CAKEWALK SONAR 6 PE (Windows DAW) - NOW WITH CONCLUSIONS!

[Anderton]

So far, so good? Is this all making sense? Anything not clear?

[DeeringAmps]

All clear here!

 

Thanks,

Tom

[MSaether]

Looking good, Craig.

 

For users who spend a lot of time with AudioSnap, I highly recommend you learn and use the various keyboard shortcuts listed below. They are sure to speed up the workflow, and you can use the key bindings feature to re-assign the shortcuts if desired (go to Options | Key Bindings and choose the "Track View" bind context).

 

AudioSnap Enable = F12

AudioSnap Add Transients To Pool = Ctrl+F12

AudioSnap Show Transient Markers = Shift+F12

AudioSnap Go to Next Transient Marker = Tab

AudioSnap Go to Previous Transient Marker = Shift+Tab

Set Measure/Beat at Now = Ctrl+M

AudioSnap Auto Stretch (Follow Tempo) = Alt+F12

AudioSnap Insert Marker = Ctrl+Alt+I

AudioSnap Reset Selected Transient Marker(s) = Ctrl+Alt+R

AudioSnap Delete Inserted Marker = Ctrl+Alt+Backspace

AudioSnap Disable Marker = Ctrl+Alt+D

AudioSnap Promote Marker = Ctrl+Alt+P

AudioSnap Snap to Transients = Ctrl+Alt+N

AudioSnap Audition Beat = Ctrl+Shift+Spacebar

AudioSnap Show Palette = Shift+A

AudioSnap Show Transient Pool = Ctrl+Alt+F12

AudioSnap Select All Markers = Alt+Shift+A

AudioSnap Clear Selection = Alt+Shift+C

AudioSnap Select Moved Markers = Alt+Shift+M

AudioSnap Select Stretched Markers = Alt+Shift+S

AudioSnap Select Disabled Markers = Alt+Shift+D

AudioSnap Select Enabled Markers = Alt+Shift+E

AudioSnap Select Promoted Markers = Alt+Shift+P

 

These default shortcuts are also listed in the online Help (Help | Keyboard Shortcuts).

 

Cheers,

 

Morten Saether [Cakewalk]

[Anderton]

Thanks Morten, that's useful info. I've learned some of them, but haven't committed them all to memory yet...

[Keith-Cakewalk]

Hi Craig,

 

Nice workflow cases.

 

What really excites me about this feature is the ability to breathe life into projects by giving robot drums some humanity. Sooo much nicer on the ears. We hear so much today on the radio that is locked mindlessly being quantized with all the life out of it that I can't even stand listening for more than 30 seconds. With this feature though, I've heard several projects from people using AudioSnap and you don't even realize that AudioSnap was used. It's just fun to listen to.

 

Now as far as your alignment with the end of the clip... There's a really easy way to nail that. Just make sure the snapper is set to clip, click the now time so it's right at the end of that clip. Then, use the Set Measure Beat at Now and type in the measure and beat it is. It will most likely even guess the right number from the get go. Then, you don't have to play with the tempo to get it to line up. Similarly, you could have done that by dragging the clip so it starts on the measure you want, use the Set Measure Beat at Now at the beginning of it to create a tempo anchor. Then, use the same command on the end, and you have an instant indication of the tempo for that clip. No guessing, back and forth, etc.

 

Oh yeah, almost forgot, you can use Set Measure Beat at Now with MIDI. So, that means you can freely play on your MIDI keyboard, then later go in and mark the measures and beats to line up the ruler. For some people/styles, it's way easier to do that than the Fit Improvisation method.

 

Enjoy!

 

Keith

[Anderton]

In the Pro Review thread on the Konnekt 24D, I was asked about latency of the drivers, and the lowest latency you could get without clicks and pops. To "even the playing field," I opened up new projects with no other tracks or plug-ins, and adjusted the samples in the Konnekt 24D applet. The following are the lowest latencies I could achieve with reliable results:

 

Cubase 4: 128 samples

Ableton Live 5.0.3: 128 samples

Acid Pro 6.0: 128 samples

Tracktion: 64 samples

Guitar Rig 2 (stand-alone mode): 128 samples

Sonar (ASIO): 256 samples

Sonar (WDM): 256 samples

 

At 128 samples, Sonar was close to acceptable, with just occasional clicks and pops, but was still unuseable. For all programs it was the same computer, same interface, etc. Are there any "magic bullet" check boxes or parameters I should try using with Sonar to bring its latency down to the other programs?

