Exchange tips and tricks for the Akai MPC4000
By mxbdev Sun Mar 30, 2014 11:32 am
dustymaestro wrote:Mxbdev, im going to gather some info from some old posts. Ill put in a thread as we can figure it out. I have read some really informative posts from Feline, Illiac, Blue Haze, MCSmooth, and yourself. Everybody is talking about gain structure, but has there own theory.

Could be useful, I've just approached it scientifically as best I could. I hevn't read everything out there about the topic.

dustymaestro wrote:The problem you describe is not getting a full dynamic response. I have read also that the gain structure effects how far you can turn a level up like the depth in mod matrix. For example you wont get a change in depth after +83.

If you wanted full dynamic response without going over 0 db, wouldnt you need tilt to be 0db at +100 in the depth field?

The issue is that the default problem sets the pads to be too sensitive. +34 tilt-velocity is a huge dynamic range. 42dB is the same as taking a sound, and doubling the volume, doubling again, a third time, fourth, fifth, sixth and finally seventh. This just isn't playable on the pads.

A tilt-velocity of +100 would be even worse, as you'd have a dynamic range of +/- 63dB! This means that if you had two hits one after the other with a tiny change in velocity of 3, one would be twice as loud!

The gain structure in my earlier post appears to hold, as does the hard cap at -3dB.

The places where you can change volume are:
  • Program level
  • Zone level
  • Keygroup level
  • Filter attenuation
  • Multi level for the program

By default, all of these are set to 0dB, so they have no affect, except program level, which is set to -6dB. This is where my second graph comes into play.

If tilt-velocity was set to 0, then the sound will play at -9dB (total of -6dB for all the above, plus -3dB hard limit). If the relative amplitude increases above +6dB then it will be capped. The slope of the tilt-velocity=+34 is such that this limit is hit at a velocity of 83 as you mentioned. However, all third party sample libraries I have seen set this to much lower value as a +/-21dB relative amplitude is not playable.

Halving the tilt-velocity value to +17 gives a maximum boost of +10dB, and effectively limits at a velocity of 100.

It is my current understanding that tilt-velocity always is based around a value of 0dB gain at velocity 63, and the gain can be calculated as follows.

Code: Select allGain= (tilt-velocity / 100) * (absolute velocity - 63)

Some examples:

  • tilt-velocity = 34
  • abs. velocity = 63
  • therefore: gain = (34/100) * (63-63) = 0dB

  • tilt-velocity = 34
  • abs. velocity = 127
  • therefore: gain = (34/100) * (127-63) = 21.8dB

  • tilt-velocity = 34
  • abs. velocity = 10
  • therefore: gain = (34/100) * (10-63) = -18dB

Therefore I have decided for all of my programs to start by picking a reasonable tilt-velocity (usually ~17), use the above formula to find the maximum gain from velocity (~11dB for tv=17) and shifting the program level to this amount (-11dB). The end result is that a hit of 127 should go though the program and out to the multi at a level of 0dB.

I really hope this is clear, like I said, it's easier to explain with a graph! :)
By dustymaestro Sun Mar 30, 2014 6:51 pm
For drums I turn off the routing in the mod matrix. I use it to taste for everything else.

To understand the gain structure you have to start adding or subtracting total db.
It's at -3 db(default) right from the start. You can add or subtract db at various places already mentioned. Program db + multi db + tilt velocity + global db + -3db = total db
Thats not the whole picture but hopefully you get the part that the sum of all those levels effect total db. Simple.

He's saying keep the default values. Except the tilt velocity to amp in the mod matrix. A good place to start if you use the mod matrix is at +17 instead of the default +34.

Im starting to think there is no setting where you can just leave it there. Its easier for me to view this as a mixing board. Multiple channels going to a buss than a master. You just have to mix as you go. In this case turning down levels as you go. You just have to know the limitations and structure.
Last edited by dustymaestro on Mon Mar 31, 2014 2:02 am, edited 1 time in total.
By mxbdev Sun Mar 30, 2014 7:28 pm
dustymaestro wrote:For drums I turn off the routing in the mod matrix. I use it to taste for everything else.

Really? You turn off all volume/pitch/etc modulations? Surely all your drum hits sound exactly the same each time if there is zero modulation.

dustymaestro wrote:He's saying keep the default values. Except the tilt velocity to amp in the mod matrix. A good place to start if you use the mod matrix is at +17 instead of the default +34.

