Hi Jared,
Don’t be hard on yourself, this is complicated by its nature. Sampled audio data (which can be huge in size) and the Voice data (which is small) that uses the sampled audio as its sound Source, are treated separately.

It does so because the Voice data is so small in comparison to the audio data. By ‘Voice data’ we are referring to all the parameters except the Oscillator wave (this includes all the envelope generators, the filters, the effects, etc). Example, the sampled data that is the Chick Corea Mark V Rhodes is over 400MB of data, the parameters that shape that audio into a playable Voice are about 2KB (tiny in comparison - and tiny is not strong enough an adjective. To put it in perspective, if you played a gig and the band got paid $2,000 that’s a good gig... but not in comparison to the gig that pays $400,000,000 - you get the idea!) there can be a tremendous difference in length of time loading/saving these two very different groups of data.

The whole concept is to keep the $400 million in the bank (Flash) then, quick as a bunny you load in Voice data that has the “catalog” to access the data it needs. So you don’t have to wait for the 400MB to load every single time you have a sound you want to play.

So it is the huge sampled audio data that gets stored on the Flash Board.
The Voice parameters (the small data) “points” to a location on the Flash Board in order to retrieve the Oscillator sound Source.

The “catalog” that the XF constructs so that it “knows” where to find the audio is the Waveform List.
Waveforms are collections of samples.

Say you have a set of samples that are each note of an 88-Key piano... the the individual samples are organized into key and Velocity map that tells them when to play... this entity is called the Waveform... a Waveform collects the samples into a useable grouping.

So in the simple example the 88 individual samples are collected into a Waveform called “Piano”.

Ok... fortunately you do not have deal with individual samples... you can select Waveforms.
Many Voices can reference the same Waveform- the Waveform becomes just like selecting a Waveform Source in an analog synth. If you set your MiniMoog to a Sawtooth Wave... you have selected a sound Source to sound across the keyboard. In a sample-based instrument you select a Waveform made up of individual mapped samples.

The Flashboard Waveform Lists, which you can view by going to [UTILITY] > [F6] FLASH; You can have 2,048 Waveforms per Flash Board... which can be organizing over 8100 individual audio samples, each.

Now to your questions....
We are going to assume that when you say you have “32 samples” you actually are referring to 32 Waveforms on that Flash “Waveform List”
We are also going to assume that these are located on FL1

Since it appears you want to keep just a few (3 Waveforms) and discard 29 Waveforms... you might just want to clear all data, then load just the data you want. I am suggesting this because it it easier to load what you want than delete, delete, delete... what you don’t.

While you have the entire 40th Anniversary data loaded... go through it. Make note of the Voices you wish to assemble from the library. If your memory is good, fine. Don’t be afraid of writing them down, if you don’t think you’ll remember all that you wish to keep.

The XF is able to LOAD individual data from an ALL data file.
From any ALL data file you can select LOAD TYPE = “voice”
During this process it will load all the Waveforms (and the samples they organize) to the XF
The Audio data is installed (burned) into Flash, the Voice data is loaded to the USER Bank location you choose.
The XF creates a “catalog” (the Waveform List) that is responsible for managing the relationships been the User Voice and the data it needs to sound properly.

When you go to the 40th Anniversary Data... you can LOAD just the Voices you are interested in... the XF system will do the rest.
If you select a Voice that requires a Waveform, you simply need to direct the Voice to the User Bank location of your choice, the XF will gather the Waveform(s) necessary to make it sound properly and it will install (burn) that data to the lowest numbered empty space on the Flash Board (FL1 or FL2) you direct it to...
Because you will be loading data from a single file the XF manages what gets loaded. Once the XF writes a Waveform, if you select another Voice that uses that same Waveform, it simply updates the catalog. In other words, it will ‘know’ that it already has the Waveform loaded in memory- it will not duplicate the Waveform. It will ‘repoint’ the Voice to use the Waveform location where the data it needs will be found.

It is able to do this because the Voices are coming from the same file. It keeps track of what it has loaded already.
This way, as you go through the 40th Anniversary data, you can pick just the Voices you want - Load and install just their data.

The key thing that will make this all possible is the setup of the LOAD options.

Method
You will want to load USER 1 and 2 with the data you refer to as your “original voice/performances” - make sure they are in place.
Next you need to have the 40th Anniversary file available - make sure you have selected, by name, all the Voices you want to gather from that collection... you will be loading these as individuals.
Format the Flash Board
Load the individual sounds
Press [FILE]
Move the cursor into the DIR (directory) area and highlight you 40th Anniversary File (.X0A)
Using the right cursor arrow move to the TYPE box
TYPE = voice
Move the cursor down and set the target... where you want to load this Voice.
If something occupies the location you will see its name (so you know what you are overwriting)
You want to load to USER 3...so you might set this to “USER 3: 001”

Using the left cursor arrow twice, move the highlight back to the 40th Anniversary File
What the screen is telling you is...
You are viewing Directory entry: 40th Anniversary File
You are set to select a single “Voice” from that file and direct it to the internal User3 001 location.

Press [ENTER]
“Executing....” appears briefly as the XF reads the file and prepares to show you the list of all the Voices contained in that file.
A list of 128 sounds in USER 1 appears. The [USER 1] button lights
You want to view User 4... press [USER 4]
A list of the 128 sounds in User 4 appears
Using the Data Dial or down cursor highlight the voice you want to Load to User 3, 001.

As soon as you have highlighted a Voice, the LOAD Options for directing the Waveform to FLASH appears
You can select to point the data to FL1

Press [SF2] LOAD
The XF will place Voice data in USER 3, it will install the Waveform to your Flash Board

Rinse and repeat loading each Voice to a different User 3 location (002-128)
Again, if the XF determines that it has already loaded a Waveform, it will just load the Voice data - repointing the Voice to the Waveform location (0001~2048).

Once all your data is loaded and all Waveforms installed to Flash, SAVE an ALL data file “with Samples”.

Hi Jared,

you may also have a look at the John Melas Tools for the Motif XF, and the Waveform Editor in particular which is great for what you want to achieve, it’s organizing voices and waveforms nearly itself, eliminating unused and duplicated waveforms etc. and can read/write the native Motif XF files. Much easier using this than working with the Motix XF alone IMHO

Best regards Georg