EMUSIC-L Digest Volume 28, Number 6 This issue's topics: A Week with the Wavestation Your EMUSIC-D moderator is Joe McMahon. Please send articles to emusic-L@auvm.american.edu (or EMUSIC-L@AUVM.BITNET). Administrative mail should be directed to xrjdm@scfvm.nasa.gov. Digests will be distributed semi-monthly to monthly. For faster response to questions, subscribe to EMUSIC-L@AUVM.BITNET (or emusic-l@auvm.american.edu). The EMUSIC-L discussion group can alo be accessed via bit.listserv.emusic-l on Usenet. Subscribers to this list may send articles to EMUSIC-D@AUVM.BITNET (or emusic-d@auvm.american.edu); they will be reviewed and forwarded to EMUSIC-L. Back issues are available from xrjdm@scfvm.gsfc.nasa.gov. This is a manual operation, so requests for things like "all back issues" will take a long time. ------------------------------------------------------------------------ Date: Thu, 16 May 91 16:29:52 BST From: Nick Rothwell Subject: A Week with the Wavestation [Something I posted to rec.music.synth this morning. - Nick.] Well, I've had this Wavestation for just over a week now, so I thought I'd let y'all know how I'm getting on with it. Lucky you. To save you jumping to the bottom and reading the conclusions first: The Wavestation is a pretty good machine, with some nice ideas, but compromised by a weak filter section. It has a particular niche in my setup, but isn't the do-absolutely-everything-from-rich-warm-analog- -to-wonderful-new-sounds machine which Korg seem to think it is. KEYBOARD and SOUND ON SOUND thought it was fantastic; I wouldn't rate is as highly as they did, but it does the job, and it works properly, unlike the keyboard it replaces. THE PHYSICAL: The machine is small, light, well constructed and looks good (thank goodness for that). It's a lot smaller than the VFX which was a complete tank. Controls are minimal but perfectly adequate since the software interface is very good. Good buttons which don't sink into the front panel (good-oh), nice ratcheted infinite-travel optical data entry wheel, etc. etc. None of the glitch problems of the VFX. No labelling of back panel connectors on tpo of the machine - it would spoil the styling - which is a minor niggle. Nice graphic display although a little more contrast wouldn't go amiss. I notice in the manuals that the pictures of the WS have a WS logo on the top and back; I presume the case artwork was changed at the last moment to just say "Wavestation". SOUND QUALITY: Hmm... Dry. Very Dry. Martini Rosso, Sahara Desert, that kind of dry. The D-50 sounds positively soggy beside it. The WS will deliver pretty good rich strings but dry ones. I must try to synthesise a rainstorm on it, 'cos it'll be a dry one. The WS lacks punch to an extent - it hasn't got the clarity of the VFX, which is a real shame, because I loved the VFX for its bells and orchestral lead lines. The Waldorf can beat it for clarity and punch as well because, well, let's face it, that's what PPGs were bred for. And, I haven't managed to get a deep, window-rattling bass out of the WS yet; again, the VFX was pretty good at this. You can get some punch out of the WS for leads and the like by pulling various bandpass tricks with the FX section, but apart from that, I don't know. The pads I have on the D-50 are brighter and clearer than those I have on the WS. And wetter, of course. FACTORY PRESETS: My god, they don't half stink. A few lead lines and dozens and dozens of breathy new age pads which are totally useless to anybody who doesn't record for Windham Hill. They're trying to make the WS sound like an M1, with a few D-50 imitations here and there. I listened to about half of them and then wiped them from my machine. Korg UK are now giving away free RAM cards of PD patches to WS buyers - I suspect the factory presets are crippling WS sales in the same way that they crippled the Waldorf. PROGRAMMING INTERFACE: Very nice indeed. OK, so maybe the trend is back towards knobs and sliders a la JD-800, but that's not really possible for a machine as sophisticated as the WS. The WS has few controls but uses them logically. The page heirarchy is, for the most part, sensible, and you can define your own markers and jump between them. You can type in all parameters from the keypad, including non-numeric ones like effects configurations, which is great - I have this habit of typing in envelope times and things rather than using a knob or slider. Getting the 4-voice configuration by typing "2" is a bit weird, though. VOICE ARCHITECTURE. Ah, yes. Here we go. Well, as Metlay has said, the vector mixing is what characterises this machine. The wave sequencing is certainly nice to have, and it boosts the power of the machine considerably, but the vectoring is central, I feel. Regarding the wave sequencing - main weakness is that sequences are either always tied to MIDI clock, globally, or run at the WS's internal tempo (100bpm, or