David Baay

Yes, I do exactly the same, and have never encountered a problem. CW always generates a unique file name and, as you surmised, keeps track of which files are referenced by earlier versions of the project. Even files for clips that are no longer referenced by any project will remain in the project's audio folder indefinitely so long as some version of the project was saved during the session in which the audio file was written. This is why you might want to peridocally run Clean Audio Folder on an individual project or use the Cakewalk Audio Finder (CWAF) Tool to clean all audio folders at once to discard unused audio. Your situation is pretty unique; I don't recall seeinganything quite like it being reported in the history of the forums, and can't really imagine a mechanism by which this would happen to even one track much less multiple tracks or a whole project!

Cakewalk can only quantize notes and controllers independently. It has no way of 'knowing' which of possibly many affected notes the pedal CC event should follow when quantizing the notes. In my experience, quantizing controllers is generally not necessary, but it might depend on your playing/pedaling style. The amount that notes move when quantizing is generally a small fraction of the 'overlap' with the pedal down that's supposed to affect them or the 'clearance' between the note off and the next pedal up so its unusual that quantizing moves a note out of or into the influence of pedal down event that was or wasn't supposed to affect it. I guess I'd need to see a specific example of a performance for which you got an audibly bad result.

Go find Melodyne in Utilities > Plugin Manager and confirm that it's in a default location for VST2 (usually Program Files\Cakewalk\vstplugins\) or VST3 (usually Program Files\Common Files\VST3\). If it's under some other path, you can just add that path to Scan Paths under Options.

Yes, Freezing bounces processed audio to a new file, and the frozen clip references that file. Unfreeze will change back to referencing the original file, but won't delete the frozen/bounced file until you close the session. A Quick Unfreeze will set a flag so that when you re-freeze, it will just reference the previously rendered file instead of rendering a new one. Any file rendered in that session that's not referenced when the project is saved will be deleted when you close the session. The original file remains referenced internally so if you were to delete it while the track is frozen (and CbB is closed), you'd get a "missing audio" error on attempting to open the project. You should be able to look in the audio folder for the project and confirm the original file is still there, but if something has become corrupted with the file references in te project, you'd have to import it back into the unfrozen track to restore it.

Interesting. I was just going by this post:

What version were you on before updating? I'm not aware of any recent changes in this area. Can you share examples of raw and looped (i.e. before and after) clips with tempo info and the tempo of the project to which you're importing?

I'm curious that the waveform is displaying its color and showing selected when you have the track filter set to Volume automation. Normally that would cause the clip to be grayed out and unselectable. Also, what's that white line at the leading edge of the clip? It looks like a clip fade, but it starts slightly outside the clip, and a clip fade would disappear on freezing. Maybe it's just a video artifact...? I can't repro anything like it in any case. Is it just this project? Does the issue survive a re-open/re-launch/re-boot?

Product description says: "VirtualMIDISynth is a software MIDI synthesizer implemented as a Windows multimedia user driver, accessible as a standard MIDI Out device." This sounds like it's emulating an onboard GS/GM Wavetable synth chip. The 23.09 EA release explicitly eliminated support for this because it's not supported (or is losing support...?) in Windows 11.

Can you share a copy of one of these projects to be tried on another machine? Chances are it's your environment, but it could be something in the configuration of the project or CbB Preferences. In any case it's doubtful the files are corrupted in some way, and it should be fixable. If it's your environment, the first thing to check would be DPC latency: https://www.resplendence.com/latencymon EDIT: The tones might be from a plugin that needs to be re-authorized. But I've never heard anything like that with the transport not running. Or it's coming from another audio source sharing the driver or the hardware itself.

Having done thousands of hours of MIDI editing using many differnt techniques over the years (and made many mistakes along the way), I have to think this is due to a procedual error of some sort - likely involving the inadvertent selection of offscreen notes as you surmised. A seemingly random selection could be the result of using Process > Find/Change or the related Edit > Select by Filter. Are all the short notes the same duration now? Different note numbers and not the same note in differnet octaves? Different velocities, Different beat:tick times? Is the clip contiguous or made up of copied/slip-edited snippets? Hopefully you have earlier copies of the project from which you can recover the original MIDI. EDIT: I should add that Process > Find/Change or a combination of Edit > Select by Filter and the Event Inspector may actually be your friend in recovering from this. You can use them to find all the short notes and lengthen to some fixed value or by a percentage of the current length.

