Kurzweil:Dead Letters 1996b

From Sonikmatter

Pulled from the KurzList by Brian Cowell

Contents 1 Calvin and Janis 2 Envelope Issues 3 Fat Kickin' Beats 4 FX Fakin' by Yash 5 GUITARS 6 Hidden Effects Part 76 7 K2500 Answers 8 K2500 SAMPLING OPTION AND INSTALLATION 8.1 INSTALLATION 8.1.1 Step One: Check the Engine Board Revision 8.1.2 Step Two: The Digital I/O Board 8.1.3 Step Three: The Main Sampling Board 8.1.4 Step Four: Finishing the Job 8.2 SAMPLING IN ACTION 8.3 CONCLUSION 9 LEGATO MONO 9.1 Idea #1: The minimal multisample approach: 9.2 Idea #2: The many-patch approach: 9.3 Idea #3: The false-portamento approach: 9.4 Idea #4: The alternate patch approach: 10 LESLIE 11 SET PHASERS TO FUN 12 SNARES

Calvin and Janis

Date: Sat, 1 Jun 1996

From: Greg Smith

Subject: disply going away slate 2

6.7 What's the scoop on Clavin and Janis?

The Janis chip replaces the Calvin chip. This is the "sample fetch" chip used by the K2000. Kurzweil made the change as part of a manufacturing update. In terms of the results for the end user, there is absolutely NO DIFFERENCE in features or capability. However, because of differences between the two chips, the Janis chip requires a slightly modified circuit board and engine firmware. You can tell what kind of board you have by whether or not there is a J after the version number displayed on power-up. If you are updating your unit to a new version of the operating system, you will need to know which version you have so that the proper engine firmware will be sent.

There is no damage caused by installing one version of firmware in the opposite machine; it will simply not work.

Envelope Issues

Date: Fri, 19 Jan 1996

From: nursingh@CLYDE.ENS.INSA-RENNES.FR

Subject: ENVCTL

I've made a sine wave keymap with just a sine wave.

Then I went into the sample editor and set the start time to a peak, so as to get a very sharp 'tick' in the attack.

In the manual, it is mentioned that ENVCTL does not control natural envelopes, and so I made a user AMPENV. When the attack is set to 0, the 'tick is heard. The 'tick' disappears when the attack time is set to 20ms.

Although you can set ENVCTL attack adjust to values > 1.000 to speed up the rate of the attack segments, the 'tick doesn't' reappear. ENVCTL does well its job of speeding up or slowing down of envelopes but as Blake Stone said the 20ms barrier remains.

The 20 ms coarse value therefore a real limitation : you cannot have realtime control over it.

A way around it is to tweak the sample start time in the sample editor. And to set the AltStart at another point within a 20ms range of the sample start.

The lack of effect of ENVCTL for the 20ms barrier cannot be heard if the sine wave does not start sharply. So Blake Stone : you were right.

Yash.

---

Date: Sat, 20 Jan 1996

From: nursingh@CLYDE.ENS.INSA-RENNES.FR

Subject: Envelope speeds

>. As a work-around, would it not be possible to use ENV2 as a pitch >control source with the envelope ONLY given the first attack segment >set to smallest time setting but with maximum sweep (ie from +100% >to -100%) with all other levels set to -100% and all other times set >to 0mS.

While experimenting yesterday, I did that on the sine wave sample, and after adding the Steep Bass DSP function in ALG1, achieved a nice 'analog' sounding kick drum.

I think it's a nice workaround : not too dificult to achieve, and quite effective.

Yash.

---

Date: Tue, 23 Jan 1996

From: David Fox

Subject: K2500 features

Blake Stone,

Yes, the K2500 (both rack and keyboard) features some extra values for the amplitude envelope. Although the basic 20 millisecond update rate is still in effect for most parameters, we have created a "special case" situation for the Att1 parameter, allowing values of 2ms, 5 ms, and 10 ms, in addition to the regular values.

As to the Sample Skipping and Alt Method parameters, these parameters are added in version 2 of the software, which is currently shipping in the 2500 keyboards (we have not yet released version 2 for the racks, but will be doing so soon.)

Fat Kickin' Beats

Date: Mon, 4 Mar 1996

From: belau

Subject: Fat Kickin beats

>I've just started to have clients come in and want to do dance >music. I'm curious how to get that super-phat kick drum the kids >today are so fond of. I tried layering 3 different kicks from my >Alesis D4 on top of stuff coming from the K2k, and got part of the >way towards where I wanted to go, but not all the way. Anyone got >any tips?

• you should try looping a kickdrum-sample, but just a single sine-wave from the end of the sample.most bassdrums end in sinewaves so it should be no problem to get good loop.

• then put an envelop (e.q. env2)with a short decaytime and no attacktime on the pitch and set depth to 1200cent.

• set the attack level to 0% so the kick beginns as it is and decay level to -100% so it rolls down an octave...

• beware of your bass-speakers: they might move arround realy heavy.;)

(if you have the resources use a highpassfilter to cut everything under 20Hz)

• the result should be a real deep bass with the beginning klick from a normal bassdrum.

