Vienna Symphonic Library Forum
Forum Statistics

180,802 users have contributed to 42,141 threads and 254,364 posts.

In the past 24 hours, we have 1 new thread(s), 7 new post(s) and 73 new user(s).

  • VE MIDI still not working when loading 3+ GB of samples (Update w/ details for VSL team)

    last edited
    last edited

    So here's the problem: I'm running a single slave machine under XP64 with 8 GB of RAM, an ASUS P5B-E motherboard, Core2 Duo E4500, and an HDSP 9652 card.  When I load more than 3 GB of samples into the VE, the VE stops responding to MIDI.  I've tried MIDI with both hardware and MOL and neither works.  Actually, it worked once, briefly, then crashed.  The same thing happened when I had an E-MU 1212m card installed in the machine.  So, my prior assumption that the problem was the 32-bit E-Mu Patchmix software is not holding up (bummer - that cost me a lot of money...).  It turns out that the HDSP mixer is 32-bit also, but I know folks are successfully using the HDSP 9652 in slaves under XP64.  They're also running MOL, so the problem must be elsewhere.  The MOL MIDI monitor shows that the MIDI is coming in to the slave computer but the VE shows no evidence.

    I can load multiple instances of the VE and get past the 3 GB limit but then I get constant crackles and pops, even with only one instrument playing. 

    All of my software/drivers/etc. are the latest.  Xp64 has SP2 installed along with all of the updates that were listed for my machine.  The VE says "v 2.0 Build 2763 (64-bit)".

    Before I throw any more money at this problem, I'd like to know what systems are working with the SE under XP64 for FULL orchestrations from a SINGLE slave machine.  I'm completely at a loss for why the VE won't work for me...

    Increasingly frustrated, but still optimistic!

    rgames

    UPDATE 11/28:

    I started from scratch and re-installed XP64, chipset drivers, ethernet drivers, SP2, RME drivers, syncrosoft drivers, MIDI Over Lan CP3, and VE 2.0.2.  I started with my soundcard buffer set to maximum (4096 samples) just to remove any problems with CPU usage.  So here's what I did:

    • Load single instance of VE, then load full orchestral template with 5.3 GB of samples over MIDI Ports 1, 2, 3, and 4

    @Another User said:

    No MIDI ports work, all audio works, MOL shows all MIDI ports functional

    • Etc....

  • When running multiple standalones how are you routing audio? Are you using the "loopback" feature of the RME software?  


  • The RME card is multi-client, so I just make certain that the two instances are using different outputs.  So, for example, instance 1 uses ADAT 1/2 and ADAT 3/4 and instance 2 uses ADAT 5/6 and ADAT 7/8.

    rgames


  •  I guess I'm wondering how you route the audio back into your DAW. You don't seem to mention what software your are using as a DAW. 


  • I guess I didn't say in the original post - I'm trying to run the SE on a separate, single slave machine.  The DAW is Cubase, but the VE is running on a separate machine.  So I'm not running the DAW on the same machine as the VE.  It's all by itself on one beastly computer...

    But the audio works fine - I can click on the keyboard in the VI and hear everything regardless of how much RAM I use.  And MIDI works as long as I stay under 3 GB of samples.

    Thanks,

    rgames


  • Update - I started trying random things to see what happens and it turns out that I can get one MIDI port to work if I hit the "Stop Engine" button and then hit the "Start Engine" button.  Unfortunately, I need four.

    Furthermore, if I repeatedly stop and start, I switch between MIDI Port 1 and MIDI Port 3 working correctly.  MIDI Port 2 and 4 never work and 1 and 3 never work at the same time...

    So the VE is definitely doing some screwey things - doesn't this seem to indicate that the problem is with the VE?

    Looking forward to figuring this out,

    rgames


  • You might want to contact RME to ask why having two instances of VE running should cause digital artifacts.  


  • Good advice - I hadn't pursued the mutliple-instance option to any large extent but if it works, I'll take it!  I've got a question on the RME forums and an e-mail to their support.

    In the meantime, though, is there anybody actually running the full SE on a slave machine under XP64?  Multiple VE's or not - I just want to know what's working...

     Thanks,

    rgames


  • last edited
    last edited

    @rgames said:

    Good advice - I hadn't pursued the mutliple-instance option to any large extent but if it works, I'll take it!  I've got a question on the RME forums and an e-mail to their support.

    In the meantime, though, is there anybody actually running the full SE on a slave machine under XP64?  Multiple VE's or not - I just want to know what's working...

     Thanks,

    rgames

    I'm not running SE, but I am running the Cube. However, as I use FXT this information is probably not of much use to you. [:(] DG

  • Sorry to hear you're still stuck Richard, depsite threads on here and ViC.  :( This seems daft in the extreme, it should be working. I'm not sure I can see why it works for me and not for you (XPP x64, RME HDSP cards, VE) but the only difference I can think of is I'm running the full cube, not SE...... dunno..... feel free to drop me a line if you want to have another go at thrashing this out.