 

I must say I was impressed by Tracktion's ability to operate reliably at 64 samples, but I'd settle for being able to get 128 samples with Sonar!

[Anderton]

I'm copying a post from the Sonar forum from Cakewalk's John McCarty. I was having problems with some devices showing up as MIDI instruments, even when I specified "Configure as Synth" under plug-in properties. Turns out this is a known bug; here's a workaround until the fix appears:

 

1. Open the "Cakewalk VST" registry key in a registry editor:

{HKEY_LOCAL_MACHINE\SOFTWARE\Cakewalk Music Software\Cakewalk VST\}

 

2. Find and select the affected plug-in's file name under the "Inventory" key

 

3. Select the "isSynth" value and set it to '1'

 

4. Restart SONAR and this plug-in will show up as a VSTi

[Matt Hepworth]

Craig, thanks! That's really laying it out well and it's helping me follow along with AudioSnap. A couple "What the?!?!" questions have been answered.

 

Excellent work!

[Anderton]

First of all, Keith's trick about going to the end of the clip and setting the desired measure and beat for the now time works perfectly. It's a nice time-saver, to say the least, when you have a loop and aren't sure what its tempo is.

 

He also mentioned something I've alluded to, which is being able to quantize to grooves. Although there's a quantize to groove tool if you want to quantize to a specific groove, the Quantize to Pool feature is the main tool for quantizing one part to another.

 

The "pool" is a collection of transient markers from one or more clips that basically determine a "master quantization grid" for a project. Clips can then be quantized to this pool of transient markers.

 

For example, suppose you have a drum loop that was played by an actual human with a great feel, and you play a rhythm guitar part along with it. You can AudioSnap-enable the drum part, adjust the Sensitivity and Threshold controls to create transients on the hits (as well as disable hits you don't want to be part of the pool), then click on "Add Transients to Pool." Click on the attachment to follow along with what we'll be covering.

 

The upper track is the drum loop with transients added. These have all been added to the pool; look along the bottom of the screen shot, and you'll see these transients extend below the next track down and form dotted lines on the clips view background.

 

The lower track is a rhythm guitar part that has also been AudioSnap-enabled and has transients at the beginning of each chord. Again, looking along the bottom, you can see that some of these transients match up perfectly with the pool (hey, my timing's not that[/] bad!), but others do not. What we'll do next is quantize the rhythm guitar part to the drums.

[Anderton]

Now the stage is set, as it were, and quantizing to the pool is simple. I clicked on the rhythm guitar part to select it, selected "Quantize to Pool" as the task, then clicked on the "Quantize to Pool" button. Done! Click on the attachment to see the final result.

 

Note how the rhythm guitar attacks now line up perfectly with the pool, and therefore, with the drum part. On playback, the rhythm guitar part didn't sound that great due to the stretching, so I selected the iZotope Radius Mix algorithm and rendered the rhythm guitar part by Bouncing to Clip. The offline rendering did its job, and the guitar sounded just fine.

 

Incidentally, as you might expect, you don't have to quantize right to the pool; there are quantize strength sliders, and you can also set a "distance from pool" parameter that causes notes further than a certain distance from a pool not to be quantized (there's also a slider for "fine-tuning" this window). It's really quite cool.

 

Again, though, let me emphasize that the more complex the part you're quantizing, the more likely you'll need to do some manual labor. For example, AudioShap added some transients in the middle of sustained chords; lowering the sensitivity got rid of too many desired transients, so I went in and disabled the unneeded transients by hand so they wouldn't be shifted to a pool transient.

[Anderton]

And I should at least provide an audio example, right? Attached is and audio example of the post-quantize to pool guitar part; note how it lines up perfectly with the drums.

 

Before we get off this topic, remember that the pool of transients can come from any clips, and in any quantity. For example, the pool could consist of transients that hit only at the beginning of each measure, and you could quantize the bass track so that only notes within 1/16th note of the beginning of each measure are quantized -- making sure the measure start attacks hit right on the beat, but everything plays as it did originally.

 

Now it may seem like a lot of work to do this kind of processing, but as always, just because you can doesn't mean you must. Yes, you can use AudioSnap to quantize everything to everything, but I think its real value is as a problem-solver.