Instrument Definitions are for managing outboard 'hardware' synths connected to physical MIDI ports. If you don't have a standalone interface with MIDI ports and keyboard synths or rack modules connected to those ports or directly via USB MIDI, Instrument Definitions are not applicable. If you don't have that MIDI hardware, most likely you need to watch some Youtube tutorial videos on using software synths in Cakewalk. EDIT: I should add that support for onboard GM "Wavetable" synths and the MIDI Out port they usually present to the O/S and thus to Cakewallk's MIDI Device list has recently been removed from CbB since they are not supported by Windows 11. So if that's what you were expecting to see in MIDI devices, that's why it's missing.

Sampling and DSP don't work that way. The only thing that matters in reproducing and processing pitch information is the sample rate; and for purposes of accurately reproducing a vocal pitch and timbre (i.e. the mix of harmonics of the fundamental pitch) you just need to sample at rate that's twice the highest frequency you care about. A quick Google says female voices don't produce harmonics above about 17kHz, so 44.1kHz has that completely covered even without anti-aliasing filters. Bit depth is how precisely you measure the instantaneous amplitude of the sample and only determines sigal to noise ratio and dynamic range; it has no bearing on the accuracy of pitch and timbre analysis and processing. So there's really nothing to be gained from doing the FFT processing to change pitch at 32 or 64 bits vs. 16 bits. The more important thing is how you manage the harmonics that make up a voice's timbre to maintain that timbre as you change pitch, and the algorithms used to extract and process attack sounds vs. sustained tones and pitch bends. The sometimes gross alterations of timbre that Melodyne and similar tools cause are orders of magnitude greater than the miniscule and largely inaudible gains in harmonic distortion and S/N ratio you get from recording and processing at higher sample rates and bit depths. Those gross errors are due to shortcomings of the algorithms that are determining exactly how the audio should be altered achieve a natural-sounding result, not the precision of the alteration process. Audio stretching algorithms have the same issues, regardless of what sample rate and bit depth they're operating at. Machine learning may well help make breakthroughs in this area.

Not really any value in debating point by point since neither of knows exactly what's going on under the hood, but I'll just broadly say: - I'm pretty sure that any audio quality problems with Melodyne output have mostly to do with inherent limitations of the technology of editing pitch/formants, and very little to do with bit rate. - I've previously tested and verified that if you don't make any editing changes to a Melodyne clip and just bounce it to a new file, it will continue nulling to near silence (like -120dB) with the original file even after a dozens of generations. If you make changes, audio downstream of an edit will tend shift position by a few samples such that you have to re-align it (with sub-sample precision) to null with the original, but in the absence of that inverted-phase reference, there is no audible effect on unedited parts of the clip. - The issues around punching and comping with Melodyne appear all to be related to passing the muted/cropped/faded/clip-enveloped state of a clip to Melodyne and refeshing it every time an edit is made so that it always displays and processes only the parts of the clip(s) that would be audible in the absence of Melodyne. It's not 100% clear that this is purely a problem with CbB and not a limitation of ARA (which is a Celemony-developed technology IIRC), but given that the problem can be worked aorund as described it does seem possible the Bakers could fix this without changes to ARA. They may just not have anticipated all the ways in which a user might edit clips after an ARA FX had already been applied, and that some of those edits might be the result of automatic slip-edits applied by punching, comping, auto-crossfades, etc.

Exactly. Having the rug pulled out from under you implies your ***** is hitting the floor before you know what's happening. That's not the case at all here. All indications are that everyone will have plenty of time to make decisions about what they are or aren't willing to pay for and act accordingly over what clearly will be a matter of months. More like having the rug wear out under you. ;^)

Thanks for the tip. I just installed it, and told it not to migrate anything from other browsers for the time being. I don't save passwords anywhere other than my head anyway. ;^)

Is it only multitimbral synths that are failing? Possibly the transmit channel changed on the keyboard, and only non-multitimbral synths that don't require setting a matching receive channel are responding.

