Martin,

thanks for your reply. I was afraid that this would be the cause of the problem. sadly, I also don't know of a way to disable this dubious feature.

DG,

thanks for chiming in.   I do run uncoupled and load stuff manually. regardless, all VIF should disconnect when closing a project and they are not and this is the problem I try to solve. This I believe is independent of whether I "should" unpreserve or not.

Please tell me what "disconnected looks like" other and a VIF loaded without the word "(connected...)" after the name and the little plug icon changing.

Paul,

I'll send screenshots. But the procedure is simple: open project - connect to some VIF - close project - on my setup all connected VIF remain connected (both localhost and on slaves) - a before and after screen shot would show the exact same picture. When I load a new song, all VIF are first disconnected before the new connections are being made. So that part works. I just can't load stuff in between songs over still connected VIF/MF. Also, maybe related, MIR doesn't unload when closing a project; only when loading the next one.

Btw, this has been so for quite some time here (more than a year) and is the same with Logic 9 and 10. Could there be a setting I'm overlooking (and please don't tell me it's Preserve...)

From what I understand, all conections to servers should close when I close a project (manual pg 32: "As soon as you close your song, the preserved instance will be disconnected"). However, they don't. So when I close a project and want to load a new one that uses a different MFrame or VIFs, I have to first force disconnect all (or load an empty dummy project to kill the connections). What am I doing wrong? Or how can I fix this?

GRHM,

are the tracks with the delay problem AUX tracks? Meaning you're using the Logic feature that one can play midi on an AUX track. If so, there's the problem. It's what I consider a bug in Logic (both LP9 and LPX). My workaround: 1 extra bar preroll (if needed) and play a random note first to defuse the delay for the real notes; delete that note after recording. lame, yes, but better than having to use Midi + AUX tracks.

Paul,

thanks for clarifying. and congratulations on yet another great product!

If I understand MirX correctly, I can see 2 possible things to be gained from it for MIR Pro users: offloading some reverb processing to slave computers when necessary (I have not yet encountered this need, but you never know...). And being able to use MIR on my mobile setup (presuming that it uses less CPU than MirPro).

Is this assessment correct?

I have the same problem. Mixing in the VIframes is possible, yes, but very inconvenient as I'd have to save each VIFrame with each project (I run VEPro decoupled for snappier response in Logic) and would have to reload them for each new project, which defeats the great preserve feature.

I submitted a feature request to Apple with a link to this thread and Martin's explanation. Those who experience this problem, please do the same: http://www.apple.com/feedback/logicpro.html

Aerosmith - yay!

seriously, here are some ideas for new features that have very promising potential to create plenty of satisfying bugs:

allow to deactivate the unused outputs of matrices and other new features when not in use so that ViPro becomes more CPU friendly (that is of course assuming that these features use CPU when not in use; and not talking about reverb and EQ obviously)

allow to change the keys for just intonation with midi controllers

return the controller curves to the settings you had in VI = allow them to be flatter

add optional randomization to round robins and humanize functions

increase the max allowed sample start offset to 100ms

and - please - allow loading of ViFrames directly from the Finder (as opposed to always first having to create an empty one)

p.pan, I understand. Maybe I'll do a setup video in the future - got no time (or muse) for it now. What specifically would you like to know?

the basic principle of my setup is simple: Main Mac and a few slaves that hold the samples in RAM, all connected via Ethernet and running VEPro which in turn loads VSL or Kontakt libraries. Everything I will change frequently (synths, Omnisphere, drums, unusual sounds...) gets loaded on a project basis on my main MAC. Whether the slaves are Mac or PC is a matter of personal preference and how important budget is. I'm sure one can build super quiet and powerful PC slaves that cost less than Macs, but the no-hassle-factor is big for me, so I chose what I'm familiar with.

One caveat I experienced with VEPro: If I have all 4 slaves connected, I can only have 18 VI-Frames connected; the more connected slaves, the fewer VI-Frames (Logic freezes if I try to connect more). I am interested in other users' experiences in this regard.

 The main thing is that every slave has a unique number (eg: slave 1 = 110; slave 2 = 111...) and that it's between 1 and 254 (if I remember correctly). check this thread - cm's first reply - he's in the know: http://community.vsl.co.at/forums/t/18601.aspx

I have Wifi turned off because I don't use it on the Minis.

David,

yes, I operate the Mini's via Ethernet only. I use a Netgear GS605 switch - I was told a switch handles network traffic better than a router. They don't cost much - about $40.

network setting: IPv4:I use DHCP with manual address. IP address: 192.168.0.xxx - last three is the number you assign to that Mac. This way I can conect the Minis to the internet if needed. You can also set IPv4 to Manually. I do this for my 2nd ethernet port on my main Mac, which is connected directly (no Ethernet switch) to another Mac Pro. For this I set IP to 192.168.2.xxx - I don't remember why, but it works. So afaik: use a switch and set the IPs manually, than your main Mac will find the slaves and especially when you reopen existing projects with VEP, it will know where to connect to.

I control the slaves with Apple Remote Desktop, but you can also use the built in screen sharing (activate it in Sys prefs-Sharing), which costs you nothing extra and also works great.  I only had to connect a monitor for initial setup. The MIni now shows me a 1680x1050 display via Ethernet, which works beautifully for editing VEP setups.

The setup is currently very stable; offline bouncing works great. etc: it's awesome!

The System takes about 700mb of RAM - seems to be the same on all my Macs. I never bothered messing with the Finder - it uses close to 0% CPU and about 30mb RAM on my slaves - not worth the hassle.

sorry, no video. VEP is the only 3rd party software running, so once the system software has started the Mac OS does not need to access the HD much on a sample slave.

