I feel that I should clarify a few fine points

,,, i love it ,, when ya ,,,,,ya know,,, "just do your thing",,, "B" ,,, ,,,,,,,,,,,,now if you were only that ,,,,'fluent: on your feet,,,,,,,

Hello B Rock. Please stay tuned for our next broadcast....I am now formulating some more questions. I have had less time the last day or so. Thank you for your patience.

Hey B Rock. Back from our commercial break.


Sorry to keep it on ADSR's, but I think this is the last clarification I need to fully understand this, so we can move on. Here we go. Tell me if I am correct on this. I know I might be repeating myself from early postings in this thread, but by doing so will help me better put the pieces into the puzzle. I am excited at this point. I think I am finally getting it. Will see what you say.

I think this diagram on ADSR's is much better than the other diagram. It helped me better understand atleast I think.

What do you think?

this diagram and ADSR in general is basically saying that you press a key and it goes into an attack stage, in basic terms, you are now hearing a sound that is at its loudest. Next you get into a decay stage. From my understanding, the decay stage is the time it takes for the sound volume to get to the sustain level. In Dimension, it could be either upward or downward sloping. The sustain level then maintains the sound at that volume level, which can be louder, lower or at the same level of the attack stage until the user releases the key. That brings us to the last one, the release stage. Release is saying that the user has now let go of the key and then determines by time when it goes to silent or reduced to zero.

Here is another one for you B Rock. How could you describe this method if you are not using a keyboard or a user pressing any keys?

Like you said: Diagrams have you believe that this is what happens all the time, which is not true and as this article stated here confirms what you said that about diagrams:"Just to confuse matters - but only slightly - a sound does not have to have all four phases. A woodblock, for example only has an attack phase and a decay phase. An organ has an attack phase, a sustain phase and a release phase but no decay phase. As well as ADSR generators, some synths, such as SynC Modular, have an AD (Attack/Decay) generator."

To take this further in dimension. Once again I might be repeating myself from my other postings in this thread. Sorry.

We are in the graph area now. To repeat again, are we safe to say that any two nodes are always upward sloping in an attack stage? Are we safe to say for the sustain level that it is two straight nodes going horizontally in the graph area? If my understanding is
correct that the decay stage could be upward or downward sloping, could you give a few examples of a decay stage?Obviously for the release stage the nodes would be in a downard slope. Am I correct?

If you can give a few examples in dimension on some node patterns in the graph area that would show instances of the four stages that might give me a better understanding.

How am I doing?

Thank you B Rock for being patient with me on this topic. Looking forward for your response.

Stay tuned again....more to come on our tv station!!!!!!!!!

could you give a few examples of a decay stage?

I got something else for you, Jeff. I don't know why I didn't try this before; back when I first thought of it. I missed your last request for some Decay examples. I'll go one better: you're going to dig those up yourself.

Load up an instance of the ol' venerable P5 sampler: the DS864. I hope that you have good eyesight, because the "DS" part stands for "Don't See". Show Instrument, so we can take a look. It'll load up the default Rhodes preset, and you're going to mangle the crap out of it.

But first, press F1 while the DS has the "focus". It'll bring up the Help file, so that you can check out the DS864's Envelope section. It's got a great little diagram there that's similar to your link, but it has a few extra sections to explore. Those would be the Hold section(s), and the secondary Decay parameter. This'll give you a good idea of what's what when you run into those. If you want to emulate a "standard" ADSR envelope in the DS, you just leave the Delay, Hold, and Decay II sliders at their lowest settings (either at 0% or 1 ms.).

OK, you've got your bearings somewhat, so let's take a look at the envelopes. There's two Filter EGs, an Amp(litude) EG, and one for Pitch. Starting to look familiar? These are DADHDR envelopes (it should be DADHDHR, though: seven-stage EG). It stands for Delay, Attack, Decay I, Hold, Decay II, and Release (Sustain is implied as a level, and partners with Hold). When you look around, you'll see a few knobs in the EGs. The designer felt it best to illustrate the level parameters as knobs, and that's a nice separation for explanation purposes. Notice in the diagram that the times are displayed horizontally, and the levels? It's beginning to take on another Dimension.

