David Baay

4th approach (or maybe just expanding on the 1st approach described above): - Set up just one audio track to record and echo the stereo output of the synth connected to one pair of inuts to your interface (preferably always available) - Shift+Solo (override) that audio track so that it won't be muted by soloing MIDI tracks that drive it or other audio tracks in the project. - Create a MIDI track for each channel of the synth that you want to use with output channel assigned accordingly. - Do a rough mix of the parts using MIDI track volume, panning and controllers as needed. - When satisfied with that rough mix and the arrangement, solo each MIDI track in turn, and Bounce to Tracks with Fast Bounce disabled, and Audible Bounce and Live Input enabled in the bounce dialog, specifying a new target track for each part. - Archive bounced MIDI tracks as you go, and when you're done, disable input echo on the audio track so you're not adding to the noise floor of your mix with his from the idling synth. - When using a hardware synth for drums, you can put all the MIDI parts for each kit piece in lanes of a single track, and solo lanes when bouncing those parts to get multitrack audio. One thing to keep in mind: For any parts that are really low in the mix, you might want to make note of the audio live level of the soloed part, temporarily raise the MIDI volume up to get a better signal to noise ratio out of the synth while bouncing, and then take the level of the audio track back down after bouncing using its Input Gain or Volume controls. This will help to minimize the effect of summing of noise floors from all the separate bounces. This would also be a reason to consider doing all the drum mixing with MIDI, and recording a single drum audio track.

Rendering Audiosnap stretching (either in place or bouncing to another track) will often move transients a bit, depending on the nature of the audio, and which stretching algorithm you're using. This is a function of the algorithm, and should be the same in any DAA that uses the same algorithm. I've had at least one case where Elastique really messed with the timing of a single transient; changes of a few ticks should not be a problem in most cases.

Are you using the Volume fader in the vocal track to change the volume? And all tracks go to the Master bus, and from there to interface main outs? And the vocal track does not have input echo enabled? I can only imagine one possible routing setup within Cakewalk that could cause this to happen, and it's so unlikely, I won't even descibe it. But if your using onboard audio interface with input set to 'Stereo Mix' or 'What U Hear', (i.e. recording/monitoring from output of the interface), and the vocal track is echoing input, then the track volume could affect the whole mix. But you should also have bad feedback in that case so... I dunno.

I was addressing this: "This forces me to create a new instrument track, copy-paste the MIDI data and delete the two unwanted tracks while I could have simply used the existing instrument track and associate a different VSTi to it." Replace synth does exactly that . You didn't mention anything about changing synth channels at the same time in our orginal post, but you can do that via the Inspector after you've replaced the synth. Bottom line: it works the way it works. An instument track is by definition a combination of a MIDI track and an Audio track connected by an associated instrument plugin. Changing that to allow an empty instrument track will require re-defining what an Instrument track is as an object or maybe introducing a dummy , internal 'placeholder' instrument that Cakewalk uses to fill the void under the covers. Or you could just to that yourself by using some lightweight 'placeholder' instrument in your templates. Beyond that, any changes you want to see in the way it's implemented will need to be entered as feature requests in the Feedback forum.

See "Replace Synth": https://www.cakewalk.com/Documentation?product=Cakewalk&language=3&help=Browser.11.html

No, Cakewalk has no tools to track down a specific plugin's contribution to a dropout. But it's likely that no single FX is causing the problem; it's the total load exceeding the CPU's ability to keep the audio buffer filled at the current buffer size. There should be a dropout code that you can check here: https://www.cakewalk.com/Documentation?product=Cakewalk&language=3&help=NewFeatures.22.html Depending on what you find, you might also want to double-check DPC latency which may have been affected by new drivers being used. https://www.resplendence.com/latencymon

No, I was wrong about that. Larry straightened me out. Buying with JamPoints does not invalidate license transfer.

Cool. Thanks, Matthew.

I'll probably download it all over a few days while working, but not try to install everything at once. I'll install one or two pieces a week, and it'll be like Christmas all year long. ;^) If I'm lucky, they'll release an installer before I'm done.

