Sure Steve, you can loop the audio back into any audio interface; the interface doesn't know it's coming from its own output. The only potential issues are the obvious ones: you might not want to give up a whole 8-channel lightpipe just for two channels; and you have to be careful not to set up feedback loops.

What I was actually suggesting is looping the Mac's optical S/PDIF back to its own in, and adding the built-in audio interface as part of an aggregate device. I haven't tried that to see if the latency is okay, though; I have been using the Mac's digital in to monitor my MBox 2 dongle for the Pro Tools software, and the latency isn't very good. But that may not be representative, since I'm going through an RCA-optical format converter and haven't tried lowering the buffer.

And thanks for both the compliment and subscribing. [;)]

By the way, both Jerome and I from my armchair were suggesting that making the stand-alone see separate outputs shouldn't be a big deal. It occurs to me that it's using the same output driver the system sound uses; am I right that other software hooks into Core Audio through a different pathway? That might not be so trivial a change.

Oh, and I should also report that Soundflower is the shizzle. I installed it yesterday, and it works really well.

@jbm said:

...
Also, have you heard the recent rumors about Clovertown-based (8-core) Mac Pros in November? Makes my geeky bones shiver.

J.

8 core Mac Pro??????? Where did you hear these rumours???

[[:|]]

M a y a

It was on macrumors.com -- a pretty good place for the latest gossip. The Clovertown processor is a four-core cpu, so two of those, and... welll.... [;)]

http://www.macrumors.com/pages/2006/10/20061025231946.shtml

The mid-November part of the rumor may be far-fetched. Basically, they're just pointing out that this is when Intel launches the Clovertown chip. So, if a Mac Pro revision were ready to go, as it were, it would have to wait for Intel's release... So, they're saying basically "any time after mid-November".

The Clovertown has already been unofficially tested as working in the current Mac Pros (the test article for this is linked from the article above).

cheers,

J.

Maya:

The rumours about 8 processor core Macs based on Intel's Clovertown chip have been rampant on Mac rumor sites for several months. You can check this out at:

http://www.appleinsider.com/article.php?id=2176

http://www.macrumors.com/pages/2006/10/20061025231946.shtml

http://www.anandtech.com/mac/showdoc.aspx?i=2832&p=6

http://techfreep.com/eight-core-mac-pro-a-reality-with-clovertown.htm

However, the story told by more conservative observers is that Intel will begin to ship the chips in large quantity in November and that the first Macs that will incorporate them may be shown at the MacWorld show in January and become available sooner or later after that. Then in the late spring OS 10.5, which is purported to be a fully 64bit (but compatible with 32bit applications) version of OS X is scheduled to be released. My own hunch that release of OS 10.5 may be followed by a release of a 64bit version of Logic.

I wouldn't rely on Macrumors to decide wether you should wait to buy a new computer or not. They're known for speculating a lot. Of course Apple will come out with 8-core computers. Nothing new here. What's total speculation is the launch date.

Apple came out with the Mac Pro in August, I highly doubt they are going to introduce an update like this 3 months later!

By the way, according to C-Net, the "Clovertown, a four-core processor, will start shipping late this year and hit the market in early 2007." It is very unlikely that it will be manufactured in large quantities before at least January, AND anyway the CPU will be probably too expensive when it comes out...

Someone posted this on the Logic TDM mailing list (which I still skim even though I'm not using TDM anymore).

http://www.xcellent-productions.de/downloads/activity2.png


[:)]

....... right [:D]

J.

So why does my G5 2.0ghz crash when VSL server reads 1.4Gb RAM (real mem)? Is this not normal?

Miklos.

@jbm said:

Hey Stephen,

Well, as I say, I haven't actually got it working with Logic. It just works taking direct hardware midi input, and sending audio directly to my RME card. So I don't know how it will work with Logic -- it just works in principle.

However, if I were going to seriously try, I'd probably try using the IAC bus or MIDI Patchbay for MIDI and Jack OS X or Soundflower for audio. Give it a go and let me know what happens!

Has anyone got any further with this?

I'm keen to use the Standalone along with VI plugins in Logic, but I have no idea how to do the MIDI and audio routing. (Something that works like Rewire to get back into Logic?)

I now have a Dual Quad with 8 GB RAM so I'd really like to maximise my resources here.

@jbm said:

Hey Stephen,

Well, as I say, I haven't actually got it working with Logic. It just works taking direct hardware midi input, and sending audio directly to my RME card. So I don't know how it will work with Logic -- it just works in principle.

However, if I were going to seriously try, I'd probably try using the IAC bus or MIDI Patchbay for MIDI and Jack OS X or Soundflower for audio. Give it a go and let me know what happens!

Has anyone got any further with this?


My apologies, I didn't see beyond the first page of the thread so missed all that great stuff re: Soundflower, jack and aggregate devices. Thanks everyone for your excellent theories, testing and explanations.

I now have two standalones working nicely.

1) output to Built-in Digital out connected to Optical S/PDIF in of Ensemble

2) Soundflower (Ensemble + Soundflower aggregate)

I've spent 3 hours trying to get Jack happy and no luck. Maybe I'm just being thick. Should I be able to use it to route a 3rd standalone output assigned to Ensemble, so that the Audio routes to a spare Ensemble DAW input (e.g. to 13/14)???

