with even shorter envelope delay times – you may be pleasantly surprised!

Tempest Recipes: Bells and alien idiophones

Based on a request from DSI forum I had a new look at making bells. I did try this before but I wasn’t convinced of the results. If I want a bell sound I typically go to FM synthesis. But I had another look and came-up with some nice results, especially on the weirder side (as the title suggests). Below is a basic recipe. Perhaps not the most natural sounding, but a start to your own efforts at least.

  • Triangle in both Osc 1 and Osc 2.
  • For this patch we want an harmonically pleasing sound so that we can play the bells chromatically. Start by pitching Osc 2 down up to C4 and Osc1 an octabve higher at C5.
    If you want more inharmonic, clangorous church bells, start with the pitch Osc 2 at D#1 and work your way up.
  • Detune Osc 2 to approx 30, or more for a more inharmonic sound.
  • Set the Amp envelope – AD mode with zero attack, and a fairly long decay (60). Give Amp envelope amount a good boost at 100. Now go to ModPaths and self-modulate the Amp envelope decay with a positive amount – around 30. This should already have at least some bell-like characteristics – play up-and down the keyboard. Similar to a toybox / glockenspiel maybe? Your ears can guide you on those paths. But a dose of LFO action here serves-up some quality variants.
  • LFO1 mapped to Osc2 frequency, random shape and a tiny amount – 1-2 shoudl be enough to introduce a slight randomness to the tail.
  • Filter 4 pole – Close cutoff just slightly (120) zero resonance and slowly turn-up the Filter FM (Audio Mod) knob. There are tons of cool analogue sounds here – everything from changing the tonality of our bell to weird-ass glitches as the resonance is turned-up. Modulate the filter parameters with both LFO’s and see what I mean.

In fact, at this stage I got totally diverted away from bells into crazy modulation land – with lots of great results. If you’re stuck on bells the above will at least get you going I hope.

Oh, and don’t forget to pile-on both chorus and reverb – both really vital here. You could also try compressing the snot out of the output – the more I use it the more I love Tempest’s compressor.
🙂

Tempest Recipes: Tuned woods, sticks and clicks

The classic way of synthesizing woody sounds is to use square wave oscillators. In extension to our cowbell recipe, the relative tuning of two Square waves throws-up many useful, if not entirely natural, woody sounds.

Quick and dirty, first we’ll play around with a single square wave so as to get a flavour of what’s possible and, critically, to set the envelopes correctly.

  • Osc 2 Square around D#2 to start. Initially set osc mix to 0/100
  • Amp Env amount around 60, decay from about 50 but play with it. This is a good place to add a little peak. For clicky stick sounds, reduce the decay to zero and slowly turn-up the peak. But keep in mind that real-world sounds do not end so abruptly, and are usually accompanied by some decay and/or reverberation.
  • 24db LowPass around 60 (just enough to lose the harmonic ‘fizz’), No resonance, No Filter FM. Filter Env decay 40, Env amount 40.
  • Go to Modpaths and modulate Amp Envelope Decay with the Amp envelope for a nice curve. There’s a lot of tweaking here – negative values gives more hollow tones, positive values shorten to a click – a sound in itself that has some applications.
  • To tighten-up the sound, set the HighPass filter to about 40 or so.
    Now bring-in the Amp Feedback to gives a pretty convincing ‘knock’ to the sound. Usual warning here – used the fixed velocity to set the Amp feedback to provide the heft, but just before ”squeaky’.

Now is the time to start changing Osc2 pitch and to introduce Osc1 as another square wave, playing with their relative pitches and rejoicing at the beauty of our world. Take your time. When you have a good balance between the two levels/pitches, use this nice undocumented feature to tune all oscillators whilst preserving their relative pitches: press and hold the oscillator select button until all four Osc light appear. The Osc Pitch knob will now tune all Oscs simultaneously. Very handy. By this method we open-up even more tonal options. It’s also good to detune the Oscs – even quite extreme detuning has it’s uses here, especially at higher pitches.

  • To add some spice, we can Set LFO1 to modulate PWM of either or both oscillators. Plenty of tonal variations here – you should try the full range of LFO frequencies and amount.
  • Another LFO trick is to modulate the frequency of one oscillator with a high LFO rate. This also gives a huge variety of interesting tones. For a dullish hit, similar to hitting wet cardboard, try using the Random LFO shape set to Osc2 frequency rate 150, amount 100.
  • Filter FM without the resonance makes the sound more flabbly, and effect you might like.

During the process above you have undoubtedly discovered that effective sidestick tones are achievable at higher pitches, whilst at lower pitches we enter usable bass patch territory (remember our Donk?).

The logical extension of this is that our woods can be played chromatically, but these tuned sounds will require extra attention to envelopes and keyboard tracking parameters to maintain a consistent progression up-and-down the keyboard. If you are willing to spend the time here you can make very convincing Marimaba / xylophone type sounds, not to mention everything in-between (which I personally find more interesting).