thereabouts). Better would have been an explicit tempo per sequence, with MIDI sync local to each sequence. Oh yes, the start point modulation is programmed into the sequence rather than the patch. Hmm. The top-level is the performance which is an 8-way patch configuration. Each part can be transposed, tuned, tied to a system or user tuning table, the usual stuff. It can also be a MIDI transmission part. Splits and velocity switches are programmed graphically. Oh yes, there's a delay parameter which also works for MIDI transmission so you can hold a note with sustain pedal (I think) and external modules' voices will come in at predetermined delays. I rather like this. Voices are held if you change performance. Good. Were this not the case, I wouldn't have bought it. Back to the voice architecture. The samples are OK, although I haven't investigated them fully yet. The choirs and strings are very good, the acoustic instruments are weaker. Korg make a great deal about the machine's wonderful, versatile modulation matrix. Bleah. It's respectable but not world-beating. As someone who's used to the Waldorf and shared a room with five Xpanders, the WS isn't all that hot modulation-wise. The vector mix position follows a 2D envelope and can be further modulated, which is good. As for the rest of the voice architecture: the voices have independent signal paths and can be routed to the four effects busses via a matrix (good), and there are two LFOs per voice (good, one is never enough) but there's only one spare envelope available for modulation (assuming you don't want to dedicate it to the filter). Each oscillator has a private pitch ramp section, but another envelope for fancier detuning would have been very useful. A nice idea would be to make the joystick position, or even better the vector position, available as a further two modulation sources, but no, they missed that trick. The joystick is only available elsewhere for effects modulation (and you can't disable it for voice mix). The vector position is private to voice mix. Minus several million for good ideas, guys. It's possible I'm missing some tricks here - after all I've only had the machine for a week - perhaps the trick is to vector mix between voices with wildly different modulation routings or something. FX bus routing is pretty good. Pan control is OK but rather strange - you have to do the routing on the effects buses, and in serial configurations you can't pan the voices. There's a hard-wired (boo hiss) modulation page for keyboard/velocity control of pan position, and the WS doesn't respond to controller 10. Oh well. Not too bad, actually; routing the voices left and right and using the vector mixing gives plenty of stereo animation, as do the effects processors. If I want flashy pan control, I'll use the Waldorf. Filtering: disappointing. Low-pass filters only, no resonance although there's an excitation parameter which seems to affect the upper harmonics in a not terribly useful way. The filtering is where Korg dropped the ball. Perhaps the M1 and T series has gone to their brain, and they thought they were building another sample-playback machine. So, instead of a wave synthesiser like the PPG where you can do amazing things with wave sweeps and filter resonance, we have something more like a flashy programmable sample player. Resonant filters would have made the Wavestation totally devastating. bandpass/highpass filters would have added considerably to its versatility - I used highpass filtering on the VFX for all kinds of things. But no, the WS has a set of perhaps-we'd-better-put-some-in- -so-that-we-can-call-it-a-synthesiser lowpass filters and that's it. The WS comes loaded with a large number of single-cycle VS waves, but without resonant filtering I'm at rather a loss for things to do with them. The best trick I've found so far is to hard-sync them to (a silent) oscillator A for PPG-like effects. Perhaps because of the lack of good filtering, the WS requires a lot of animation in its voice patches. I'm using a lot of layering, detuning and envelope pitch control to thicken up the sounds. And this eats up voices. The WS can run multitimbrally with sixteen channels each playing eight patches, but I think it's being a bit optimistic here. I think voices are going to run a little short if I try doing too much with it. EFFECTS PROCESSORS: Unremarkable but respectable. It's certainly good to have two of them which can be run in parallel - excellent for multi-timbral setups. The exciter and phaser algorithms are pretty good. The chorus algorithms are passable, but not as good as the VFX, whose 8-voice chorus configurations were fantastic - you could run a boring bass tone in one end and get Peter Gabriel's Passion out the other. The reverbs don't seem too remarkable. What the machine is missing is the ability to run the effects in parallel and then choose between