Ah, yes, I can. I had not clicked on the license button. Perfect! Muchas gracias, Larry. Made my day. 😎

In a perfect world, I would not have bought Modo Bass just two weeks ago, and everything in the package would be new to me because I'd be crossgrading from AmpliTube Leslie... but this is still a slammin' I deal can't pass up. Had hoped to transfer MODO Bass to my son-in-law as a gift, but bought it with JamPoints, so no-go, I guess. Really bites to have copies of things you can't even give away. Will really have to quit watching the Deals forum after this, though. After years of spending very little on new music gear or software, I've shelled out for quite a few deals in the last month or two. And it will take many more months to get to know it all even skin deep. Everything was on the order of 60-80% off even the usual discounted 'street' prices, though, so I'm pretty psyched to have gotten so much for so little. Thanks to all for taking the time to share, and especially Larry of course. Over and out. No, really, I mean it. Seriously, I have plugins to learn, and music to make, and fellow musicians to help. And I still have a day job (thankfully). So that's it. I'm gone. For good... 'til tomorrow, at least.

I don't have that update. I take it you had similar audio performance issues...?

Open the Console View, and pull the bus pane divider to the left. Then pull the additional pane divider at the far right back to reveal the hardware output bus to confirm you have signal there also. And hover over the abbreviated name to see the full name and confirm it's going to the ouput you're monitoring.

ThreadSchedulingModel=2 is recommended for quad cores and above. You should use that even if it makes no difference in this case. What sample rate are you running? TTS-1 supports only 44.1, 48 and 96 kHz, and at 96, you have to enable Light Load Mode in System > Options. You should definitely check your DPC latency. Spikes to 50% and above in a light to moderate project (in terms of plugin load) with buffer at 128 sounds high even for that older CPU. Always good to know, what 'normal' is in your system for future reference even if this issue turns out to be TTS-1-specific.

Something made the click volume decrease as well. Possibly it was delivered to you with a fade already applied - maybe specifically intended to obscure the click. EDIT: Nulling out the click could be made a lot easier by having a copy of the raw .WAV file with the click running out further.

Yeah, it's both - natural sustain decay with a volume fade on top of that. You can see the clicks fading in this screenshot of the waveform with scale maximized:

Got De-Click working without crashing. It helps but leaves behind the room sound of the click. Also tried inverting the click from a near silent section at the end with the guitar gated out per Lars' approach. As he said, it worked well in that area, but a fade has been applied to the click along with the guitar so the level is wrong for nulling in earlier areas. And it's a challenge to get the sample-accurate alignment you need to null. Figured out the click is at 149 BPM, but still had nudge each copy by a few samples. Not happy enough with any of my results to share anything. If Lars can make it work, he's your guy. ;^)

Well we're of to an inauspicious start, but there's hope! My Izotope suite has a 'De-Click' tool that looks like it might do the job, but it just crashed Cakewalk on a test fade+click clip I created! If I can get around the crash, I'll try it on your track.

Izotope Rx might have algorithms for removing 'pop' type noises that could reduce it. I have the limited set of Rx tools that comes with their Music Production Suite , but haven't had a need to use any of it yet. If you can post the fade out snippet, I could see what's possible. Or someone with the full Rx might try it for you.

If you can reproduce it, you might try recording just the click bleed, invert the recording, and mix it with the dry guitar track to null it out.

Please, everyone, start your own threads for performance issues. Aside from ThreadSchedulingModel=3 (which is not a default setting), every case likely has a unique root cause, and there are far too many users experiencing no such issues to conclude that Cakewalk has an inherent issue with its audio engine independent of the environment in which it's running, the content and configuration of the project, and preference settings.

Not sure about the Nvidia stuff. I might have disabled that, but don't recall there being more than one, or that it made any difference. Did you check the ThreadSchedulingModel? You might also turn off Plugin Load Balancing if it's been enabled. Quickest solution might be to use the Reset Config button in the Configuration file page. It backs up AUD.INI and replaces it with a default configuration (in lieu of manually renaming AUD.INI as we used to do). Then compare them if it solves the problem.

You can can access the Time+ offset control via the MIDI tab at the bottom of the Inspector for an Intrument track.

Anyone had any luck getting responses from Soundspot? My 'failed' order with them a couple weeks ago was in fact paid by Paypal, but I never got any confirmation from Soundspot though I'm continuing to get their marketing e-mails (yes, I checked Junk folder). I entered a support request with screenshots from their website, but heard nothing for a week. Just emailed nathan@soundspot.audio per contact info in the the Paypal transaction. Will see what happens, and dispute the transaction of I don't hear something in a few days.

Another shot in the dark that occurs to me: I gather Cubase installs something called Generic Low-Latency ASIO Driver that can cause symptoms like this. Any possibility you've installed/re-installed/updated Cubase recently?