I'm not going to try to make you a believer. Just reporting what works for me. Ultimately you'd have to try this with your own setup, DAW and working style to see if it would work for you. Minis work for me and when I did the math, they were cheaper than buying a used MacPro with similar CPU and RAM with the benefit of being smaller and producing less heat. There are similar PCs out there btw (don't ask me for links, I did not keep them).

Paul,

I saw that setting and it's brilliant. What I am arguing for is to add an option for randomization that is not prohibited by the optimization function. Current alternates work great for lines and shorter repetition patterns. With longer patterns the sound pattern of the repetitions starts to emerge (even with the perf rep patches) so that the pattern of the repetition starts to sound mechanical.  I'd trade having control of exactly how the samples are triggered, so that they sound the same every time I play the song, for a randomization option that I can use for such long repetitive patterns. The same would apply to the new humanize function.

    It's just an idea to make the awesomeness even awesomer imo.

Yes, I use a MacPro 8 core 3GHZ as the Master. The Minis are connected via Vienna Ensemble Pro, which is absolutely totally awesome! So, the only hardware you'd need is an Ethernet switch. I installed 7200RPM HDs in the mInis and run all samples from the internal HD.

      Examples of what I run on them: One currently runs LASS - bank presets of 10 main articulations, all sections; Kontakt set to 30kb DFD preload, which amounts to 4.3GB RAM + about 2GB for the Kontakt banks. At a buffer of 256 the CPU will max out when I use all divisi sections simultaneously in a busy section, but it works fine if I use higher buffers.

    The other Mini currently holds Chamber Strings + Appassionata Strings + Vienna Saxophones - I use custom presets, but they are approximately as big as the Level 2 presets. Also uses about 6GB of RAM; similar situation with the CPUs.

VSL Chamber strings + LASS sections is a wonderful combination btw.

Cyril,

I use two Mac Minis (late 2009 model 2.26GHZ with 8GB RAM each) connected vie VEP - cost me about $1100 each. They work great,a re super small and super quiet. Their CPU will max out if I run everything I got loaded (eg VI orchestral strings + Appassionata Strings) with performance legato patches, but that has not caused any problems in real live situations so far.

please excuse the OT - I know the question was about PC slaves.

a way to accomplish something similar to  (and almost as effective as) a master pan would be the ability to select multiple channel strips in the slot mixer and have the changes made to one powerpan translate to all others (maybe when holding a modifier key like you have it for volume in the VEP mixer) - hey, that would be cool for VEP as well.

This goes slightly off topic:

Hi Herb, 

 Are you saying that optimize currently checks which alternations have been used and discards the rest? That's very effective, but couldn't the optimize function be adjusted such that when a note is being triggered at a certain velocity, that velocity layer is preserved with all its alternation samples. I would gladly trade the minor loss in optimization (I hardly ever use it anyway) for the benefit of having randomization of the alternation samples as it can be done in Kontakt. I'm no programmer of course so forgive me if I don't consider code intricacies or other considerations.

I am just trying ViPro with the demo version. Lots of awesome stuff!

 The adjustable controller curves have changed - I can't make them as flat at the top as they are in VI. I liked that to add just a little bit of Expression smoothing to linear velocity cross fade settings (both on Ctrl11). It is not only the display that changed, the curves sound different. With VIPro, the flattest setting, which is displayed as 0.20, results in a velocity reduction that is too strong. Is this a deliberate change? Is it possible to match the previous curves?

looks great, lots of good ideas.

What I'd love to see is a way to assign different Perform settings to patches. A choice between 2 options would be great, so I can have long articulations (like legato) use Expression (or any other controller) for velocity switching and volume and use Note On Velocity for short articulations (with Velocity X-Fade disabled). This makes playing realistic performances much easier. I currently have 2 VIs for each instrument - one for long articulations, the other for short ones - just so I can have 2 sets of performance controls. Having this option within VI would simplify this setup greatly.

I had this happen too. If I remember correctly it was because the Beta versions I had installed were not identical.

I know I'll be the outsider on this thread but I'm quite happy with my recent MacMini purchases. I bought the basic model (2.26gHz), updated the internal HD to 7.k RPM and the RAM to 8 GB - costs about $1130. The actual RAM for sampling is then closer to 7GB, but that's still pretty good I think. They are super quiet, super small and from what I read very energy efficient. I run Snow Leopard, VEPro with VSL and Kontakt on them without any issues. So I think if you don't need one big machine but can go with "smaller" (7GB RAM is still pretty sizable) increments to your setup, this is a viable option.

gblume,

thanks for posting this and creating the examples on your webpage. To me the difference is glaringly obvious and I am strongly for implementing this system in VSL. The Hermode Tuning examples sound much more natural. For me it is an immediate emotional response, not an analytical one. It's like the HMT examples take a certain pressure of the listening experience and it makes it easy to just relax and get absorbed into the music - which is one thing that is often a challenge with music created with samples. I think this is an important step towards making sample based music more believable and emotionally compelling. I was aware of Hermode Tuning in Logic before and was frustrated that I could not use it fully.

    I dread having to go back to my TMP samples now :-(

please VSL, implement this. 

If licensing Hermode Tuning would cost serious extra money, I'd be happy to pay for an upgrade that includes it.

I bought a new MacMini and installed the RAM upgrade ;-) (bought from macsales.com)

The specs are outdated. The latest Mac Mini firmware, which comes with the new ones, supports 8GB RAM. I currently load about 6.5GB of samples into RAM. runs great.

I just bought a Mini (2.26GHz model), loaded it with 8GB RAM (yes, the new ones handle 8GB!) + a 7.2k hard drive. I'm running VSL chamber srings and half of LASS in VEPro on it without problems so far. It cost me about $1400 - it's small, super quiet and no hassle for me, which is why I chose it over a PC slave.