Kind of ;)

Set program level to -11dB and tilt-velocity to amplitude to +17 in the modulation matrix. This gives full dynamic range without a huge volume difference between multi and program/sample modes. Obviously adjust to taste and note that if you have multiple samples one the same pad/keygroup drop the levels on these according (two identical playing samples on zones 1 and 2 each at -6dB is the same volume at one sample).
By dustymaestro Sun Mar 30, 2014 8:41 pm
I dont like the modulation matrix on drums. I like it to sound like the classic mpc dry sound. Ill start from there. Off the top of my head I use velocity, pitch, filter, envelope, make copies, and use different quantize values to make the drums go. I also use the q link sequencer.

Heres some more info on some tests on the gain structure done by a person named Muser.

AKAI Z8 Level Settings.

for this example I used a Normalized Tampura sample. any differences in using a full scale white noise sample may show some different values.
but this seems an adequate enough test for now. This test is carried out at 16Bit - 44.1kHz.

I'm not yet satisfied that I have a proper grasp of exactly how the structure of the Z8 engine is set up and what the best ways to use the modulation section is to drive the Amplitudes and Levels. I am not quite sure if the Amplitude destination is the overall programs Amplitude OR a given KeyGroups Amplitude. I suspect the later. if it is the KeyGroups Amplitude then it is the KeyGroup Amplitude which would need to be lowered to (say) -60dB before (velocity >> Amplitude) is used in the modulation section, for example. Also this approach would effect ALL zones in THAT KeyGroup. it may be better to have the Amp Envelope or the Velocity or Velocity Tilts set up to control individual Zone Levels instead. especially if you intend to get creative.

Given the sensitivity of the Gain staging's ability in this Engine to either induce cramped dynamics / level distortion & Filter distortion & envelope release clicking artifacts (if not very careful) the Z8 Engine is at one time very tricky and sonically very powerful, because of these structural design features.

Anyhow.. here is a little of what I found in my first stage.

With the Z8 Master Level setting set @ 0dB <--- This is on the Utilities GLOBAL page. press the Utilities button.
the Tampura sample was normalized.
in this case the maximum possible output level showing on the Z8's output meter <---- This is in the Record section.. it has a dedicated button.
(for a a single normalized sample) is -3dB.

This is with the program (Level) & Zone Level & KeyGroup Level & Filter Attenuation (All set at Zero).

IF I add a copy of this single Zone set that SECOND Zones level to -9.2dB you will hit Zero on the meter.

IF I then set both these Zones to -3.5dB & increase the programs level to +0.1dB then this also hits Zero on the meter.


For this Test I set the Z8 Master Level setting set @ -6dB <--- Utilities GLOBAL page. press the Utilities button.

IF I then set both these Zones to +6dB & increase the programs level to -2dB then this hits the meter @ -2.5dB.
NOTE: adding a third Zone copy allows for this -2.5dB reading on the Z8 output meter to then exceed Zero on the Z8's master Meter..

NOTE: working again on only Two Zones and not Three.
increasing the program Level any more has NO further effect.
-2.5dB can NOT be exceeded by increasing the programs volume while using these two zones.
ALSO when increasing the KeyGroup level, it also is not possible to exceeded this -2.5dB Level.

Interestingly, adding extreme Resonance on a 2-POLE LP Filter and altering the Filters frequency at this point shows that this limit then creates a possible distortion into the signal as the resonant gain in dB hits this headroom wall. This happens (without) overshooting -2dB in the Meters. However, IF the Filters attenuation is dropped to -18dB then, the distortion clears up (but) the meters then overshoot Zero. This indicates that the gain structure in some way limits overshoot in one condition and allows to occur in another condition. odd.

I am not sure if this Filter attenuation is the same as the KeyGroups Level or if it is a Filter attenuation.
The reason I don't know IS. you can have NO Filter selected and set the attenuation to -18dB and the KeyGroups overall level seems to drop.

This is as yet still unclear to me, in regards to the Gain and Level structure in this Engine configuration.
By Clint Thu Jun 05, 2014 5:39 pm
helloitsmeagain wrote:this thread is so awesomely in depth it's INSPIRING the shit out of me.

Corrected for you. :wink:
By manji Mon Jul 21, 2014 3:26 am
Not sure if this has been listed, and isn't specific to the 4k and can be done on any mpc that can have multiple programs.

Just to help with my basic songwriting and making sure my mpc is ready to go as fast as possible, I have a process that helps me find drums through a ton of sounds very quickly.

The first thing I do is load up huge folders of samples. Kicks, snares, hats, low bass hits, what ever it is you consider to be a basic beat construction kit. In the end, I have a couple hundred samples in memory.

