Key off volume when setup for velocity response follows a volume relative to your strike (note on) volume. If you wanted to offset this volume over time, you could. You could use motion sequence or perhaps LFO. I think motion sequence may offer slower ramp down speeds although both should be able to trigger when a key is pressed and then slowly lower volume over time. The downside is that these are not individual by each key pressed - so you're holding down keys while playing staccato then your held down keys may lower the volume for the staccato notes. It would work best for single-note playing. If you solved this problem by using each note on (trigger type) to bring/reset the volume back up - then this would impact held notes played simultaneously with staccato and change the volume ("incorrectly" ) of them - the held notes would have their release volumes brought back up when the staccato note is played. So there's a narrow usage that would have this approach work.

What you would want for individual note control would be a note-on amplitude envelope that is used for note-off XA elements. But the amplitude envelope for note-off notes will start when you release the key. There's not an option for such an envelope that would follow each individual triggered note on key-on and apply to the key-off notes.

Is it possible to adjust the volume of key of samples according to key touch/ push time.

The volume of the Key Off Elements is not only influenced by the Note-On velocity, but can additionally be influenced by the current Amplitude Envelope Generator Level of the sounding Elements in the same Element Group.

Part of the XA CONTROL feature is Element Grouping which can determine how an a Element will behave. For the “Key Off” articulation the "Element Group“ can influence the behavior as follows:

If you place a Key Off Element in a separate Element Group, it will trigger independent of the other Element’s AEGs — allowing you to create a situation where the Key Off Sound is not influenced, at all, by what happens over time to those Elements being played by initial Key On; it will simply trigger when key Release is reached. It’s output level will not be influenced by the AEG Level.

If however, your Key Off Element is in the same Element Group as the sounding Elements, the initial output level of the KeyOff Sound will be determined by the current amplitude level reached by those Elements. In this case, if the main sound has completed its path (the Key is still pressed and all Time and Level points have occurred) the output level of the Key Off Sound will be equal to that final Level.

In other words... If the main sound has reached “Decay 2 Level” and that Level = 0, a Key Off Element in that same Group will output = 0
If the main sound has not completed the envelope when the note is released, the Key Off Sound will reflect the output Level reached at the point of key release. A staccato note would likely cause more output (assuming the sound has a front loaded Attack). Sampled envelopes typically reach maximum Level shortly after Key On... therefore, a staccato note will have a high amplitude Level, and will create a high amplitude for the Key Off Sound.

An acoustic piano chord held for an extended period will eventually create a 0 amplitude for Key Off... as the AEG Level dies out, the quieter the Key Off Sound — if the Key Off Element is in the Group.

Background
With AEG “Time” and “Level” parameters you describe the change in loudness of the Element from Key On until Key Off. In general, percussion instruments (those that are hammered, struck, or plucked) will have an envelope where “Decay 2 Level” is 0. This can be understood because in these instruments eventually the triggered vibration will stop (friction with the air, gravity and other forces work against it staying in Motion).

In general, self-oscillating instruments (those that are bowed or blown) the musician can continue the vibration indefinitely by maintaining pressure on the instrument. These amplitude envelopes will have a significant (number) value for “Decay 2 Level” which is the loudness of the held sound (Sustain portion of the envelope)... a bowed string or blown trumpet will continue to sound at this Level until Key-Off. This is where a sample-based instrument will utilize the ability to cycle play a portion of the waveform (loop) until you release the key (at which point the Release Time parameter takes over; Release Level is always 0

The Key Off Sound, in this case, will reflect the level of this sustain portion of the Elements within the same Element Group.
The output level at the time of release can influence the initial output level of the Key Off Sound.
If the Key Off Element is placed in its own Group, its behavior will be independent of the other (sounding) Elements.

Use Cases
Because in the AWM2 engine each Element has all of the component building blocks to be a standalone instrument, and because each can contribute sounds under specific conditions (including Note Range and Velocity Range), you can have different behaviors for different performing styles.

Key Off can be used to create a fixed pitch ‘sound’ - the mechanical noise within an instrument, or it can be used (as is most often in pianos) to recreate the specific sound of felt dampers returning to pitched strings. It could be used completely differently to simply create trigger events as you release keys... for example, it could start the playback of an audio clip (sample Loop).

If you trigger a note with a specific velocity that value can be used to select a specific sample within a Waveform. For example, if your Key Off Element is a three-way Velocity switching Waveform. 0-60, 61-104, 105-127. The Velocity of the Note-On will determine which Key Off sample will sound, and the Amplitude Level of the sounding Elements will determine how loud it will sound (when Key Off occurs).

This allows for multiple Velocities to be used — when creating legato solo instruments, for example, you can have different “legato” velocities for soft, medium and hard strike notes.

BTW: This works. Something like the single PART pianos already use AEG to decay the sample - so the key-off sample is placed at a suitable volume relative to where AEG tapers off.