It's an unfair comparison, though, because as advanced as the DS-864's EGs appear, that's still only the equivalent (if that) of eight nodes with no curves, and limited to a few shared levels. But there's a nice pair of features in that microscopic set of three buttons, to the right side of the DS's EG's. One is Normal, which is self-explanatory. The second one (Repeat), causes a section of the EG to loop back, from the second Decay stage finish (could be like the S trick in Dimension), back to the Attack stage again, and on & on. Release is still intact. The last one (Free) will go through the stages no matter what. Now your Note On acts as a one-shot trigger, and when you finger-lift a key (Note Off) has no bearing on the envelope travel. It's going to play through, no matter what.

OK, so you need to experiment with ADSR's, right? Just move around some sliders, but start with the Amp EG. With the default Rhodes patch, the slower Attack time is going to be obvious. But slide up the Release slider, and you'll hear the lengthened trip to Ground Zero. Crank the Sustain level down, crank up the Decay, and hit the Free button, and toy around. With your current level of understanding to this point, I think that you'll start to be able to visualize an envelope shape with differing slider positions.

See that Time knob? That scales the entire envelope to be larger and smaller; and you'll get some interesting effects by varying that. It'll speed up and slow down the entire enveope contents, and it may be easily to hear what's going on with different adjustments.

Now, we'll return back to our originally scheduled program: Already in progress ...

Yeah, Jeff: all I did was copy the link of said picture off of its Properties, and paste that into the reply. Then I highlighted the link in place, clicked the image button, and did a little centering. But that particular pic was already up on someone else's server, and I just referred to that spot. It's a little different with "homemade" stuff. I have all the tools for image creation and processing (how about a 3D envelope?), the chops to do it in no time, but no server space that I could depend on to place it. I'm working on that last piece of the puzzle, and then "Watch out!".

"Holy Modulation Device, Batman! Do you think that the Riddler has run out of questions?"
"Not so fast, Boy Wonder. That devilish fiend will stop at nothing until his nodes are curved all over Gotham City."

You can be as nutty as you want to be.

"Danger, Will Robinson. Danger!"

Well B rock, while I am waiting for your response from my previous post, I would like to add a few more things. I am pretty much at a closure with the ADSR thing, which I am sure you are bored of by now. I just need one more clarification when it comes to the ADSR concept in dimension EG's.

Thank you for that example with the ds864. That was the icing on the cake when it comes to ADSR in general. I understand the concept fully.

In my post above I said: "How could I get results like above now with dimension? That is the next step." to explain further. When it comes to ADSR the DS864 instrument illustrates what that concept does. Excellent. When it comes to Dimension, unless I am wrong and I am sure you will tell me, is that it doesn't seem to follow the ADSR. As I try to produce sounds with the EG area, do you recommend me not going by the ADSR concept or thinking about it and just move nodes around until I am happy with the sound? The reason why I am getting confused with this is that as for the DS864 example you were able to use the ADSR method and shows it, when it comes to dimension it seems like it is in its own world.

Otherwise, moving on to the next step and thank you again for all your help.

I'm getting a little behind, huh? What do you know: real-life rears its ugly head again. I may have been getting a "little behind", if you must ask. (Actually, I'll be glad to see this year come to a close.)

OK, where were we? I'm assuming that you were using the default Rhodes patch in your DS864 experimentation. Unlike a straight synthesizer (and like Dimension), the original sample is going to have a huge impact on what you hear further down the road (in the synthesizer processing section). Just like a synth, you've presenting a base sound that's getting modified by the later modules. So, if you start out with a sine waveform (in a synth), then there are going to be less harmonics for the filter section to sculpt with, yet a saw waveform may give you "too much". It depends on what you're after. With a sample-based synth, the "problem" is exaggerated, because the sample itself (for example) may have a finite length that gives the Amp EG very little to work with.