EuC Belau

---

Date: Wed, 6 Mar 1996

From: "HECTOR, KEVIN"

Subject: Fat Kickin beats

>-then put an envelop (e.q. env2)with a short decaytime and no >attacktime on the pitch and set depth to 1200cent. >-set the attack level to 0% so the kick beginns as it is and decay >level to -100% so it rolls down an octave...

You might find the envelope does not move fast enough. Try adding an extra modulation source to the PITCH page called "AttStateCtl" (or something, can't quite remember what it's called) which is a very fast sweep from 1 to 0 each time a note is triggered. It is much faster than the 20mS envelope speed - but dont flame me with ENVCTL stuff !

BTW, try a triangle waveform rather than a pure sine, and put a lowpass filter in the algorithm which decays across the duration of the kick. Gives a more natural dynamic quality to the sound.

Kevin Hector

---

Date: Thu, 7 Mar 1996

From: Pawl Stevens

Subject: Fat Kickin beats

For a bass sound take a sine wave and transpose it down to the desired frequency. Select a filter, usually a low pass, and assign an envelope to the filter. Set every value to zero and select the decay. Move the decay time up and the decay level up to your liking. You might want to add some release, but that takes up voices.

Another thing to do is assign a quick lfo to the pitch page and assign another envelope to the LFO. If you can't hear the bass turn up your amp.

For a snare sound pretty much repeat everything above but with a noise waveform.

Cymbles are tough to synthesize. I just use a sample. The roland tr-909 uses a sample so no problems there.

Toms are the same as the bass sound except you assign an envelope to the pitch. Also transpose it up to a higher pitch.

For that resonant BLIP sound go to the resonance page and turn up the resonance and turn down the envelope. Also transpose up.

There is always sampling if you don't have time to make a tottaly unique drum program. And you can mix some of these techniques in with samples and get some really sweet sounds going. Also mess around with algorithms etc..

You can pretty much make ANY sound in the k2000 if you know how.

Welp that's what i know about drum programming in it's simplest form hope this is of some help.

Pawl Stevens.

FX Fakin' by Yash

Date: Thu, 4 Apr 1996

From: nursingh@CLYDE.ENS.INSA-RENNES.FR

Subject: Faking FX on the K2k - APRIL 96.

---

---

Author : Yash Nursinghdass nursingh@clyde.ens.insa-rennes.fr

---

The K2k has a global multi-effects processor. Independent WET/DRY mixes can be achieved with the previously-discussed panner algs method. Having different effects per channel is normally impossible on the K2000.

If your requirements are not too extreme, here's what you can do:

Fake program flange, delays and reverbs using envelopes, LFO's and other hacks.

1. Flange-type FX :

Use several oscillators, exactly same PITCH settings. Works best with DSP OSCs (Choose an alg where you can select SAW/SQUARE/SINE as a DSP function).

2. Chorus/phasing/flanging FX :

a. Use several similar oscillators with slight detuning between themselves :

OSC1 PITCH FINE : +11
OSC2 PITCH FINE : -13.

The best way to hear this effect is to make a simple layer, say a sawtooth on the keymap page, and ALG1. Then duplicate this layer. Now go to the PITCH page of one of the layers.

Maintain a key pressed, and then select the Fine field and vary it with the alpha wheel. You'll hear an instantaneous dephasing.

You can set the PITCH FINE values to +x and -y where x=y. However, the above settings (+11,-13) produce "beats" in the pitch variations (at all multiples of 11*13). Without beats, the pitch variation is regular and predictable. The beating effect is therefore a way to make the phasing less predictable especially when the values are prime numbers.

More appropriate PITCH FINE settings(a higher difference in pitches) can yield a CHORUS FX.

b. Use a NOTCH filter with a varying CUTOFF FREQ.

That is, set SRC1 or SRC2 to LFO1 or LFO2 on the FREQ page. Using LFO2 is a good idea since you can make it global, by gong to the common page and setting Global to ON. The phasing effect will then continually change, even when no key is depressed. Essentially, when you press a key, you're having a "window" on the phasing effect. I'll write more about what I call "timbre windowing" techniques in my text on wave-chaining and wave-sequencing emulation on the K2000.

c. Use Synchronized Oscillators, with a variable-pitch slave oscillator.

Oscillator Synchronization is a synthesis technique which can be found on some older analog gear. Use algorithm 30, for example. You need to set SYNCMand SYNCS's Keytrack parameters to 100ct/key. Set the 2 last DSP functions to LOPASS and AMP+ as an example. Set the Master oscillator (SyncM) Pitch to -12 and the Slave to 0. Set the slave's SRC1 to LFO2 and the depth to approximately 1600ct. Edit LFO2 to .15Hz and a triangular shape. Go to the common page and set Global to ON. The LOPASS cutoff should be fairly high so that higher frequencies may be heard. Now play staccato and hear the fake phasing or flanging effect continually "move".

Of course, you can then mix the ways of faking phasing and flanging described above by using different layers.

3. Delay/Echo:

a. You can tweak the AMPENV's release settings to be like the following for a simple single-tap delay:

_______
       \
        \       r2
         \      /\
          \    /  \
           \  /    \
            \/      \r3

Note the level of r2 which is less than that of the decay end-level, which adds to the realism of the echo.

b. Use a LFO : First you need to increase the AMPENV release time. Then go the F4AMP page and set SRC1 to LFO1. Set DEPTH to approximately 17dB. Set the main field of the AMP page to -17dB approximately. Choose the piano keymap for this example.