Where this doesn't work so well is where AEG is not used to decay a sample. Where the sample itself decays on its own. Therefore, AEG will not decay. The "CFX Concert" is an example where the AEG "doesn't" decay - but rather lets the sample itself taper down as long as possible. Therefore you get better sustain from CFX Concert (vs single-PART CFX Performances) - but this also means that the key-on AEG doesn't work to attenuate the key-off sample amplitude if holding the key for a long time (long gate times).

... I'm just focusing on what's happening to AEG here - not worried about the more gory details about CFX Concert where different PARTs cover different velocities or different ranges of the keyboard. What I'm saying is ignore those details. Just saying that similar (even single-PART) where AEG doesn't decay will have this kind of "limitation" requiring making compromises to the programming/performance in order to make this AEG grouping key-on to key-off work.

... picking back up: To play by the rules - you'd have to edit the note-on AEG so it does decay which will have an impact on the sustain length.

One could record piano samples that actually get louder over decay1/decay 2 (over the gate time) as you hold a key. Then use AEG to decay - where the end result sounds like a naturally decaying string that ends up matching the natural decay. This would be a trick just to allow for the key on AEG to be at lower levels as you hold the key and not impact the sustain by using AEG on the key-on elements.

Lack of compression makes this kind of thing expensive because pianos, done well, mean lots of samples. Uncompressed ends up taking lots of space. Regardless, I'm not saying to actually do this - but, as an approach, that synthetically increasing the volume of an element in order to allow to cancel this with AEG is one way to go if you want the key-off element to have lesser amplitude after holding your key-on (gate time) for some duration.

It's good to see that this does work and can be easily "played with" using an initialized AWM2 with some simple tweaking to see how it works under the best of conditions. YMMV

Mr. jason ıt is finally ok for key off sampling like Nord realistic pianos. You can:'t achive key off element especially in piano samples with long decay and release. however if you create a ghost element (e.g.cerate an empty sine wave sample in element 7 with a short decay) and group the release sample(key off) to elemt 7, and then turn to volume down to zero of element 7 ( not mute!! if you mute the elemet you can't hear the release sample). Finally the keyoff sample volume and release behaves according to the elemet 7. the biggest disadvantgae you sacrifce one element for keyoff sample I will soon load audio recording.

Octave shifting a sine wave low enough will also make it inaudible. At least on my monitors. Results may vary I suppose.

BTW: This works. Something like the single PART pianos already use AEG to decay the sample - so the key-off sample is placed at a suitable volume relative to where AEG tapers off.

Where this doesn't work so well is where AEG is not used to decay a sample. Where the sample itself decays on its own. Therefore, AEG will not decay. The "CFX Concert" is an example where the AEG "doesn't" decay - but rather lets the sample itself taper down as long as possible. Therefore you get better sustain from CFX Concert (vs single-PART CFX Performances) - but this also means that the key-on AEG doesn't work to attenuate the key-off sample amplitude if holding the key for a long time (long gate times).

... I'm just focusing on what's happening to AEG here - not worried about the more gory details about CFX Concert where different PARTs cover different velocities or different ranges of the keyboard. What I'm saying is ignore those details. Just saying that similar (even single-PART) where AEG doesn't decay will have this kind of "limitation" requiring making compromises to the programming/performance in order to make this AEG grouping key-on to key-off work.

... picking back up: To play by the rules - you'd have to edit the note-on AEG so it does decay which will have an impact on the sustain length.

One could record piano samples that actually get louder over decay1/decay 2 (over the gate time) as you hold a key. Then use AEG to decay - where the end result sounds like a naturally decaying string that ends up matching the natural decay. This would be a trick just to allow for the key on AEG to be at lower levels as you hold the key and not impact the sustain by using AEG on the key-on elements.

Lack of compression makes this kind of thing expensive because pianos, done well, mean lots of samples. Uncompressed ends up taking lots of space. Regardless, I'm not saying to actually do this - but, as an approach, that synthetically increasing the volume of an element in order to allow to cancel this with AEG is one way to go if you want the key-off element to have lesser amplitude after holding your key-on (gate time) for some duration.

It's good to see that this does work and can be easily "played with" using an initialized AWM2 with some simple tweaking to see how it works under the best of conditions. YMMV

I am working on a Rhodes patch and I have a 5 velocity layer Release sample. How would I be able to make this release sample work like it does in a real Rhodes? As the base Rhodes sound I have loaded in this performances decays in sound (in the sample, not in any settings), how I do make the release samples, regardless of the note on velocity, also play at reduced volume to match the base Rhodes?

I'm not good with all of the Montage terminology. So in a basic sense, I have a Performance with Part 1 Rhodes, Part 2 Release Samples. I hit a chord at velocity 127 and sustain it, once the samples reach near their end, how do I get the release samples not play at max volume but rather at a reduced volume to match the Part 1 Rhodes sample volume reaching it's end?

I think it sucks as I have to have this release sample to have multi-velocity layers. So I have to have the note on velocity trigger the release samples' velocity sample, but I don't want it to play at it's intended volume over time. Immediately is fine, but I need that decay in volume.