how to create a keygroup program?

hi,
could anybody explain to me step by step how to create a keygroup program on the 5k with two or more sampled notes, please?
can i sample two notes on the same octave of an instrument like c and d and just assign it in the keygroup program?
does the mpc automatically detect which note was sampled or do i have to tell the mpc: this sample is a "c" on the third octave for example?
i think the additional manual of this feature is a littlelbit confusing.

i would love to create some more kg programs from instruments of fl studio and so on to play them directly from the 5k.

thanks,

dirk

Heres how to start:

________________

To create a Keygroup program, do the following:
01. In PROGRAM Mode, select the 'PGM' field.
02. Press the [WINDOW] key.
The NEW PROGRAM Window will be displayed.
03. In the 'Program Type' field, select KEYGROUP
You can also name the program using the 'New Name' field.
04. Press [F5] (DO IT) to create a new Keygroup.
________________
Further details on setting up keygroups can be found in the manual starting on page 130.

Mojamz

@mojamz

thanks for this but thats not my problem.
creating a new and plain keygroup program is not the problem.
the question is what for notes of the instrument you want to sample should be sampled and how much notes for lets say 6 octaves and whats the best way to split the keygroups regions on that 6 octaves.

for example: you could sample 6 times the note c. one time per octave.
but is it really nessesary to sample so much notes for a kg-pgm?

My baad, sorry I misunderstood your question. I have a little experience with sampling techniques/programs
(Akai S900, etc.) but I am new here and also very new to the MPC500.

However, and although I am sure there are other more knowledgeable folks here, here is my take on what I now think you are asking:

I don’t think there is a set formula for determining what notes are to be sampled or how many keygroups should be set up. What is most important however is how realistic and close you want your finished product to sound to the instrument you are sampling. The more complex the sound of the instrument, the more samples you would have to take to make it sound realistic.

For example, a note played on a acoustic piano sounds different depending on how hard/soft it is played. When played soft you hear less attack and when played loud you hear more attack thus more overtones/harmonics. So, if you were sampling a piano sound, you would not only sample several notes, you would also sample them at different volume levels. For a sound that is not as complex you would not have to take as many samples to get a realistic sound.

You also have to consider how stretching the pitch will affect the sound. For example you might get away with sampling all “C’s” but then you might find, again based on the source you are sampling, that you have to sample all C’s and G’s to get a good sound.

It might help to study some typical examples of keygroup programs. If you don't have any available, you could use as an example, the Akai’s “GrandPiano1-144m” which consists of 46 keygroups each having 3 samples velocity cross-faded. In total there are 138 samples making up this 88 key instrument.

As you probably know, on the MPC5000, there can be up to 128 keygroups, each with up to 4 samples for a total of 521 samples. To create really good keygroup programs is tedious and time consuming and you just have to experiment to get good. Lastly, many people don’t have the time or the patience to do good sampling that’s why businesses like Loopmasters and others exist.

To get a decent multi-sample keymap, it will depend on what you are trying to sample.... if you want a perfect piano program for instance, you need to sample every note at multiple velocities. If you are sampling something like Mellotron strings, then you could get away with 2 or 3 samples per octave, perhaps sample each C and F only and stretch the C up to D# and down to A, and map each F sample from G# to E.

FWIW, unless you are trying to make PGM's to sell you make a sample for the purpose at hand.
You don't need to do unnecessary keyspans unless it's truly required for HOW you plan to use the sample.
Only be concerned with created that part of the sample that you need.
Here's an example.
How many people do you know that even if they sampled every key and had the ability to give it 20 layers of velocity would sounds as good as the average rompler?
Or even worse, sound no better than if they just did 2 or 3 octaves with 3 layers?
I make lots of multi's but they are all to go with or enhance an existing patch some where else.
I even sometimes don't loop it I will just use a longer sample that I created a certain sustain on.
Sometimes by making a sample stretch C2 to G2 or even from C2 to C3 you will get a sound more like what you are after or an unexpected result that just sounds right.
If you do want to make a full Keygroup with a large span and layers the easiest way to do it it to take a S1k PGM and plug your samples into it.
You can quickly make a list of samples from a keyboard and just plug them in to each layer.
I quess my point might be getting lost so I'll go back to it, create your samples for your ear and purpose. Please don't get caught up in believing there is a certain way to achieve results, it just doesn't ( or shouldn't) work that way.

Creating sample instruments based on keygroups is fascinating but time consuming I confess. My motivation is to have a basic set of Wurlitzer, Rhodes and Pianos in my MPC 5000, thus not having to carry the 18kg of my Nord Stage 88 all the time...

First, I created a sequence which plays all the notes I want to sample at a certain velocity (to give it a start, I took three notes per octave: C-E-G#, over 6 octaves). The sequence makes sampling reproducable and eases the workflow. First sequence was at velocity 127, then I copied the sequence 3 times and changed the velocity in the STEP EDIT mode to 95, 63 and 31.

Then I created a SONG which calls all four sequences in a row. This gives me the possibility, to record the stage piano into 2 channels of the hard disc recording section of the MPC. So, after creating the SONG switch to HD RECORD and set the channels appropriate. Then start playing the song and record all into 2 tracks (for basic stereo material).

A small comment on the sample lengths: Usually, a grand piano will need up to 12 sec samples in the lower section (C1-C3) and 2 sec in the upper section (C5-C7). Rhodes and Wurlitzer are much different, you can get along with much shorter samples.

The recorded tracks can then easily be exported as WAV files. In the above mentioned example with 3 notes per 6 octaves (C1-C7) at 4 velocities this means 77 samples to be created - this is something to be done at the computer. I used the freeware program Audacity, combined the 2 tracks to one stereo track and cut out every note played/recorded (with the comand "export selection"). Each sample is named accordingly, e.g. "C1 127 stereo", stored in a folder e.g. "/Yamaha C7/samples/stereo". Afterwards, with batch automation, I converted all stereo files to mono and stored them in "/Yamaha C7/samples/mono".

Cutting the recorded track into single samples is the most time consuming part. 77 samples take about 10 hours of concentrated work. But this is the most viral part of creating the keygroup instrument

In the above mentioned example, this made 48MB stereo samples and 23,5MB mono samples. This gives you the maximum flexibility for using your instrument. In a live environment, playing the piano in a rock band, mono samples are sufficient enough. It might be more important to have more velocity layers on board or to be able to modify the sound from soft to hard (soft sound for your piano intro, hard sound for getting through the mix in the middle section, e.g.).

You may also improve the stereo experience for your mono samples, if you use the keygroup panning. This allows you to define the panning of each single keygroup, thus spreading the mono samples in the stereo panorama. Further improvements can be achieved through using the crossfade in the middle section (C3-C5)

After having created all samples on the computer I transfer them to the MPC, rather via USB or Compact Flash. In case you want to use several instruments at once, you will need to rename the samples in order to give them unique instrument identifiers (otherwise MPC will overwrite them while loading into memory). Use a automation tool for renaming files like PS FileRenamer 2.97 from Polarsoft.

I think, creating the keygroup in the PROGRAM section of the MPC is common knowledge. Rather interesting is the question, how many layers do I need? How can I further reduce memory usage of my instrument? I will come to this later on.