Go to the LFO page. Select a min rate of 1.00 and a max rate of 9.00 approximately. Set the RateController to Data. Select Square as SHAPE. Now play a key. You should hear the sound pulsing. However, you may not hear the attack of the sound. To remedy this situation, change the phase setting of the LFO.

Of course, you can also bypass this problem by using the SRC2 and couple the LFO with an ADSR envelope so as to delay the time when the LFO comes into effect. However, using the phase setting saves the ADSR for later.

If you are using a natural AMPENV, you should go to the LAYER page and turn ThrDecay On. With a user AMPENV, you may increase the release time so that the effect may be heard after note off. Moreover, If you want the effect to last when notes are sustained, you should set the decay level to a suitably low value and set the decay time accordingly.

The effect should be pretty realistic. This is because the AMPENV is following its normal course to complete death and the LFO variation is superimposed on it. Thus you have regular surges of amplitude and the amplitude of the surges themselves decay: echo number 1 on when the AMPENV is in its release section should be louder than echo #2 which itself should be louder than echo #3 and so on...

You can experiment with various LFO shapes and vary the "delay time" with the DataSlider if you own a keyboard version or any MIDI continuous controller otherwise. It is therefore possible to use this method to have independent delays with realtime control on each channel. The "wet/dry" setting here would normally be the DEPTH on the F4 AMP page which you cannot set directly in realtime if you are using SRC1 as mentioned above. You can however set the LFO MinRate to its lowest possible value. When the RateController is set to this value, the sounds dies out without the LFO coming into effect.

In reality, echoed sounds may not sound the same as the original sound; timbre variations may be induced by the distance and texture of the reflecting surfaces. An envelope may be used as SRC1 on the Cutoff Freq of a LoPass filter so as to make the "reflected" sounds duller. Or you can use a SHAPER or WRAP or DIST function to distort the reflected sounds when the AMPENV is in its release segments.

4. Reverbs :

You may turn to the use of an additional layer with the same sound being effected in various ways - with a DIST or SHAPER DSP function for example - for fake reverbs. Or else you can set a layer to generate some radically filtered noise with a slow attack and large release on the AMPENV page. AMPMOD and SHAPEMOD can also yield some good results.

5. For those of you with the sampling option and an outboard digital or analog recording device :

It might be useful to consider recording the K2k's output effected by the internal effects processor. Which means you will not be faking FX but using real FX, although realtime control over them will be lost this way (I would say sampled or fixed FX).

The K2k's Digitech FX chip is known to be noisy. This is especially acute when the K2k's volume and the WET level are set to their maximum values on such FX as Flanging and Parametric EQ.

The ideal solution would be to have 16 independent outputs and treat each output with a different effect with outboard gear, without using the Digitech chip. In practice, most of you will probably like to try their best with their current gear (I know I do).

To minimise the noise problem, here is what you can do : If you have version 3.xx you can go to the MIDI page, then press the CHANNELS button. There is a field called outgain which you can set to 6dB, or 12dB. You can even set it to 18dB if you are not planning to play chords because otherwise, digital distortion sets in (clipping).

With the earlier O.S. version you should make it a habit to save your own programs with the Output level on the OUTPUT page set to 6dB or 12dB.

Then you should revisit the effects themselves. Choose an effect and press EDIT. On most effects you will see fields called EarlyLevel, LaterLevel, or ReverbIN, ChorusIN in the multi-FX configurations which usually incorporate a mixer. These levels should be set to their maximum values wherever possible. After this, a program which used a 50% Wet/Dry value should be reprogrammed to use less for the same sonic effect. And since the Wet/Dry setting is to be incriminated in the noise generation, the program will sound cleaner.

The final step is to set the Master Volume Slider approximately to its middle value to reduce output hiss. You can then choose a program, a suitable effect, and record it to your outboard gear.

The recorded effected sound can then be sampled back to the K2k and the process can be repeated a few times if necessary. In addition, you will still be able to process the sample with the realtime DSP function of the VAST engine.

6. Those of you lucky enough to have a K2500 with the sampling option can start having fun with sampling a sound with its effect. Since normal K2500 operation is not frozen during sampling, you can consider all types of realtime parameter changes during sampling.

In addition, the forthcoming KDFX option will hopefully enable K2500 owners to have independent effects on several channels.

Conclusion :

You can achieve quite complex fake effects if you combine the different types of fake FX described above.

Upside:

• Super-clean fake FX (no noise is added since the echo/delay is built from the sample, the DIGITECH chip is not used)
• Different types of FX per layer
• Different WET/DRY of FX per layer
• Added flexibility
• Realtime control over parameters through ENVCTL, and usual realtime controllers.
• Fakeness is not so noticeable in a Multi-Timbral mix. You don't have to spend $$$$ on that Eventide DSP4000 Harmonizer yet, because you'll be busy experimenting for some time.

Downside:

• Fake FX (wait till you hear the ramblings of those who own Eventides and Lexicons about how this post is totally useless).
• Lot of work - much tweaking by ear.
• Limited types of FX to which this hack can be applied (echo, delay, chorus, reverb - much more work plus you may have to use separate layers).
• Limited parameters for editing.
• For multi-timbral mix coherence, much more work is needed.
• Takes up polyphony (whereas the DIGITECH chip doesn't).
• Owners of outboard FX will think you're crazy.