+1 on using Zip files to archive projects. To disable Audiosnap on all tracks of a project where you haven't made changes: - Ctrl+A to Select All - Expand the Audiosnap section of the Clip Inspector on any audio track. - Click twice to toggle the Enable option from '(Multi)' to checked and then back to un-checked.

It's not among the fixes listed for the latest EA update, and I verified it's still broken. Will likely have to wait for the new Sonar. EDIT: Somewhat good news: I just found an easier - and completely effective - workaround. Right click the track and Duplicate Track with all options enabled except Linking. All the sections of the clip will display in Melodyne and sound as expected in the new track. It should be a perfect copy of the original in every other respect so you can just delete the original. EDIT 2: Correction: You have to take both actions to fix it: Slip-edit the clip, Undo, and then Duplicate. Duplication alone does not change anything.

But you can't use live VST FX that way. I input monitor my keyboards all the time with reverb, chorus, flanging, delay, distortion, etc from plugins. With a 64-sample buffer I get a measured RTL of 4.5ms ( 3.6ms if the keyboard has ADAT out). This amount of latency is a total non-issue unless maybe you're singing, and I don't sing. ;^)

It's not in final release yet.

All plugins should work both ways. You might just need to make sure CbB knows all the paths that SONAR 2017 is scanning if any are non-default. Might need some clarification of the routing to answer your second question. If the Interface is not set to pass input signals to the headphone out (i.e. Direct Monitoring), and the tracks in SONAR do not have Input Echo enabled, you should not hear anything other the direct, acoustic sound during recording. What interface?

What exactly are you dragging, from where to where, and what are the exact symptoms of "no luck"?

The curse of familiarity is making everyone blind to the basics: Live input signals have to be echoed to the output of the track in order to get to a bus or output. You do this by clicking the Input Echo button in the track header: Off: On:

Two tips: - Drag in the lower half of the clip with the default Smart tool to get the Selection tool; dragging in the upper half will move the clip with the Move tool. - Holding Ctrl gets you a copy instead of a Cut, and you will probably want to hold Shift as well to constrain the drag to preserve the timing of the clip.

That might happen if you inaadvertently put two notes on one beat requiring an impossibly high tempo or, in this case, if you kept the three songs together and had a lng gap between the end of one an the start of the next that would require a tempo less than CbB's minimum of 8 bpm. Also, Fit Improv requires a note on every beat; notes on downbeats would be interpreted as quarters, resulting in tempos that are 25% of what they should be. But you can use that guide track with Set Measure Beat At Now, itshould give a good result. I would start by setting the Now time at the downbeat that should be 5:01, hitting Shift+M, and entering that measure and beat and OK. That'll give you a reasonably accurate average tempo so CW can accurately guess the other beats. Select the MIDI clip and use the Tab-to-notes feature to go out another 4 measure and Set that 9:01 which CbB should guess correctly so all hou have to do is hit Shift+M and Enter. Then you just keep going: Tab, Tab, Tab, Tab, Shift+M, Enter, Tab, Tab,Tab,Tab, Shift+M, Enter... as fast as you can. You can speed that up even more by binding a single key to SM/BAN in key bindings. I use it so much I bound S for "Set". If you make a mistake or get a "tempo out of range" error, Ctrl+Z to Undo and make sure you're setting the right note to the right measure/beat. At some point, you should stop and listen to the piece with the playback metronome running and see how it sound. If the performance was reasonably tight, a tempo every 4 measure might be enough. If there are rough spots, you can go back and set intermediate downbeats as needed or even every measure. The beauty of SM/BAN is that you can set as many or as few points as needed to get what you want, and you can even set fractions of beats (just remember that they're decinmal values, so an 8th note between to beats is .500, not 480 ticks). And finally, Set Measure Beat At Now is only as precise as your precision in setting the Now time to the transients. You can Tab to transients (Shift+Tab to go backward) in an audio clip just as you can Tab to notes in a MIDI clip, but CbB may not place the marker exactly at the beginning of a transient, and you will likely need to enable Audiosnap and raise the Threshold slider in the Audiosnap Palette to disable superfluous markers for transients between beats so you're not Tabbing 8-16 times to get through a measure. Clear as mud? Sorry, this stuff is always easier to do than to describe. ;^)