I then make a ton of kits. Turn on a movie or something and make 20 drum kits. But, I make sure to keep all the basic components like soft snare, hard snare, soft kick, hard kick, booms, closed hats, open hats, clicks, etc etc on the same pads. This is very important.

How I personally file all of this away is I have a folder for Drum Kits, then genres like rap, dnb, acoustic, brush, percussive, etc etc. And then in each one, honestly, I just number them. Rap Kit 007, Rap Kit 008, etc etc. For me that really helps me remember stuff, weirdly enough.

Point being, be able to load up a huge collection of them as easy as possible. I have 20 kits in each sub folder. I can then load up 20 kits, go snag some coffee, come back, and have A LOT of kits at my disposal.

So why keep them on the same pads?

Because each time you make a sequence, you can now scroll through kits, as the beat plays, and hear how each kit would play whatever you're working with. It's a way to just try kits out quickly.


If you seperate out your tracks per pad, i.d. have a track for pad 1 (soft snare), a track for pad 2 (hard snare), a track for closed hat (pad 5), etc etc, you can now customize. This gives you A TON of options extremely quickly and can help you figure out what you want to use. On the 4k, because the memory is so large, I just keep all the kits I'm using loaded. However, if you'd like, youc an then create a custom kit from there in order to save on memory/disk space.

And also, since I'm sure the vast majority of us beat our drums in live, you can look how to seperate out pads to tracks. I work on a template, so Track one is ALWAYS High Hat 1, Track 7 is ALWAYS Kick 2, etc etc, which makes the kind of obtuse process on the 4k a lot more of just a muscle memory thing than anything.

Seq Edit>select pad on "Pad" field>copy to>make your selections from there. You can replace, make sure to grab the whole sequence, grab what you want, etc etc.
By manji Mon Jul 21, 2014 3:35 am
A way to get the most realism out of your sampler for acoustic instruments is to build at the very least programs with 4 levels of samples for velocity.

But since the 4k can have 2 programs playing on one track, you can build 8 level programs for extra, extra realism.

I do this with sounds I just adore in Extreme Sample Converter, which is extremely easy and contact me if you need help with. It only writes Keygroup programs, so it's perfect for anything outside of drums and one shots, which are extremely easy to build.

However, so that it doesn't completely hog your memory, I also build 1 and 4 level program as well. If I have a lot of sounds loaded up and I'm just messing with sounds, I'll use those. If I'm about to multi track out of my mpc into my recorder, I'll empty the memory and load up the really large program. If you have done everything write, you shouldn't have to adjust anything or resequence anything.

Combine with how extensive the programs in the mpc can be edited and it can really be surprising. Good for gigging and whenever you don't want to lug your computer around.

Extra realism out of your mpc.
User avatar
By MoreBuck$ Mon Nov 17, 2014 3:16 pm
2 programs playing on one track? :hmmm:
By manji Mon Apr 27, 2015 3:20 am
Ive been doing this for a while, and dont know if I ever mentioned it, but if you buy a 2.4 ghz WIRELESS (NOT bluetooth) keyboard, you basically get a wireless remote for the 4k. The transport functions dont work, but I control my rec/play from my kronos, its worth looking into if whatever youre using can too.

anyway, its really nice. A workflow I like on the kronos is to mute whatevers on ch 16, use it as a control channel, and with the wireless keyboard you can sit at your workstation, rec/play, do everything you do sitting at the machine, etc.
User avatar
By JUKE 179r Sun Jul 09, 2017 9:48 pm
I was needing a system check (Panel Test Mode) to test and verify that everything works correctly on a newly acquired MPC4000.
It was located on Page 9 the MPC4000 service manual.

Panel Test Mode
To enter the Panel Test Mode, press the GLOBAL key while holding down the SHIFT key in the GLOBAL page [F1] of the GLOBAL Mode.
• Moving the Q-link knobs changes the graphics accordingly.
• Tapping the PAD shows the amount of pressure applied on the bar graph.
• Switching the FOOT Pedal on/off changes the graphics accordingly.
• Turning the JOG wheel increases/decreases the value in the Data field.
• Pressing the key shows its name at the bottom of the screen. For those keys with the LED, the LED is turned
• Pressing the [F1] key stops the key test.
• After stopping the key test, pressing the mode key will abort the test mode.
• While the key test is stopped, pressing the SHIFT key and [F6] key shows the A/D converted value of the
Q-link/ PAD.