Yash.

nursingh@clyde.ens.insa-rennes.fr

GUITARS

Date: Thu, 4 Apr 1996

From: terry Tabor

Subject: Guitar samples

> Speaking of emulating instruments.. > I want to emulate a good distortion guitar, along with some power >chords, etc. I'd like to achieve the kind of gritty guitar sounds >you find in a Ministry recording, for example. > Has anyone had any success creating good electric guitar/bass >sounds on the K2K, of for that matter any keyboard synth? I'm quite >disappointed with the guitar sounds on every synth I've listened to. > >Ian

Ian, You've struck one of my favorite chords! Yes, I have had great success with guitar of all types (lead, rhythm, chunks, clean stuff). My approach is somewhat different than most synthesized guitars I have heard.

First, I start with some good "clean" guitar samples. Nothing fancy, just good clean samples. My personal favorites are a clean Strat, a clean Gretsch Tennessean, a Chapman Stick sample (off of an O3R/W), and also a crunchy Strat (also off of an O3R/W) which has alot of "pick" character at the start of the sound. I usually layer two or three of the different samples adjusting the relative layers and filter settings to emulate different pickup combinations. Usually I use two of the clean guitars described above as the main instrument, and add the crunchy strat to add bight to the start of the sound. This is not necessary for rhythm sounds, but adds alot of character to lead sounds.

The only thing special about the individual patches is that I set the release to go full scale and release to zero in 0.06 seconds, and I then set the filter cutoff envelope to go to full scale immediately on release. This little touch emulates the sound which is made when either the guitarist takes his fingers off of the strings at the end of a chord, or mutes the chord with his right hand.

If I want to add feedback, I use another voice to fade in a sine wave at some pleasing harmonic.

The final touch is to then pipe all of the layers to the "B" outputs to an outboard processor. The one I use (with great success) is an Alesis Quadraverb GT. I think you could use any of the popular guitar amp simulators with good success. With the GT, I can then get almost any sound that a guitarist might like from the grungiest Paul/Marshall to the brightest Strat/TwinReverb sound.

Another key is that you have to learn to play like a guitarist. When you play a chord, start with your left hand and quickly role the chord up.

Don't use many notes. Rock guitarists mainly play just the root, fifth, and octave for the most part. And, they rarely if ever use a minor third with a heavily overdriven sound. When you are playing heavy screaming leads, don't worry if you make mistakes. Part of the character of electric guitar involves the slop which inevitably creeps in (i.e., a string inadvertantly left ringing while the player is "really" playing on another string, two strings struck with the "wrong" one immediately muted, etc...).

My personal experience is that the only way to get good electric guitar sounds from keyboards is using clean samples through external guitar type processors. I've never heard a sampled dirty guitar sound which sounded good to my ear. They are just too static.

An added bonus is that most guitar processors also have excellent distorted Leslie effects as well. My Alesis Quadraverb GT is a great example. It can get the anything from a cool Jimmy Smith sound through a cutting Keith Emerson to a grinding John Lord organ sound. Like the guitar samples above, I use fairly plain B3 samples, layered in different ways, and through the GT. I can even setup the Q/GT so that the pitch wheel or sustain pedal (or any other controller for that matter) controls the Leslie rotor speed.

As for the success of my guitar sounds, I have had several guitarists (including two of the best in the local scene) comment the guitar playing on my recordings, and asked who the guitarist was! I don't think you can eliminate the need for guitarists with these methods; the guitar is such a versatile instrument, with myriads of playing techniques which are difficult if not impossible to emulate on a keyboard, and certainly not with a single patch or a small set of samples. But using these techniques, you can get more than adequate guitar sounds, and easily cover tons of the bread and butter guitar sounds needed for 90% of the music out there.

Hope this is useful to you.

Terry T.

---

Date: Fri, 5 Apr 1996

From: RABITZ Josef

Subject: Guitar samples

Terry,

your discussion about achieving a realistic guitar sound is great. just some additions from my experience :

>If I want to add feedback, I use another voice to fade in a sine >wave at some pleasing harmonic.

1) don't use the internally generated sine - it has harmonics and thus does not melt like it should. instead use the ROM Sine sample, it's ok.

2) the sine sample has a different phase like the guitar sample and moreover it might be slightly out of tune. I had a better result using a resonant filter placed at a typical fixed frequency simulating the feedback through an acoustic amp. Thus the feedback is different for each note like it is in reality. some tones do better, others won't feedback.

3) a tip for the solo guitar vibrato : never use a LFO, it's too regular, you will not find such an accurate wave in nature. use the pitch wheel instead. use the pitch wheel only between zero and higher pitches. most of the guitar players tie the strings, that raises the pitch only.

so long - josef !

Hidden Effects Part 76

Date: Wed, 3 Apr 1996

From: Wim Verheyen

Subject: hidden effects

There are 47 effect presets in the K2000 (1-37 and 100-109), however the Digitech effects processor seems to have 128 presets. To reach these hidden presets, type in ID# 200-809 in the effects page, you will get a message "Not Found", but the effect is there. (but you cannot edit it)

Then David Fox explained :

As stated by someone else on the list, the Digitech 256 comes with 127 preset effects. Since the Digitech has 127 slots, there are 127 possible effects that you can call up on the K2000. To spread those numbers out over the 10 banks, they are accessible using the following numbers.:

1-37 100-109 200-209 etc up to 900-909

We overwrite the first 47 presets with our own effects settings (0-37 and 100-109). We create our own objects in the Setup ROM which cover up the ones Digitech created.