 

For example, I have a bunch of loops from Dr. Walker's "Cologne Cyclez" sample CD that really kick butt, but most of them use swing. I found a bass line loop that would have worked perfectly with it, except it didn't have swing...but with Quantize to Pool, it did!

[Keith-Cakewalk]

 

Originally posted by Anderton At 128 samples, Sonar was close to acceptable, with just occasional clicks and pops, but was still unuseable. For all programs it was the same computer, same interface, etc. Are there any "magic bullet" check boxes or parameters I should try using with Sonar to bring its latency down to the other programs?

 

 

I assume it was the same sample rate so same latency.

44,100 2.9 ms

48,000 2.6 ms

96,000 1.3 ms

 

However, was the bit depth set similarly? In the Global Options, Audio Data tab, try different settings for the file bit depth. If the test file you're using is 16-bit, matching the Import and Render depths may help.

 

Another setting to check is the I/O Buffer size. You'll find it on the Audio options dialog, Advanced page. For some systems, better performance can be gained with a lower value, others need a higher value.

 

Keith

[Anderton]

Well, well, well...I tested Sonar 6 with the Creamware SCOPE interface and got reliable operation over ASIO at 3 milliseconds, and with the E-Mu 1820m at 4 milliseconds, also with ASIO. This was a somewhat more complex testing environment that the other tests, as it had a few drum tracks and an AmpliTube2 plug-in.

 

Therefore, it seems that issue is related to the driver in the Konnekt 24D. I'll follow up on this in the Konnekt 24D Pro Review thread so as not to hijack this one.

[abzic]

Hi all

Great info here! I wonder if I can create the pool from a midi track.

Thanks

Andres

[Bud Ice Man]

No, you can't, you must render to audio first. This is the one area in which Beat Detective is up on Audio Snap, but rendering is really easy, so...

[Anderton]

AudioSnap also has a REX-like function in that it can slice a clip into individual beats (by beats, I don't mean quarter-notes necessarily, but individual slices of sound), then quantize the start of each beat (each of which is now its own clip) to the grid or the pool. The advantage to doing things this way is that no stretching is required to do quantization. The disadvantage is that you need to have beats with fairly clean breaks in between; also, after quantization, a beat might move forward a bit and if the next beat doesn't move, there may be a gap between the two. Whether or not this matters, of course, depends on the particular material.

 

This basically works like the other AudioSnap functions we described: You set up the transient markers so that each beat begins with a marker. You don't want any markers in between beats. Then, you click on the "Split Beats into Clips" button (the one that looks like a pair of scissors). This splits the clip at each transient marker; click on the attachment to see the clip after it's been split.

[Anderton]

Now you can quantize the beats. Remember, each one is essentially a clip, so what we want to do is quantize the start of each clip to our desired quantization grid.

 

Click on the attachment to see how this is done. The crucial part is that under "Change," you must select "Audio Clip Start Times" and not select anything else to change. Also, note the "AutoXFade" parameter. If a clip moves forward into the end of another clip, you can have them crossfade automatically over a certain period of time. The screen shot shows 20ms, but for the clip shown here, I changed it to 5ms as the clip was a relatively percussive guitar part with chords stabs.

[Anderton]

We're almost there. Click on the attachment to see the final result. You can see the start of each clip is now perfectly aligned with the beat, but also notice the small gaps and crossfades.

 

At this point I cleaned this up a bit by eliminating the "silent" spaces (they actually had some noise) and doing a quick fadeout each beat. As I wanted to convert this into a single clip, I selected all the clips and did a "Bounce to Clip." I then slip-edited the end so the end landed exactly on the beat, and bounced again. Perfect! I had my loop.

[Anderton]

I spent quite a bit of time last night after posting this creating some guitar loops, and came to the conclusion that the "Split beats at clips" functions is the most squirrely of all the AudioSnap functions. One problem was that splits would sometimes occur even at transient markers I'd disabled. Another was that sometimes the reverse would occur: Splitting at clips would split in one or two places, but not at markers that had been promoted.

 

There may be some pilot error involved, but I'll be working with this function some more so if it's a problem at my end, I should figure it out...if not, I'll file a bug report. Bottom line was that for the most reliable results, it was usually best to right-click on each marker where I wanted a split, and select "Split Beat."

