David Baay

I got a CAL script working pretty nicely for this. It's got some awkward structure in places because I haven't worked with CAL for a long time, and couldn't recall how to do some things (like getting more than one thing done in the 'else' part of a conditional). And I'm sure it could be made to run faster but it works. I haven't tested much in the way of corner cases, but could share it after I've hammered on it a bit more. As a test project, I found a particularly bad example of a 3-minute improvisation with 49 re-triggers. The highest velocity 'ghost' was a 16 with most well below that. I was playing so lightly that many notes had velocities in the single digits all the way down to 1 and some of the re-triggers were at higher velocity than the initial note. The gaps between them ranged from 1 to 9 ticks so there's no mistaking them for deliberate repetitions. It was an interesting exercise and will be a useful tool so thanks to the OP for getting me motivated to do it.

Really sorry to hear that. See you next week.

This a really good point. I often record without a click and then use Set Measure/Beat At Now to align the timeline to measures as necessary to get everything close enough to the grid to be able to do a percentage quantize if needed. Even with very tight quantizing, leaving the tempo variations from one measure/section to the next in place helps preserve the live feel. And working in the opposite way, it will be a lot easier, and probably more effective, to change tempo every few measure than to edit individual note timings within the context of a fixed tempo. And ramping sections up/down with Insert Series of Tempos could also be effective, and sound pretty natural if done correctly.

For a simple one-to-one mapping of input channel to output channel (or Omni to 1), and MFX is not needed. The forced output channel assignment in the track will take care of it.

If it's an Instrument track, you will need to go to the MIDI tab in the Inspector to access the output Channel field. If it's a pure MIDI track, the Channel widget is in the track header when showing 'All' or "I/O" controls.

There's also the issue of continuity for users who have a well-established workflow that depends on the old implementation. Minor changes like removing the phase buttons from the track view (eventually restored by popular demand) have generated long acrimonious threads on the forum in the past.

Many moons ago, I fixed a bad key on my RD-300s by swapping a strip of velocity sensors from the top of the keyboard where I almost never play to the middle where the problem was. If it turns out the RD700NX's problem is consistently on certain keys, I might try that, but I suspect it's more widespread.

The fact that that there are three notes in a row on the same key exhibiting the problem in the OP's screenshot suggests there might be a problem with the sensors on that particular key. It's also interesting that the velocity of the 'ghost' notes is at or near zero. I'll have to check some recordings and see if any particular keys are more common offenders on my keyboard and what the velocities are; I didn't think they were that close to zero.

This is very interesting, My Roland RD700NX does this sometimes when playing very lightly. Usually the second note is such low velocity that I can just ignore it. But if my OCD gets the better of me I will sometimes go through a recording and manually 'glue' them to restore the full duration . It typically doesn't affect more than dozen notes in a 4-minute piece so is not a lot of work, but I've recently been considering writing a CAL script to fix them for me. I had assumed it was a defect in the keyboard design that it could re-trigger so easily, but I just Googled and found that the RD-700NX has a 3-sensor keyboard. It might still be considered a defect that it's so hyper-sensitive, but it's good to know it was designed deliberately to better emulate the repetition mechanism of a real keyboard action. It does play beautifully otherwise.

Might be easier to find the specific issue if you can share a copy of a project to be analyzed.

One man's "core features" are another's "superfluous relics". ;^) I forgot about the phase buttons because I haven't used that dialog in anger for a decade. or more. Yes, there are some dated dialogs and UI elements scattered around Cakewalk. In most cases I suspect this reflects their relative unimportance to the majority of users. The development team has always been pretty small, and there's a limit to how often they can re-visit existing functionality that might not be pretty but is perfectly serviceable. I don't really have any objection to your suggestion except to the extent that they will take precious development time away from adding new features and continuing to polish what I consider to be the 'core features' needed for recording, editing, arranging, mixing and mastering songs. And in these areas, Cakewalk generally is as polished as anything out there.

Ah, yes. I didn't realize the un-labeled boxes at the bottom are for phase inversion.

If you need more input level to TH3 to get the desired sound while recording, you can add a gain plugin ahead of TH3. An empty FX Chain will let you add 6dB.

True, But that can't currently be done from the Process menu, either. It seems this post is as much about adding new features as it is about presenting them differently, and not just about Gain as the subject suggests.

Yes, but if you're not assembling an album, you can just select in the timeline and export as Kalle Rentaho suggested. The mistake the OP made was selecting all the tracks first which is not necessary. CW will export everything in the time range if nothing is selected. You can also set a punch or loop region and use the option in the export dialog to get the time range from that.

As Mark said, automation is the more modern way to handle this. The beauty of digital audio editing in a modern DAW is that it can all be non-destructive. You can continually tweak your edits as needed and immediately hear the result without having to commit them until everything is sounding like you want. But you're using the old-school shoot-first-listen-later destructive tools. Just use automation (and all the other non-destructive editing tools) and when you've got what you want, bounce/export the audio.

This only happens in Comp recording mode and is a necessary feature. In order for Comping to work properly, every lane has to have clips with matching boundaries so that the content of only one lane is heard at any given point. Marled's solution of using a punch region to define the boundaries for all takes and stopping the transport outside it is a good one, Otherwise the usual procedure when you don't want any part of the last take it is to just delete the last take lane and then heal the split in one of the remaining takes, either by sweeping through the whole clip, starting from outside it, or by selecting the two clips and Ctrl+clicking one of them with the Comp tool.

Disabling the Volume widget will stop CC7 from being sent when playback starts but won't have any effect on what happens when playback is stopped. In any case, the issue here is CC1.

... and duration. Much of the "mechanical" sound of step-entered MIDI is due to a lack of natural use of duration as well as velocity. I've often posted that I can hard-quantize the start times of a live performance without totally killing the "feel" so long as the velocities and durations are preserved. My advice is to get a keyboard controller if you don't already have one and spend all the time you would have used trying to humanize your step-entered MIDI learning to play it. I maintain the even with very limited ability to play a keyboard, you will still get better (and faster) results by recording an imperfect performance (at lower tempo if necessary), and removing the 'excess imperfection' than by trying to do the opposite. And the more you do it the better you'll get and the less you'll have to fix. Plus it's just way more enjoyable!

PDC [Override] can only work on audio/instrument tracks whose output path does not go through the PDC-inducing plugin. If the FX is on the track you're monitoring or on a bus in that track's path to Main Outs, there's no way to bypass the delay with the FX active.

I agree it would be inappropriate to implement this just to work around NI's non-standard MIDI implementation, but I have long wished for the ability to selectively zero controllers anyway. The only controller I typically need or want to have zeroed on stop is CC64 - Sustain. In most cases all others can stay wherever they are.

FYI - Chrome froze twice attempting to download. Life's too short to drink cheap beer... or use cheap synths. ;^)

Yes, I actually removed that from my original post because interface buffer options don't generally go high enough to be a problem, but BounceBufferSizeMsec definitely can. Glad that sorted it.

Yes, I'm switching the Edit Filter to Clip Gain first by Shift+clicking one of the envelopes. I did not consider that a 'hoop' ;^)

Works without jumping through any hoops here if I lasso-select (right-click+drag) the multiple clip gain envelopes (use Ctrl to add non-adjacent clips to the selection).