But the other 80 effects (#s 48-127) are still accessible by typing in the appropriate K2000 number (a number between 200 and 909). The K2000 wil say Not Found because there is no Kurzweil objct in ROM at that number. But it still sends a program change to the Digitech, calling up tha number. You must type the number in (ass opposed to scrolling with the wheel) because when scrolling, the operating system is designed to skip over "Not Found" object numbers. You could also send a program change (with FX Mode set to Master and sending the command on the FX Chan) to call up the effect.

Although I don't have a list of what those effects are, they are the stock presets that come with the Digitech 256, so you could find out by asking Digitech.

Keep in mind there is nothing "special" about these effects - they are simply the presets that Digitech chose.

This is the list of hidden effects Matthew Bassett listed with the appropriate K2000 locations.

K2K#	EFFECT NAME	CONFIGURATION
200	Medium Chorus	Stereo Chorus
201	Slow & Sweet	Stereo Chorus
202	Slim Machine	Para+Cho+Dly+Mix
203	I Love Leslie	EQ+Chor+4Tap+Mix
204	Rapid Sweep	Stereo Flange
205	Animal Flange	Para+Fla+Dly+Mix
206	Tubular Flange	Stereo Flange
207	Animal Flange 2	Stereo Flange
208	Chorus Room	Chorus+Room+Mix
209	Chorus Delay	Cho+Dly+Room+Mix
300	Slap Chorus	Stereo Chorus
301	Delayed Flange	EQ+Flan+4Tap+Mix
302	Swimming Delay	Para+Fla+Dly+Mix
303	Fat Chr Thin Dly	Para+Cho+Dly+Mix
304	Rotary Organ	Stereo Chorus
305	Flange Solo	Stereo Flange
306	Flange Pan	EQ+Flan+4Tap+Mix
307	Soap Opera	Fla+Dly+Hall+Mix
308	High Synth Lead	Fla+Dly+Hall+Mix
309	Crisp LeadSynth	Para+Cho+Dly+Mix
400	Piano Chorus	Chorus+Room+Mix
401	LA Chimes	Delay+Hall+Mixer
402	Fat Synth Bass	Fla+Dly+Room+Mix
403	Keyoard Cho/Vrb	Chorus+Hall+Mix
404	AdagioForStrings	Fla+Dly+Hall+Mix
405	Sharpened Edges	EQ+Flan+4Tap+Mix
406	Guitar Solo 1	Cho+Dly+Hall+Mix
407	Full Bass	Cho+Dly+Room+Mix
408	Guitar Delay	Cho+Dly+Hall+Mix
409	Lead Guitar 1	EQ+Chor+4Tap+Mix
500	Lead Guitar 2	Cho+Dly+Hall+Mix
501	Row Yer Rowds	Cho+Dly+Hall+Mix
502	Stereo Image 1	Stereo Chorus
503	Metal Guitar EQ	Para+Chorus+Mix
504	Monster Gate	EQ+Gated+Mixer
505	Ambient Snare	Ultimate Reverb
506	Fat Snare	Gated Reverb
507	Big Snare Chambr	Ultimate Reverb
508	Gates Of Hell	EQ+Gated+Mixer
509	Bright Drum Room	Ultimate Reverb
600	Darker Drum Room	Ultimate Reverb
601	Kick Chamber	Ultimate Reverb
602	Big Kick Room	Ultimate Reverb
603	DelayVerb	Delay+Hall+Mixer
604	SplashVerb 1	Delay+Hall+Mixer
605	DelayVerb 2	Cho+Dly+Room+Mix
606	Echo The Left	Cho+Dly+Room+Mix
607	To The Right	EQ+Reverse+Mixer
608	600ms Reverse	Reverse Reverb
609	400ms Reverse	Reverse Reverb
700	200ms Fast Gate	Gated Reverb
701	Absolute Gate	Gated Reverb
702	350ms Gate	Gated Reverb
703	Chorus Rev Right	Chorus+Hall+Mix
704	Delay Chase Room	Delay+Hall+Mixer
705	Chorused Bedroom	Chorus+Room+Mix
706	Low Boost Graph	Graphic EQ
707	Mid Boost Graph	Graphic EQ
708	High Boost Graph	Graphic EQ
709	Parametric #1	Parametric EQ
800	Parametric #2	Parametric EQ
801	Parametric #3	Parametric EQ
802	Sweeping Hall	Fla+Dly+Hall+Mix
803	VocalVerb	Cho+Dly+Hall+Mix
804	Dark Reverse	EQ+Reverse+Mixer
805	Slow Combed Room	Fla+Dly+Room+Mix
806	Space Flanger	EQ+Flan+4Tap+Mix
807	Super Stereo	Para+Cho+Dly+Mix
808	Crystal Hall	Room Simulator
809	Big Chorus Room	Cho+Dly+Hall+Mix
900	On The Town	Fla+Dly+Room+Mix
901	Dark Bird	Chorus+Room+Mix
902	Light Flange Rm	Fla+Dly+Room+Mix
903	GeneralVerb	Cho+Dly+Hall+Mix
904	Funky Bow	EQ+Reverse+Mixer
905	Vox Whispers	Delay+Room+Mixer
906	Slightly Combed	Fla+Dly+Hall+Mix
907	Into The Abyss	Cho+Dly+Hall+Mix
908	Pink Melon Cafe	EQ+Chor+4Tap+Mix
909	Chorus in a Hall	Chorus+Hall+Mix