[johndefiore]

 

Originally posted by Anderton I spent quite a bit of time last night after posting this creating some guitar loops, and came to the conclusion that the "Split beats at clips" functions is the most squirrely of all the AudioSnap functions. One problem was that splits would sometimes occur even at transient markers I'd disabled. Another was that sometimes the reverse would occur: Splitting at clips would split in one or two places, but not at markers that had been promoted. There may be some pilot error involved, but I'll be working with this function some more so if it's a problem at my end, I should figure it out...if not, I'll file a bug report. Bottom line was that for the most reliable results, it was usually best to right-click on each marker where I wanted a split, and select "Split Beat."

 

 

I had similar experiences with "split beats at clips". Sometimes it refused to split at all the beats, just picking a few that it liked. I tried to get a Cakewalk employee to show me what I was doing wrong at the AES show, but everything worked flawlessly there when I tried it on their system (of course). The funny thing is that it's working fine on MY system now and I don't know what's changed. Anyone else having issues?

 

Regards,

 

John

[Keith-Cakewalk]

Apologies for that behavior.

 

We've identified a few issues with split beat that we'll be addressing in a future update. For now, clear your selection after doing the first split before doing the next ones.

 

Keith

[tekrytor]

A quick question regarding Sonar 6.

 

Can I get the same recording and playback transport control features using my GNX4 (or similar units) in Sonar 6 that I enjoy with ProTracks Plus?

 

Thanks,

[Anderton]

That's an excellent question! I suggest you go to DigiTech's GuitarWorkstation.com site, where there are instructions on how to install the DigitechRPX.dll plug-in for control surfaces (you can download the plug-in from there as well). It should work with Sonar 6 (it worked with Sonar 4 just fine) as long as Cakewalk didn't change the control surface plug-in format. I'll try this myself, but you might get a chance to check this out before I do...if so, please report back and let us know.

[tekrytor]

Thank you Craig,

 

Actually, I asked because I am considering upgrading to Sonar 6 (do not own it yet) and GNX compatability would be a definate PRO factor in my decision making process. But without owning Sonar 6, I can't just try it out myself.

 

Your excellet review process here is also a factor, I've been a spellbound reader of yours since the 70's with your GP columns and your "Electronic Projects for Musicians" book, etc; so your reports in this forum are highly influential for me.

 

So, keep on keepin' on! It's great to know YOU're trying what we're buying before we buy, too! Especially at the higher end of the market.

 

If I might add, it's also refreshing to see that Cakewalk is actively online here. Too many companies are missing the boat by avoiding popular user groups for their products. Another PLUS for Sonar 6! My upgrade to Sonar 6 is not far off at this rate.

 

tekrytor

Steve M

[Anderton]

My hard drive died awhile ago, and although I didn't lose any data thanks to backing things up (that was a close one), I needed to re-install my programs from scratch. I hadn't re-installed the GNX4 software yet, so your post gave me an excellent excuse

 

As I was a little rusty on the details of installation (I installed when I first got the GNX4, and it stayed on there until the drive died), I went to the tutorial I did on hands-free recording with Sonar, and downloaded the control surface plug-in and the command to register the plug-in. (This is all documented in the tutorial.)

 

I installed the USB drivers, and the "beta" ASIO drivers. (These have been beta for years now, so I doubt they'll ever be updated. They work, but you can't adjust the latency; you're stuck with 10ms. However, this may not matter, as we'll see.) Then I set up "DigiTech Hands-Free" as the control surface, and it fired right up -- hitting the record footswitch created the track, record-enabled it, and started recording... the Stop and Undo functions worked, in fact the whole thing worked right up to spec. That was quite encouraging, I must say!

 

There are a couple things of which you should be aware:

 

1. When hands-free creates a track, it record-enables the track but does not enable input monitoring ("input echo"). Therefore, you'll probably want to set a mix of the guitar with effects, and the return from Sonar via USB. That also gives you zero-latency monitoring, as you're hearing the guitar while you play it rather than through Sonar.

 

2. If you hit the GNX4 Stop footswitch after recording, recording for the track is disabled. But if while recording you hit Play, that's equivalent to hitting stop except recording is not disabled. If you then hit Record again, you'll continue recording in the same track, as another layer. This is a handy technique if you want to build up a ton of sounds in one track.

 

Anyway, thanks for asking me about this...setting it up reminded me just how cool it is to use the GNX4 as an interface with hands-free recording options.