  • Yeah - I got it to work ONCE with the E-Mu card and ONCE with the RME card but I have no idea why.  It seems that it does not work 99% of the time, but every now and then, the stars align...  Couldn't that indicate some type of memory addressing problem - sometimes certain sections get overwritten but every now and then the OS moves that chunk of code/data somewhere else?  Dunno - my code writing chops are pretty basic, so I'm just shooting in the dark here.

    The main difference I recall between our setups was that you were using only two MIDI ports per slave machine - is that correct?  I'm using four.  And I'm using the SE, but that shouldn't make a difference.  So our memory use is different, and if it is a memory problem, it could be the difference that causes it not to work for me.  A lame explanation, but the only one I have right now.

    Quick question, though - does your HDSP Mixer show up as a 32-bit process under XP64?  i.e. does it show *32 next to the process name in the task manager?

    I'm doing the format/rebuild process once again today and adding drivers, software, etc. one-by-one to see what effect they have.

    Thanks,

    rgames


  • Update - I did a complete re-install and added the details to the original post.

    Basically, nothing changed.  Sometimes the MIDI works and sometimes it doesn't.  I can stop and start the engine, change sound card settings, etc. and that seems to do something sometimes but, overall, it's just sort of random.  The audio always works, though...

    rgames


  • OK - another update.  I bought a completely new motherboard to match the machine that VSL is using (ASUS P5W).

    Still no change, though.

    So now I've tried three different sound cards and two different motherboards.  I think it's safe to say it's not a hardware problem...

    I'm really starting to wonder if it's not related to the Syncrosoft licensing software.  It is, after all 32 bit (so it might be subject to a 3 GB limit), and the only remaining difference between my machine and those of you who have responded is that we don't have the same licenses - I only have the SE.

    I'd sill like to find out if anybody is actually running ONLY THE SE on a slave machine w/ 8 GB of RAM under XP64.  I know it's not supposed to matter from the standpoint of the VE, but I'm becoming more suspicious of the licenses...

    Anybody have thoughts on what to try next?

    Thanks,

    rgames


  • Hi rgames,

    I´m also (like Daryl) running the whole Cube on an external PC with 8 GB RAM. BUT I´m running it under Vista64 at the moment.

    I think I have a similar setup like you:

    Intel Xeon 2,66 (3070)
    Intel S3000 AH mother board
    8 GB Corsair RAM
    RME 9652
    VE v2.0 build 2781

    I´m not running MoL but using the midi ins from the RME card. The Adat outs are connected to my RME 648 on my main machine, which still is a G5 2,5. I work with this setup for about 2 months and never had any midi problems.
    Right at the moment I have loaded 3,485 GB into ViennaEnsemble (read out from taskmanager).

    So maybe it is easier to try once just with old fashioned midi cables instead of Mol (sorry if this was already mentioned somewhere, I didn´t read everything).

    Or maybe you can get a version of Vist64 Business or Ultimate to try it on a separate drive?

    goog luck
    christian

  • Hi Christian,

    Yes, I have tried the hardware MIDI and it does not work, either.  Well, actually, it works every now and then, just like MOL.  I have two 4x4 hardware MIDI boxes on order but I'm not hopeful that it will solve the problem.

    How are you able to run a full template on just 32 MIDI channels?  I need at least four ports to cover all the instruments I'm using...

    I'm really starting to believe that it has something to do with the licenses - the only responses I've seen so far have been from people running the Cube.  I still have not heard from anyone who is running ONLY THE SE this way...

    Thanks,

    rgames


  • last edited
    last edited

    @rgames said:

    Update - I did a complete re-install and added the details to the original post.

    Basically, nothing changed.  Sometimes the MIDI works and sometimes it doesn't.  I can stop and start the engine, change sound card settings, etc. and that seems to do something sometimes but, overall, it's just sort of random.  The audio always works, though...

    rgames

    rgames,

    My other two posts explain a bit more. But I'm now having similar behavior. I was able to get the VE to respond to Midi once I pressed teh Stop Engine Key, but only the one time. It hasn't worked since. Funny thing is, as of now, my other slave (which is an exact duplicate except for the motherboard) is not behaving this way. I've loaded build 2781 on it and 6.5 gigs of RAM (all Chamber and Solo Strings) and it seems to be ok so far.

    By the way, my slave machines are a bit different from yours, so the difference may help in eleminating some guesses:

    Gigabyte motherboard

    8 GB RAM

    Frontier Wavecenter PCI Card (2 midi inputs/1 ADAT out/1 ADAT in)

    Midisport 4x4 midi interface

    Sonar 7

    Mahlon


  • last edited
    last edited

    @Mahlon said:

    By the way, my slave machines are a bit different from yours, so the difference may help in eleminating some guesses:

    Gigabyte motherboard

    8 GB RAM

    Frontier Wavecenter PCI Card (2 midi inputs/1 ADAT out/1 ADAT in)

    Midisport 4x4 midi interface

    Sonar 7

    Mahlon

    Not to sidetrack your post, but what drivers are you using for the Midisport? I can't find any 64bit ones.