K2500 Answers

Date: Mon, 1 Apr 1996

From: David Fox

Subject: K2500 answers

The following replies are from one of the engineers at R&D:

>>I've figured out how to setup the K2k for multi-timbral use. The question is: Is there any way to save this multi-timbral/multi-MIDI channel setup so that I can just load it in (rather than reprogramming it every time I want to come back to it).<<

Someone else suggested saving master tables, another approach would be to put the sequencer into "multi" when song #1 is selected, then hit: record, play, stop, save. The resulting song will contain 16 program changes (as well as volume and pan settings) which will set the K2000 up when selected. This sequence can be edited if desired, giving a lot of flexibility.

>>The big difference I find is when the rez is cranked up, the K2000 (Emax and lotsa others that try too) gets really crusty i.e. BZZZZT instead of Bzsszsszzss.<<

If you just try to add a lot of resonance at a freq where there is significant energy, you will get digital clipping which may or may not be what you want.

What most people don't realize is that you can set the "Pad" parameter in the first DSP block to create headroom for the cranked rez. 18 dB of pad will usually do it, then go to the output page and increase the "Gain" parameter to compensate. I'm not saying it will sound like an analog synth, but it will get rid of any "crusty" tendancy. This is true for the EQ DSP's too; if you try to boost without padding, it will clip, although it may just sound "mushy".

>>Anyway, when the K2500 is in setup mode, does it still respond to notes sent in via MIDI? If so, is it still 16 part multitimbral even though it only has 8 zones while in setup mode?<<

Someone else already answered this, but yes. I just wanted to add that if a note is received on the MIDI channel that matches the "Local Keyboard Channel" (if one is designated), it will play all the zones in the setup. Any other channel will play normally.

>>Unfortunately, when you leave setup mode, the original programs assigned to each MIDI channel will be reinstated which can cause confusion if you've just recorded a part on your sequencer with the setup and all the sounds change!<<

Partly true, but the only channel that gets "reinstated" is the one that was current in program mode when you went to setup mode. The other channels are left assigned to whatever programs were assigned by the setup.

>>I'm still gonna have to a) audio record a beat or part of the song b) lay down a beat in Cakewalk (ugh) and use the JV to play it while I sample the vocals.<<

A method I've used is to create a reference beat by looping a sample of a finger snap (or whatever). The length of the loop determines the tempo. I play the looped sample while the sequence is running, and if needed, tweak the pitch of the sample til the drift is acceptable. Then I play the sample by itself, hit record, the loop keeps playing, providing a stable reference for whatever I'm recording. Of course a huge drawback is that you can't hear your song while you're singing. To get around that, I strum guitar over the click to create a scratch guide. Then I play that sample while I sing. OK, it's not pretty, but it doesn't require buying more stuff.

K2500 SAMPLING OPTION AND INSTALLATION

Date: Wed, 3 Apr 1996

From: Blake Stone

Subject: K2500X Sampling Option Arrives!

Hurrah! My SMP-K2 sampling option arrived yesterday, and was rushed home to prepare for the ensuing operation on my beloved. I thought I'd post a brief synopsis of my experience for those interested.

INSTALLATION

The upgrade ships with a fairly detailed description of the upgrade process, but be warned: this is not designed as a user-installable upgrade. Unless you are feeling extremely confident, I would recommend letting a certified service center do the work for you. Of course, free to ignore my own advice, I proceeded to do the work myself.

Step One: Check the Engine Board Revision

Early in the process I was relieved to discover that my K2500X is a late enough revision that I did not need to dig out a soldering iron. Some early boards require the removal of a resistor. If this had been the case I would definitely have closed her up and visited my nearest service center.

Step Two: The Digital I/O Board

There are actually two boards that form the sampling upgrade. The digital I/O board is a smallish beast containing minimal support for the optical digital out and XLR digital input and output. It screws right to the rear panel of the K2500 so that plugging cables in results in zero stress on the board itself. Nice design. A pair of cables connect the digital I/O board to the main sampling board later in the process.

Step Three: The Main Sampling Board

The main sampling board contains all of the A/D hardware and logic, and includes four external ports. Included are left and right balanced 1/4" low impedance inputs, a stereo unbalanced 1/4" high impedance input and an optical digital input.

As previously mentioned, two cables tether the sampling board to the digital I/O board. Another two connect the unit to existing K2500 hardware.

Installation would be pretty straightforward but for two problems:

- The ribbon cable connecting the sampling option to the engine board was already connected ... using the wrong end of the cable.

- It's pretty cramped inside the keyboard where the sampling board resides.

The former problem was easy, I just reversed the cable. The latter problem took some effort.

