Environment CPU heavy?

[BJG]

Just trying to confirm with other users. I really haven't used the environment in any advanced capacity until recently getting into Bidule and routing my bidule data back and forth through the IAC busses.

More info;
Macbook pro 2.66Ghz i7
8GB 1067MHz DDR3
Logic 9.1.6 (64bit)

I'm using cable switcher faders set as text style. I have an IAC bus plugged into it from the physical input. I set the range on the cable switcher to 0-2. 0 is plugged into the sequencer input, 1 an instrument object, and 2 to nothing (dead end). Multiply this by however many objects I create in Bidule to perform different tasks and have the ability to turn on/off the data. Typically so far I'm not exceeding using more than maybe eight.

With these objects set up in the environment I'm finding my CPU use increased by up to fifty percent!

I would have thought the mere routing of MIDI data wasn't such a CPU intensive process no? I took a session with various bidules performing things like LFO's etc, and printed that data to the instrument tracks in Logic. When opening up a new session and importing those tracks, the CPU use is down significantly despite the same real time parameter modulation happening. The only difference is it's playing back from a region instead of being routed from Bidule. I kept Bidule open and synced to MIDI clock, so it's still running and sending data through the IAC busses, it's just not connected in Logic anymore.

So, is this just a new unexpected thing I'm learning about Logic's environment now that I'm finally using it? Is the environment just way more CPU intensive than I thought?


Thanks!

[David]

Hello BJG,

You're describing the use of IAC busses to route data from Bidule to Logic. Do you use any IAC busses to route data from Logic to Bidule? What kind of tracks do you have in your Arrange area: audio, software instrument, external MIDI? (if yes to the latter, what are they routed to?)

[ski]

Descriptions about the environment are rough to troubleshoot. Can you post a screenshot of your Clicks & Ports layer?

[BJG]

I use IAC busses to back and forth.

I have Logic send MIDI clock to IAC 16.

Typically I have external MIDI tracks pointing at IAC busses, and in Bidule the inputs are those busses. So for every software instrument I have in Logic that I want to have the performance data re-trigger processes in Bidule (envelopes, arps etc), it is played from an external track, routed through Bidule, then back through the physical input to the instrument. If I want to print that data, that's why I have it connected to cable switchers. I can route it directly to the sequencer input with target instrument record armed to achieve the same modulation only recorded now in the region.

[ski]

OK, so are you dead-ending the IAC "returns" at the Physical Input?

[BJG]

I set up a quick smaller environment that is more pic friendly. Here are pics of the environment and the Bidule setup.

Bidule;




Environment;



I couldn't get this screen capture to fit here.. but that one cable switcher gives the idea. Besides going to the instrument, it's going to the sequencer input, and that third option is "dead end".


Arrange;





I may be going about this in a weird way but it's kind of new territory for me. This way, the external tracks represent being able to arm a track running thru BIdule's processes. All I have to do is create a cable switcher for each Bidule process I wish to route. I can switch it to an instrument that it's cabled to, then switch it to the sequencer input and just record arm the instrument itself if I want to record that data.

The thing is, while this is going on it's hitting the CPU rather hard.. after I print all this down and scratch the environment routing, the CPU use goes down significantly, even tho the same parameter modulations are occurring.. the difference is being routed, to just paying back from a region.

[ski]

OK, this is great!

So... on the Physical Input's (PI) right side you see all of the Bidule and IAC bus MIDI "returns" into Logic. If you're sending MIDI to IAC busses, data goes out to your external processes and returns to Logic on one or more of those busses. So what goes out comes back in (even if it's in a modified form via PB), and thus there's possibility to generate a feedback loop. My feeling (just a hunch) is that you've got just such a situation going on and perhaps that's why your CPU is spiking.

Assuming this is what's going on, you have to prevent the feedback loop from occurring, and this requires a bit of explanation about the PI and it's SUM output...

All MIDI sources incoming to Logic are represented on the PI's right side (these are called ports), and here, "sources" includes MIDI output from PB and IAC as well. ALL data presented at the PI is merged, or "summed" at the SUM output of the PI, and this is normally cabled to the Sequencer Input (SI). When you record-enable a track you're patching MIDI presented at the SI to the instrument assigned to that track.

The way to divert MIDI data from any port away from the SUM is to connect it to any environment object. This is commonly referred to as "dead ending". Your IAC #4 is essentially dead-ended, but none of the other ones are (and thus there's the possibility of a feedback loop going on). So if you were to create a monitor or ornament and cable a port to it, that data stream will no longer reach the SUM output and thus prevent the possibility of a feedback loop. If you need to then utilize that data you can cable it directly to an instrument.

