When I load it as instructed as an ALL file it seems to overwrites everything in USER bank 2.

When you use Load Type = All the data in the file actually overwrite all User data. That’s why it is called an “All” data file.

To load a single VOICE from a File:

Press [FILE]
Press [F1] CONFIG
Press [SF1] CURRENT
Select your USB stick
Press [F3] LOAD
Set the File TYPE = “voice”
Move the cursor to highlight the file name

When you press [ENTER] the file will “open” to reveal several lists of VOIES within that file.
Use the [USER1], [USER2] to view the different USER banks contained in your file

Select the VOICE you want to load
Target the internal location to which you want to LOAD it
Press [ENTER] to execcute the load.

The ES will load the VOICE and the samples it needs to play that Voice.

The sound bank I want to load creates 16 voices using the same 21.4 meg piano sample. If i load voice by voice, the Motif ES seems to duplicate the sample for each voice creating a new User Wave for every voice. Doing this, I quickly run out of DIMM. Alternatively, if I load the sound bank as an ALL file, it does seems to only load and store the new sample once, but, as you said, it overwrites all USER banks. Is there any way to only load and store only the 16 new VOICEs without duplicating the same sample 16 times?

If not, I guess all I really need is the sample. I can then add my own effects, EQ, etc... to my liking. but before I can do this, do I need to know the architecture behind the sample i.e. how many separate elements/waves are put together across the keyboard to make up a full playable sample? I believe there is only one USER wave (same one) behind each of the four voice elements . Is there a way to simple load the complete sample itself? Is that called the waveform?

A Waveform (in AWM2) is a collection of samples.
If you were to sample every note of an acoustic piano, you would have 88 samples.
These would be “collected” into a single Waveform.

Say that first set of 88 samples were all done with a very soft piano strike… you might name this Waveform “Acoustic Piano Soft”
If you were now to sample every note but a medium strike… you would call the Waveform “Acoustic Piano Medium”
If you were now to sample every note but a hard strike… you would call the Waveform “Acoustic Piano Hard”

Each Element references a Waveform (set of sample). So a three Element Voice containing these three collection, could be setup so that
Element 1 responds to Velocities of 1-59
Element 2 responds to Velocities 60-99
Element 3 responds to Velocities 99-127

Each Waveform can actually contain 256 samples.

When you set the LOAD TYPE = ALL or ALL VOICE, the synth does all the housekeeping — preventing duplicates.
Loading single Voices will indeed load the same data over and over.

There is a way around this… I maybe sketchy on the exact steps, it’s been many, many moons since I’ve even seen an ES.
LOAD All the data to the ES.
SAVE your own .W7W (All Waveform) file… this will create a file that only includes a single version of the custom Audio data.

When you power down all Audio data (Waveforms and Samples) disappear… the Voice data is still there.

You can manually reunited your 16 Voices with their Waveforms by powering up and loading the All Waveform file.

OK - I will give it a try...

Yeah, let us know. I remember that dealing with duplicates became a thing to be solved in the XS/XF… I don’t recall with clarity the solution for the ES. Too long ago… yikes.

Let me ask a related question - whenever I load an external voice library, the ES always fills in any blank USER voice bank ## with the factory set voices. Is there any way to stop the factory set voices from always loading as fillers. Utility setting?

Let me ask a related question - whenever I load an external voice library, the ES always fills in any blank USER voice bank ## with the factory set voices. Is there any way to stop the factory set voices from always loading as fillers. Utility setting?

No, not really.

Back in 2003-2007 (the Motif ES era), the User banks always include “placeholder” data. There are two User Banks of Normal Voices (Bank1 is original Voices; Bank2 is made from duplicated presets)… and 1 User Drum Bank of 32 Kits. Every file will include Voices in those locations (holding the place).

Current products, like MODX/MONTAGE, have the ability to manage its own internal memory (when you store as a new program it will create and occupy the next available location). There are so many Banks of 128 a new method is used — neither Voices nor the “placeholder” system are used.

Any Motif-series file will have “placeholder” locations that are occupied with either a factory duplicate or an Initialized program. This is why it is very important to find and read any documentation that is made for each commercially available library. It will typically tell you where they placed the “new data”.

Using the various LOAD TYPEs, you can usually target to load “All”, or individuals or “1 Bank” (a Bank being defined as 128 programs). Now most libraries will have 128 programs or less, and very, very few would provide more than that — given how expensive it can be to produce this type of data. But it is the nature of the internal synth platform that utilizes “placeholder” programs