Careful arrangement of cables, both old and new, is the key to a clean installation. I needed to remove one cable tether (okay, a piece of tape) that had been installed at the factory, and reroute a number of cables several times until I was satisfied with the arrangement. Take the time to do it right!

I seated and removed the board a half dozen times before I was satisfied, so make sure your service center takes the time to do it right.

Step Four: Finishing the Job

Once everything was in place, I snapped everything back together, turned her on and tried the sampling option diagnostics. The loop-back tests required cables I didn't have, so I skipped those and moved on.

The sampling upgrade shipped with v2.13 of the K2500 operating system, so I could have upgraded if I hadn't already done so. Since it was included with the kit I assume it fixes some problems relating specifically to the sampling option.

SAMPLING IN ACTION

Once installed, the sampling option is a breeze to use. Its very much like the K2000 option, but cleaner.

Cleaner in what way? Specifically, sampling isn't an operation that locks up the keyboard. You can set up your sampling options (including triggering threshold, monitoring, gain, sampling time, inputs, et al) and then do ANYTHING during the sampling process. Dial up a program or two. Play them while making editing changes. Play a song from the sequencer. If you're doing self-sampling the sampler catches it all.

When you're done, return to the sample page and press stop. No messy reseting the keyboard to prematurely terminate a sampling session! Hurrah!

Selecting the various inputs was a trivial operation, but I haven't yet tested them all. I'm assuming it's as easy as it looks.

CONCLUSION

It's shipping. It works. The sample-while-playing feature is a gift from above.

All in all, the sampling option is a welcome addition to my favorite beast.

Blake Stone

LEGATO MONO

Date: Mon, 22 Jul 1996 From: Darwin Grosse Subject: Legato Mono Mode

>Hi my name is lester pickens i need to if there is any trick i can use >to make smooth transitions between notes using the multi sample >keymaps in rom like keymap 7 Jazz Guitar or my own they click between >notes. > > Everybody says the clicks i hear is the switching between >the samples but in the K2000 v3 guide it says: > >By using GHiKey or GLowKey as the only source for pitch tracking you >can creat mono like layers in a polyphonic program I don't know how to >do this can you tell me how to get this effect or any other trick >using Funs or anything. Some Midi controllers insert pitch bend to get >from note to note with no clicks is there some way to do this in real >time > >Any trick or hack will do Ill try anything > >I play a midi guitar controller and need smooth >6 channel mono mode Please Help >

OK. First, thanks for giving me a swing at this. Surprisingly, it's an issue I've run into while working on some sample CD stuff for someone. Basically, the GHiKey/GLowKey thing is specifically for a multi-layer patch, where you want a single "mono" voice -- like a chordal synth pad, with a lead synth sound on the highest (or lowest) voice. Not so helpful under these circumstances.

The trick/hack involved has to do with subsuming our (as synthesists) desire to have a perfect set of samples for dead-on sounds. There are various ways around this problem if you are willing to make some compromises in sound...

At this point, I'm going to spit out these ideas, and let you (or someone else) try them out -- my schedule is so locked down that I'm unable to spend sufficient time defending these approaches. However, there is'nt any reason why they should not work... they just might not!

Idea #1: The minimal multisample approach:

This first idea (one that I've actually used) is to recognize the problem -- the K2000 will not portamento smoothly through multiple samples. Instead, you get an annoying click as the portament slides through each section from the last note to the new note. A drag. However, with a little experimentation, you will find that this does not occur when you do a pitch wheel move. Herein lies the key to some solutions...

This first approach involves minimizing the number of samples you have in any single multi-sample. Obviously, if your sound is based on a single sample, you don't have any problems -- since there are no sample "borders" to cross, there will be no clicking. The downside to this is the dual troubles of 1) a sound doesn't sound "natural" when transposed to an extreme, and 2) depending on the sample rate, the sound generally cannot be transposed greater than an octave or so.

In order to use this approach, take a sample set like the Jazz Guitar, and monkey around with the multi-sample break points. Currently, the samples are about 4-5 semitones apart. If you are willing to take a bit of a sonic hit, extend each multi-sample to cover a larger distance -- say 1 octave. The change in tonal characteristics will be significant, but the problems with clicking will be greatly reduced. If you are using a sound for a specific tune (or whatever), you can often make a "song-specific" key mapping that will get you through the fire.

Idea #2: The many-patch approach:

In your case -- a MIDI guitar controller, you can take advantage of the "multi" mode to do some of this. Make a patch for each string -- using a single keymap for each. Since most guitars are struggling to provide a two octave range for each string, you could easily stretch a single sample to work for each string. For ease of switching, you could map the six patches to a single "setup", and have fun...

The downside? No portament when switching strings. A big drag, but at least controllable from your playing...

Idea #3: The false-portamento approach:

I've had to use this for sample CD work, and have found it to be sort of useful.

Basically, the idea is to simulate a portament with a pitch "slide". Although it is not useful in the most extreme cases, it can provide some of the sonic cues of portamento without the cross-sample clicking.

Basically, set up an envelope (or an AR, if you like) that provides a stable source of modulation. Then apply it to the pitch with SRC2, using some other modulator (like a pedal) to turn it on and off. This generally provides a useful "slide" function that is controllable at play-time.