Hope that's clear (if not at least a little helpful towards your situation). If not, post back.

[BJG]

I'm aware of the SUM feedback loops. If you look closely at the environment picture, I've disconnected the SUM all together, it's connected to nothing.

Instead I connect the Caps Lock Keyboard or my MIDI controller (when it's plugged in) directly to the sequencer input when I'm trying to run things like this.

I did connect it to a monitor, that was "dead ended" just to see if there was any activity. None.

Thanks for the help so far BTW!

Hmmmm... so am I right in thinking this isn't supposed to be any sort of hit on the CPU? There is a problem here?

[ski]

I'm aware of the SUM feedback loops. If you look closely at the environment picture, I've disconnected the SUM all together, it's connected to nothing.

Oh jeez, how did I miss that? (I'll call the optometrist first thing in the morning, I promise!)

But what's feeding into your Sequencer Input? Can't see that in your screenshot.

It's good that you tested to see if there was any activity, just in case. So.... Hmmmm.... I create some pretty big environments and can't say that I ever noticed a CPU hit. And that's what leads me to suspect a feedback loop (which will indeed cause a CPU spike). But... Let's see if anyone else chimes in with brighter ideas than me, or at least more in focus.

[BJG]

It's good to hear that it seems to be an isolated problem on my end. whew' I was thinking "no way this can be a CPU killer can it?"

I was beginning to feel like the wind was getting taken from underneath my sails!

I really a musician so I've always been sticking to tweaking inside synths themselves and straight playing to a channel in the arrange. But I've had some time off between projects and now I'm really getting into modular stuff and going outside of synths themselves as part of patch design/creative process.

Baby steps I guess? Like I said I may be going about this a weird way, but it makes logical sense to me so far with the routing. It took me awhile to narrow other problems down like timing issues and avoiding using Rewire. (the timing/triggering of somethings on Bidules end is horribly buggy when synced via Rewire) I wrapped my head around it enough to start really writing something and then the CPU troubles started as I finally started layering in more instruments in real practice..

I've been using Logic since version 5, and I'm really comfortable and speedy with everything. I wasn't exactly afraid of the environment, just didn't really have a use for it outside of the arp object. I love no matter how much you get to know the program, there's always more there learn. So this new fangled exploration of the environment and modular applications is really refreshing/inspiring to me... just got to make it work in real practice!

Thanks for all the help again... I'll be up all night trying to narrow this down.

Great forum BTW! Super helpful and focused, I can't believe I never really paid attention to it till now.

[ski]

Y'know, it could have something to do with IAC itself and/or Rewire. Wish I could point you more in the right direction, but hopefully someone else will chime in and help you figure out what's going on.

Great forum BTW! Super helpful and focused, I can't believe I never really paid attention to it till now.

Oh yes indeed, extremely focused.

focused.jpeg (4.91 KiB) Viewed 181 times

[BJG]

Another quick question while I've got that pick of the environment up.

I like the text style cable-switching object, but is there a way to make it less sensitive when dragging up and down to select the values outside of regular OS X system preferences and track sensitivity?

It's hard sometimes to get that middle value of "1" without shooting past it. I like it better when it's limited to only two values because then I can just click on it to toggle the two, but at three I have to "slide" through them and it's a bit fidgety.

[ski]

No, unfortunately there are no facilities to adjust that kind of thing. About the only option available is to set the text object to "behave as menu" (there's a checkbox for this in the text object's text entry area). But in the event you don't like that action, getting the text entry area to open again to disable it is very difficult. Not impossible, just difficult. I find that I have to double-click wildly on the object to get it to open again.

[BJG]

Awwww yeah that is PERFECT!! Didn't see that little check box!

[BJG]

I think I may have narrowed this down!

It looks like I may have mis-understood how IAC works? I thought it behaved more like a "virtual MIDI interface". I guess each IAC bus likes to be a one way ticket in or out of Logic.

If your targeting an IAC bus from an external track, you can't use that IAC bus anywhere else for a return on the Physical Input. As soon as I made sure every connection going to and from both programs was not sharing "directional" duties, everything seems to be playing nicely on the CPU load!

Here is a pic of a new environment template I made to get me started. I've got a ton of monitor objects in there so I could keep my head in it and get visual conformation things are working in order.





I'm going to re-scheme my connections outputs in Bidule and from the PI to make more sense to build from here. It would of been better to think of everything in "pairs" i.e.----> External track to IAC 1, BIdule receives IAC 1 and returns that to IAC2.... External track IAC 3, Bidule returns to IAC 4 etc etc..

Oh man this is really neat!

[ski]

Cool, glad you're getting this sorted.