That said, I plugged in your numbers (with the Rhodes), and couldn't verify all of your results. Those times that you're using (1.7 ms. jumps out) are very small. [Edit: Oops! Read twice, Tom. Comment once. Looks like I mistook seconds for milliseconds. I'll try this again later.] The human ear has difficulty discerning anything directly with anything less than 5-10 ms. (there's a psychoacoustical exception with the attack time). Delay times and the like are another story, and the fact is that small Decay times will affect the latter stages, so it's easy to confuse what's actually going on.

There's another possibility: you're using pattern-based notes to trigger the ADSRs, and the length of those notes can change the whole feel. You don't need a MIDI controller, but it's easier to determine instinctively where the sustain portion ends and the release times begin when you're physically lifting your finger from a key. Here's a suggestion: create a "test pattern" that includes all manner of notes. No matter what the length, give it a measure of its own to develop. So you might start with a note that's two measures long, follow that with a one-measure note, then give a whole measure to a half-note, another measure to a quarter-note, and on down the line. It'll be easier to hear the differences.

The following won't apply to the DS864 (for the most part), but Dimension's sample content often only spans the comfortable note range of the original sampled instrument. You may hear nothing at the extremes. Speaking of which, I tried hard to duplicate the "slight sound" left over in the "-INF" track setting. I'm just not hearing that.

Hmmm ... where would you use all of Dimension's envelopes? Let's see: many plucked and wind instruments will exhibit an initial Pitch change in the attack portion; depending on how hard it's struck or overblown. Filter characteristics almost always change over time with acoustic instruments, and that would cover the cutoff frequency and (slight for a natural effect) resonance. Amplitude? That's going to always change no matter what; even slightly in an organ simulation. Panning is more of a movement effect, so you (most times) would leave that in a static position, but there are cases where a shift in soundstage position (instrument movement) or an effect desired would demand this. Think of a Leslie rotating speaker, for example. Most places, though, you may only be using a single or few EGs on any given Element, unless it's all-out synthetic creations. But you could find a need in a hard-plucked guitar through a Leslie, as a single example.

OK, we're ready for Dimension's EGs. You know what traditional ADSRs do, and Dimension's build on that. But they go well beyond that with flexibility. Remember that I told you that you can simulate an ADSR in the Dimension graphs? If you want an ADSR, you build an attack stage with two nodes, add a decay stage with another single node, then the sustain stage with either the Qwerty commands or another node, and modify the slope of each stage by adjusting the curves. Say you want to mimic the DS864 and it's DADHDR EGs. Add a second node at a level of zero at the point in the timeline where you want the Delay to stop, and start the Attack stage from there. Then add another node to represent the Decay/Hold transition, and adjust the curves between.

I'm sure that you're getting the picture now. You take the basics of ADSR and synth development, and use the knowledge of those standard modules to create any type of EG you can think of. Hmmm ... I'd like a mandolin-style picking attack that repeats for every note played, but I want the attacks to build up from moderate to strong, to simulate a "fade in" of louder up & down picking strokes at a rapid pace. So you place several (maybe dozens) "attack" stages very close to to each other in the timeline, and slightly change every other level to simulate the up versus down pick attack (direction effects timbre & strength of attack in the natural world). Yet these "pairs" of attack stages can be gradully increased in level in relation to the next pair to mimic the fade-in. That came out more difficult-sounding than I intended; I hope that it was a good example.

You've heard of "Build it, and they will come"? Well, Dimension's EGs are: "It's already there. Dream it, and you can build it." Or "See it or hear it, and you can recreate it." You were right: Dimension is in its own world with this. Others come close, but this one takes the cake (or Cake takes it. They know what they've got here.).