Conceivably, you could create some FUN's that would adjust the amount of envelope modulation based on the "last key" and "current key" functions, but I've not tried this yet...

If this idea sounded like Greek, take a look at my tutorials for information on pitch modulation.

Idea #4: The alternate patch approach:

This is probably the most player-friendly approach -- simply avoid sample use when you will be using monophonic portamento. Going back to the olden days, an awful lot of synthesis can be done with simple sine, square and saw waveforms.

Since several of the algorithms allow you to use these as synthesis tones, and are not subject to "sample breaks", they are generally the best bet for monophonic synth tones that will be using portamento.

An note: I started synthesis before the sampler-wars began, and remember the days of getting usable organ, guitar and piano sounds out of a DW-8000, Juno-60 or Arp Axxe. Again, they aren't as plush as the Kurzweil multi-samples, but will often do in a pinch -- especially in live scenarios.

Just spend a bit of time with your synth programming to get the tone you like.

Going back to your original question -- there is no real way to adjust the K2k to slide multi-samples correctly. If you think about it, you can see why -- the "state" that a sample is in is not easily translated to a new sample unless all samples are fired simultaneously. This would eat polyphony when in mono mode, and would render the K2000 a straight, non-multimbral synth whenever any patch was in mono mode. The alternative would be to provide some sort of a cross-fade function, which is not currently available with the K2x series of synths, and would probably be a disaster in any case.

Others may know -- how is this dealt with on other machines? On the ones I have, there is simple no portamento functionality...

[ddg] Darwin Grosse

LESLIE

Date: Thu, 13 Jun 1996

From: Lynn Cooley

Subject: Leslie fx (toggled)

> Is there anyway I can set up footswitch 2 press on/press off >switch? I've messed withe Gospel Organ patch so that I can turn on >the leslie fx by stepping on FS2 but as soon as I take my foot off >it, it stops the leslie

Seeing as i'd been wondering about that myself, and i didn't see any replies, i tried it. The hardest part was to dredge up memories of making a latch out of NOR gates, but after that it was easy sledding.

You need a function that stores an input until the next clock.

That's supplied by Kurzweil with the "Sample b on a" FUN. Then you need to make that function see its own output, negated. I used the "|a-b|" FUN to do that, setting "a" to 1 and "b" to the first FUN's output. The negated output is then fed back to the first FUN's "b" input, so that it samples the opposite of its own value every time the trigger input comes along.

There's an example patch in ftp.sweetwater.com/incoming, called ORGFTSW.KRZ, described in ORGFTSW.TXT. I used Foot Switch 1 because it was all i had connected at the time, but it's easy to change on the FUN page.

Remember to choose appropriate values for the Sustain or Sostenuto functions on the LAYER page.

The author wishes to acknowledge the significant contribution of the Kurzweil engineers who supplied us with this amazing instrument in the first place.

Lynn Cooley

SET PHASERS TO FUN

Date: Mon, 11 Mar 1996

From: nursingh@CLYDE.ENS.INSA-RENNES.FR

Subject: K2000 phaser effect?

>I dont think so, but if you use the NOTCH filter with a narrow width >and an LFO on the FREQ it gets pretty close !!

or else two layers with slightly different pitches or one layer with 2 or more DSP oscillators with slightly diff. pitches.

or else use the osc sync. algs. (say 30) and keep the master osc constant and vary the slave osc's pitch with a slow trianlge LFO.

Yash.

SNARES

Date: Thu, 7 Mar 1996

From: Joe Albano

Subject: Snares

Mark Schaffel writes :

> hey joe, read your post on the kurz users group.. > > regarding snares you have in your keymap...do you have differing >types of snare hits (w/ slightly different nuances) spread over >different keys or just the same snare in order to facilitate easier >playing?

Hi Mark -

Yes, I do have multiple samples w/slightly different nuances.. My snares have up to 9 different samples (all fff) - 2 (similar) center hits, 6 rimshots (since that's where I usually hear the most tonal variation when drummers play), and 1 (soft) sidestck.

The two center hits (middle of the drum/no rim) and 3 pairs of 2 rimshots are combined into 1 sample each (rimshots 5 and 6 are slightly stronger accents):

c.snare 1 followed by 2 = 1 sample (4 keys)
rimshot 1 followed by 2 = 1 sample (2 keys, above and below the c.snares)
rimshot 3 followed by 4 = 1 sample (2 keys, above and below the c.snares)
rimshot 5 followed by 6 = 1 sample (2 keys, above and below the c.snares)

This is all 1 layer - it's set up so each pair randomly alternates (using AltSwitch = RandV1), and when played from a kit of pads with a separate pad (c.snare) and rim, I can get up to 2 variations for the c.snare pad trigger and 4 (1-4) for the rim trigger (though this requires 2 extra layers for the rims).

A little velocity-control of attack transient level (not time) to simulate the tonal range from hard to soft strikes, and a little random pitch variation as well, and the result is quite natural sounding..

Needless to say, to find this many varied samples from the same drum, I had to do my own (though I think the Clearmountain, Erskine, and Gadd CDs may provide enough variants to accomplish something of this sort).

Well, I hope this was of some interest..

Cheers - Joe Albano

Back to Dead Letters Office

Main Page :

Documentation :

FAQ :

DSP Blocks :

KDFX :

Algorithms