    DG

  • Mahlon - I'm sorry to hear that you're having the same problem but I am at least relieved to hear of someone else repeating it :).

    I don't think the hardware has anything to do with it (I've tried a lot of different hardware and it all behaves the same with the same set of samples).

    So you have one machine that works: it has the solo/chamber strings.  What are you running on the other machine that is having MIDI problems?  And here's a test you might do: swap the samples between the two machines and see if the problem follows the machine or the samples.

    Oh yeah - how many MIDI ports are you using on the two machines?  Is the functional machine using fewer MIDI ports? 

    Hopefully VSL will chime in with some suggestions soon...!!! 

    rgames


  • DG,

    The midisport 4x4 is on the DAW main computer, which is running XP32 and Sonar. The (2) slaves are running XP64 with the Wavecenter PCI cards receiving the midi in. Now, there might be a problem with the Wavecenter 64 bit drivers and midi, not sure. I couldn't work on it anymore last night, but this evening I will try some scenarios to see if either I can reproduce what happened or chalk it up to user error somewhere.

    rgames,

    Each slave computer has 2 midi ports which come in through the Wavecenter PCI cards. Both ports on each computer are enabled in VE, but I'm only using 1 port on each computer (Port 1 IN) at the moment.

    On the machine (Computer VSL01) which has worked correctly so far:

    App Strings ,L1 and (L2 in demo mode)

    Chamber Strings, L1 and (L2 in demo mode)

    Solo Strings, L1 and L2

    Harp L1, and L2

    On the machine (Computer VSL02) which gave me the glitch last night:

     Brass 1, L1 and L2

    Woodwinds 1, L1 and L2

    I will try your suggestion of switching the sounds tonight if I have time. So far, it's only happened with the one VE Project file which runs about 6.5 gigs on Computer VSL02. I tried reloading the matrices individually to make a new project file, and as of bedtime last night, MIDI was working on that project file. But I didn't have time to reload that file in one 'go' -- but I will tonight.

    Thanks,
    Mahlon


  • Well, rgames,

    After several hours, I'm making some headway in finding at least a little consistency. VE definitely stops receiving midi. Whenever I load more than 13 channels in a standalone VE and save this as a project, the project won't re-open with midi working correctly. In fact VE is not recognizingn any midi at all when I open a project that has 14 channels with a matrix in each. If I have 13 channels loaded and save this as a project, it re-opens and works fine with midi. This has been consistent and is why my other slave hasn't displayed the problem -- because it's only had 10 channels loaded at most. IWhen I loaded 14 channels into it, the midi stopped working upon re-opening the saved project. Incidentally, if I don't save the project file, I can load up to 16 channels and everything works ok. It's just once I save the file, and re-open it that the midi stops. So, I guess I will contact tech support with this info and see if there something weird going on.

     Below (if you want to read it) is a bit of stream of consiousness writing I did while experimenting. It may help explain in more detail about my setup and what I did. I'm going to bed now.... my hard drives have had a time of it tonight...and so have I.

    Mahlon

    1. I have a 5.7 GB VE project file with 15 midi channels, one matrix loaded into each. This project file will load into VE but midi doesn't seem to be working. Midi is leaving Sonar 7 and going out the DAW's midisport 4x4 and then to the slave machine. But VE does not pick up any midi! (incidentally, I can play the patches from the VE keyboard by mouse-clicking, so I know my audio is OK)

    2. However, if instead, I start from a clean slate (nothing loaded and full RAM in VE), I can individually load the same matrices as above into VE by hand, and the midi DOES work correctly. It's just once I save it as a project midi won't work. But this hasn't been true for all projects -- smaller projects with less midi channels load ok.

    3. So, to experiment, I've decided to load the individual matrices one at a time, AND SAVE a new project file after each additional matrix. In other words, I load my matrix Flute 01, then save the project as Project Flute 01, reboot the computer, start VE, then reload the saved Project Flute 01. If all goes well, I add another matrix (say, Oboe 01), save both matrices as a new project (Project Flute 01 Oboe 01), reboot, start VE and reload Project Flute 01 Oboe 01, then add another matrix, save, etc. etc.

    ...you can imagine how long this is taking once several instruments are loaded..... keep imagining....I'm loading up midi channel 13 right now, and so far so good (even on such an unlucky number)....

    Ok. It broke on midi channel 14. After saving the 14th matrix, and then reloading that Project (with 14 midi channels), the midi stopped responding. Now, I'm thinking that why my other machine may be working fine is because I have only 10 midi channels loaded. hmmm.......just confirmed, the other slave loses midi connectivity at channel 14 loaded.....