As for the velocity parameters: Yeah, they are primarily intended for keyboard use, but why let that stop you? The default automation parameter in the pattern Editor is MIDI: Velocity. When you click in some notes, they'll all come in at the same velocity (64 of 128, I think; halfway). But you can click on that middle Editor tool, and draw in different values for velocity in your patterns. You can also insert a Velocity MIDI FX plugin to a track, and get a variety of velocity responses by playing around with those settings. Or go to the Process menu, and use Scale Velocity to get some variations. Once again, it might be easier to get the tactile feedback by inputting "harder and softer" notes with a MIDI controller, but don't think that leaves you out in the cold. Change up some velocities, and then see how that affects the Dimension parameters mentioned.

I'm not bored with the EG discussion. I want you to be clear on this. Get this, and you're well on your way to programming custom patches in Dimension, and have a comfort level in doing the same with any other synth. Things are getting a little crazier on my end, Jeff, so there might be these little delays in response. But we've come this far, so let's stick with it until it's second-nature and absolutely clear to you.

Phil: Do you mean that type of horn "after-crescendo" that builds up after you release your fingers from the keyboard? Or more like a second attack stage that swoops back up while you're still holding onto the chord? For the first, the MIDI Matrix comes immediately to mind, and using the Release Velocity parameter to control a second level stage. That parameter use is KB model-dependent, though. What were you using (again)? Maybe it's a pattern, and this won't apply. I'm not sure about the Note Off stuff in P5: it's inconsistent at best.

The second could be accomplished with an initial high-level node, a second at mid-level, and a third (constituting a second attack stage) back up high again. Then adjust the time placement and curve shape to taste. There are other ways to accomplish both scenarios, but I'd better narrow down the intent before I go crazy on this one. Edit: Duh! Would've been nice for me to tell you that this would be done in the Amplitude EG in each Element, to override the "hidden" EG that takes overall level to zero. Sorry, Phil: My brain is fried here lately.

Is it obvious what you're after in your posted song? We've got some poles down here today, so the ol' dialup connection is a little more frisky than usual. I'll see what I can do, but the downloads are hard-fought this weekend.

ORIGINAL: b rock

Phil: Do you mean that type of horn "after-crescendo" that builds up after you release your fingers from the keyboard?

That'd be it, yep.

OK, Jeff: I've got to stop you right here before you hurt yourself. You don't need to have the envelopes with identical layouts. More often than not, each envelope will have it's own distinct shape, and the node counts and curve shapes will vary wildly. With the graphical envelope, you're simply describing level changes over time, and this shape will change a lot from one type of envelope to the next.

The reason that I came up with that bulleted "recipe" above was to coax you into drawing an association between the actual resulting shape of an envelope, and how that might affect a particular parameter destination. I thought that if I mirrored the same envelope across 5 different destinations, it would be obvious how the shape affects the sum total. I'm not sure why you keep having problems with the filter envelopes, but I concentrated on those in the recipe first. I used a high-harmonic base sample, and a steep filter to accentuate the filter action.

Cutoff and resonance go hand-in-hand: the CF is going to give you a base frequency in the overall spectrum as a starting point, and the resonance will feedback some of that original frequency back to the input of the filter again. The envelopes take that base frequency, and move it around; allowing some higher partials in, or lower; depending on the position in the envelope. But there's another thing to consider: there are different types of filters, and different ways to configure those types, and some are "steeper" than others. For now, let's just say that a shallow filter is more subtle (and harder to hear, unless you know what to listen for). A steeper filter [more "poles"; that's the "p" part] is going to hit you over the head. I won't get into the electronic aspects of this, but you can hear a steeper filter's actions more easily than another, and a little bit of [stationary; using the GUI knob] resonance will accent it even more.