Ensemble doesn't have a loopback mode, i.e. it can't route outputs back to inputs without a cable. As far as I know the only ones that do that are RME's interfaces, but Metric Halo is adding a loopback feature to their interfaces.

The Ensemble sure does sound nice, though.

I think we need a step by step tutorial and links to the relevant sites / third party software, plus some diagrams and screen shots would help ! [[:)]]

This is a very clever and excellent work around by the way, congratulations to the masterminds behind it. I agree, VSL should release a standalone version with the option of selecting a different output - we're not asking them to implement this workaround as an officially supported feature, just quietly allow a standalone to support *choosing* the output (as opposed to multiple outputs) - they can even say "we don't support this work around we are just adding a tiny little bit more functionality to the stand alone in case people would need to use it" [[:)]]

Of course by now I imagine they are hard at work on the 64bit version of the server??!! But then again who knows when Leopard + Cubase / Logic + VI will be a workable up to standard and reliable set up?? It could take many months after the initial release of leopard before it's all bug free, and it will probably be the Cubase / Logic side of things that will take the longest to be worked out. Armchair speculation of course...

Miklos.

I'll put one up on www.VirtualInstrumentsMag.com shortly. We've had yet another letter on the subject, which I'm answering in the magazine, and I'll just stick it up there as a perennial article.

Give me a few days...

And of course I don't mean to preempt VSL in any way. [:)] That's in addition to what anyone cares to post here.

By the way the latest Metric Halo software's loopback mode works very well.

Hey you remembered I have a 2882 [:)] what a memory! I don't know my age or the year most mornings let alone the day of the week.

Thanks Nick you're a champion for the cause! Your good work is much appreciated.

Miklos.

As somebody who has just bought a new 8 Core Mac and would dearly like to be able to use more than the 3-3.5GB of memory Logic allows you to, this whole thread is like stumbling over Atlantis!

I'm going to try out the Soundflower/IAC approach outlined earlier in this thread, but I wondered if there wasn't a more stragiht forward way.....

Why can't one simply run a separate VST host on the same Mac, run the additional VI plugins in that, and route the audio into Logic via Rewire. This way 1) no additional audio drivers and routing is needed, 2) no additional MIDI routing is needed 3) you can save and load all your extra instances of VI, with their settings intact, in one operation and 4) you can have as many separate stereo pairs running via REwire into Logc as you want.

I must be missing something obvious here (and I admit I haven't tried it yet), because this seems like the easiest way to have double the amount of RAM available to VI. I've been experimenting doing exactly this on our WindowsXP machine, which runs the brass element of the symphonic cube, hosting the VI plugin in Acid (theoetically Reaper, Live, Fruityloops etc. could also be used), and routing it into ProTools HD on that machine using REWIRE. Seems to work a treat.

Please tell me why this can't be done on a Mac, so I can stop dreaming and get stuck into the whole Soundflower and IAC adventure.

Many thanks

Jules

Just going over all the details of this whole issue again, I'm wondering if the reason my suggesstion is not going to work so well, is that both sets of plugins (those in Logic and in the separate VSTi host) will use the same VI server application for hosting samples, meaning actually there's no benefit to having a separate host.

Does anyone know if that's the case? I suppose that's why the focus has been on standalone instances, which address memory as a separate application ..... oh well, it was worth suggesting.

Jules

Trailerman I think you are right. But here is another way of doing things on an 8core machine. Even at low latency you should get 20 instances of VI running at once, so why don't you just write 20 instances at a time. Then freeze say ten of them, continue adding / writing, freeze a few more. The 8-core should freeze pretty fast. If you use Logic, it will retain all the samples in memory - not much help. But if you use Cubase, it actually removes the samples and the VI instance from memory (but still remembers all the settings for when you unfreeze) - so you can literally write as many tracks as your hard drive can feed audio into the 8 core mac - 128 stereo tracks? Then you don't have to worry about a memory limit anyway. Of course you will not be able to have all tracks live at the same time, but being able to have 20 tracks of your choosing live at one time is not a bad trade off (but you will probably need to have many of those tracks optimised at some point as well, or else have less tracks live at any one time). Depending on what you load per instance 20 tracks is not unimaginable, but if you load a lot of samples on every instance, maybe 8-10 tracks would be more doable from a memory standpoint, still, again it's not a bad scenario. Freezing on an 8 core shouldn't be too painful. On my dual 2.0 G5, it's VERY painful...

Miklos.

Trailerman, the answer to your question two posts ago is that the plug-in gets one helping of RAM regardless of how many hosts it's running in. The reason is that the plug-in is actually a MIDI/audio router to and from a separate program running outside the host. That means its RAM space is separate from the host's, but there's only one program running on the machine rather than lots of plug-ins.

On the other hand each instance of the stand-alone program gets its own memory allotment. You have to rename each instance, of course - Copy of V.I. Player, or whatever.

Come to think of it, I wonder how hard it would be to make the program launch multiple instances of the VSL plug-in server application? That would let you run multiple hosts, and it would also solve the "no individual audio interface outputs addressed" issue...

Hm. Another one for the feature requests list?