FX1, FX2 or dry to outputs C&D (i.e. three-way routing) which is a shame. MIDI SPEC: As a receiver, the WS seems fine, though I haven't played that much with it yet. It can receive on several channels via a number of multisets. Here, the multisets carry effects settings which override the performance settings, so I find myself switching between Multi and Poly mode quite often (easy in Performer which I've programmed to send the 11-byte SysEx message required. I'd have used the MRC, but its maximum SysEx message length is, yes, 10 bytes). I think the VFX handled this better. It only had one multiset but you could tell it to adopt the effects configuration of any patch using patch change messages. Having to program (or copy) them into multisets on the WS is a bit of a pain, especially since there are only 16 of them. Multisets don't have names. And, they can't be selected via patch changes - you have to use a short SysEx message (I've programmed my MRC to do this on reception of controller #0 messages). As a master keyboard: excellent. Very impressive. It can hold voices on remote modules as you change performances, keyboard splits, transposition values, whatever. The VFX thought it could do this as well, but it was wrong, alas. What's even more impressive: the WS keeps track of whether it has access to each MIDI channel, for its global controllers (pitch wheel etc.). So, if I change from a performance with MIDI channel 1 in Part 1 at the top of the keyboard to a performance with MIDI channel 1 in Part 5 at the bottom, it knows that it can still see channel 1. Once I select a performance with no part to channel 1, it realises that it's dropped the channel and resets the pitch-wheel and controllers. I'm impressed. A slight hassle is that it insists on resetting volume as well, which is an inconvenience for various reasons. Anyway, as far as I can tell without more extensive tests: the WS will make an excellent live master keyboard, since it seems to use the same trick as the VFX - multitimbral reception independently of local and MIDI control from a performance selected manually or by footswitch. The lack of poly pressure I shall live with - reliability is slightly more important. STORAGE FORMAT: Performances refer to patches indirectly, patches refer to wave sequences indirectly. This means you have to keep track of what refers to what. The WS has two banks of performances, patches and wave sequences, and it's best to keep each bank self-contained so you can move it from one bank to the other or to/from cards etc. and keep the references intact. This organisation means you can't audition performances without the correct patches in place, and can't audition patches without the correct wave sequences. A minor hassle but one to be aware of. SYSTEM EXCLUSIVE: Sensible, which is high praise from me, I can tell you. It's documented fairly fully in the reference manual - I was worried for a while because the WS is American-designed, and Americans never like to give their users useful technical information. Writing the librarian modules for patches and performances was pretty simple. I'm still not sure whether names are 15 or 16 characters, though - they seem to be 15, although the data formats suggest 16. One annoying thing: the WS insists on reporting something like "SYSTEM EXCLUSIVE RECEIVED SUCCESFULLY" (sic) every time it gets anything, so you can't audition patches and see the names on the display right away. The librarian for multisets should be easy, although the things aren't named which is a shame. Wave sequences are going to be interesting because the things have variable length. I think I can do it with minor structural changes to the inheritance heirarchy in my editor/librarian, but it's going to be interesting. RELIABILITY: Seems fine so far. The hardware feels solid, which is a change from the VFX. The software, despite its complexity and sophistication, seems good and solid as well. I'm impressed. Even the little details of MIDI external control seem to have been thought about and got right, more so than for a lot of so-called master keyboards. I've managed to crash the software a few times by sending fast combinations of SysEx dumps, so I think there's a wee buglet lurking in there somewhere, but apart from that it seems totally solid. This is a new experience for me. CONCLUSION: Nice machine, on the whole. Korg really dropped the ball with the filters, though. The WS is no does-everything machine, and is weak in a number of areas, but it's well-designed, solid, and very versatile in the areas it does address. It'll do. Nick. -- Nick Rothwell, Laboratory for Foundations of Computer Science, Edinburgh. nick@lfcs.ed.ac.uk !mcsun!ukc!lfcs!nick ^^^ "The tabla is an organic instrument. We use the hammer for tuning. ^^^ ^^^ And also for teaching." ^^^ ------------------------------ End of the EMUSIC-L Digest ******************************