Be careful with the resonances; you've found this out already. Another thing: the cutoff is the line of demarcation, but filters are built differently, and use this dividing line in different ways. A highpass [HP] filter passes everthing through above the CF; a lowpass [LP] under the CF, and a BP [bandpass] centers a relatively narrow band at the cutoff [and reject [BR] subtracts this "band" from the center; leaving the rest to flow through.] There's others: check of the possibilities in Dimension. But that's another discussion entirely, and I've listed the most common types. You can also place filters in series, or parallel. Suffice it to say that you (usually) need a sound source rich in harmonics to subtract anything from it in the first place, and that filter types, routing, & configurations play a huge part in how obvious the filter action is. And by extension: in how obvious the Filter envelope action(s) will be. This is something that will come; you just have to play around with it enough to get what's happening underneath.

To serve as a quick ADSR notation: You got everything right, with one caveat. If you don't define a Sustain stage (using "s"), the envelope will play through from start to end. If you want to define a "Release stage" in this scenario, you need to set the last node at a zero level. The envelope will be the same length *always* if the note length that triggers it is at least as long as the envelope's total time. If you define a Sustain stage with the Qwerty, the envelope will go through all the previous stages until it reaches the Sustain node; then hover there until the actual Note Off arrives [the note is finished]. Then it will go through all the stages [nodes and curves] after that sustain point, reaching a true Release stage finale only if the last node in the timeline is at a zero level. It's a little difficult to get your head around this; just let it simmer for a while.

Pitch envelopes are fairly easy to distinguish; that's why I started using them as primary examples for demonstration purposes. One type of destination has nothing to do with the others [except for the cutoff and resonance pair, and perhaps you could stretch it to (loosely) include the amplitude and panning pair). It bothers me that we keep coming back to "only the first two nodes work". You have to look at the Time parameter displays in the envelope sections, and relate that to the rest of what's happening in Project5. Your actual note length will have an effect on what you hear; I think we've covered this. Your overall project tempo in P5 also plays a role. It's defined in Beats Per Minute, and that sets up a relationship between the actual note lengths and real time. If you've set up a bunch of nodes, and the display stretches over 1000 ms. [by default], everything is happening within a time span of only one second.

If you have a short note length and a fast tempo, you're skewing that relationship towards a higher ratio of "beats" over time. All I can suggest for right now is to make sure that your envelopes extend past a few seconds while you're still learning, that your triggering notes are long enough to complete the envelope cycle, and that the project tempo is slow enough to let you hear everything. This problem that you're having really gets to me, because it means that I'm not explaining things clearly enough for you to understand. I hope that I can do better as this thing progresses; maybe I'll come to a "Eureka!" conclusion for explaining this.

I think that the reason the DS864 was easier for you to set up is that all the ADSR-type stages are pre-defined. That takes some of the grunt work off of you, but the tradeoff is in lost versatility. With Dimension, you have to do the heavy lifting yourself, but once you get it down, it'll take only scant seconds to emulate any other envelope configuration or shape. Try what I suggested: Set your initial "attack" node where you want it; up or down, and place a "Release to zero" node at zero level; way at the end of the timeline. If you want, drop that "s" red line somewhere close to it; way towards (but not at) the final Release point. Then fill in the other "stages" and nodes between those two extremes. I'm counting on this helping you to construct the envelopes easier, with that last "Release to zero" and sustain point in place in advance.

The light bulb's going to come, Jeff. With your persistence and enthusiasm, it has no choice but to light up soon. Stick with it, man. You're almost home.

that type of horn "after-crescendo" that builds up after you release your fingers from the keyboard

Didn't forget about you, Phil. I've got to give my two index typing fingers a break. <g> I'm trying to remember your KB controller; I seem to think that it was a hardware synth at one point, and that's a good thing here. It just may provide for true Note Offs, and you can take advantage of that Release Velocity parameter in Dimension's MIDI Matrix.

But I'd rather write up a general recipe that's work for all KBs in all situations, and the "crossfade Elements" thought of mine comes up hollow. I can do what you're looking for easily in some of my hardware, but I need to translate that to Dimension in a few different ways, to decide on the easiest way and the fastest method of programming that in. I will get back to you ...

Hey B Rock, I got some great news. I figured it out!!!!!!!!!! Thanks to you and the good ole piano roll.

Whew! My index fingers are thanking you profusely! I was this close to becoming a monk ...<g>

Before this one gets away, though, let's clarify the terminology. Piano Roll refers to the editing section in in P5's (or other app's) Editor section, and implies using the Free mode. It's the part of the Editor interface where you actually draw the notes in, like the roll of punched paper that you'd see in a 19th-century player piano. It allows you to input notes of varying lengths into the timeline. Were the Editor in Step mode, this would be closer to an emulation of an old analog sequencer in hardware. The note lengths are all the same size, and the relationship to tempo speed determines the "note length", in a way.

What you've found to be useful in your explorations is called the Editor's preview keyboard. It serves as both an input device for notes (along with selection and de-seletion of existing notes), and as a display device of what's happening in the Editor (incoming MIDI notes from a KB, or notes played back in a pattern). I won't drive this point into the ground, but you need to grasp what the common meanings are so that we're on the same page. And I don't want you to have to "unlearn" anything later.

I can see how you could easily come to that conclusion with the naming convention, though. This one's my fault. If I'm to take on an explanation of this depth, I should put myself into the exact same position as the reader; as objectively as I can. That means that I should try to approach the problem with fresh eyes, and in the process, use only what you had at hand to complete the project. That means foregoing any MIDI keyboard input, and then I might've realized the value of using the preview keyboard in the discussion. I'll remember this one.

You don't really need less notes when working with envelopes, but you do need a note of a certain length to hear the entire length of the envelope develop. Later on, you can use this to your advantage, because inputting notes of different lengths actually changes the amount of the envelope that's exposed. So, if you've got an envelope with multiple decays in it (for example), shorter notes will reveal only one decay, but longer notes reveal the later decay stages, too. You can control this with note length, and it becomes a performance technique. It makes it difficult in the learning process, but it opens up the possibilities later in the learning curve.

I'm so glad that you've now got the ADSR thing down, and made the connection between that EG structure, and the graphical node system used in Dimension. There's the groundwork complete; all that's left are the special nuances and applications. I'm tempted to "celebrate" by turning an envelope into a low-frequency oscillator [LFO], and a mini-step sequencer, but now I don't want to rush things out. Let's see if you have any more fine points to clear up, and let's not jump ahead to the unusual before we cover common practical applications.

And there's bound to be something that I missed. Look at the length of this thread! But I can see the light bulb is ramping up to full brightness; isn't it?

Phil, I was overthinking this one too much (surprised?). The "after-crescendo" is extremely easy. I was using the trumpets f.sfz from Dimension Pro (it may be unique to that version).

There was already an Amplitude envelope in the program. It has just two additional nodes (beyond the initial "0.0" timeframe one). One is out 60.2 ms., at a maximum level of 1. That matches the "instantaneous attack" of the initial node; also at the maximum level of 1. The "s" Qwerty command has dropped the "red line" sustain point marker there, making it the pivot point between holding your note, and what happens after you release your finger (Note Off). The third node drops to zero after 100.7 ms of TDIF; landing at zero level at an overall Time of 160.8 ms. It gives you a fairly rapid release time, which adds a bit of a reverberant feel to the trumpet (accentuated by the small Room reverb in FX).

Enough background. I included the specifics in case this wasn't in Dimension Semi-Pro, and so that the envelope could be reconstructed. So here's the simplest way to change it to an "after-crescendo" buildup:

1). Add a single node after the last node (beyond the "Release to zero" node). Drag it out to around two seconds or more at a maximum level of 1.
2). Drag the new curve downward. You may also want to bring the Release to zero node before it up in level slightly, and adjust it's position forward in time for a smoother "decay" and "buildup" combined curve.
3). Adjust the recipe and season to taste.

Note that the crescendo will occur on every note played, so it can get a little unrealistic (tape-reversed trumpets) without impeccable performance technique. There are other ways to skin this cat that are a bit more controlled (and slightly more complicated), so let me know if this is what you were after.