Stuck MIDI notes...arrrrrgggggghhhhhh!

[Tobor]

Another random thought....

The latest version of Band in a Box has some fairly complex sequences and a lot of MIDI activity going on when you throw in the soloists, etc. Going back and forth to a PSR 3000 through the MTPAV it works like a champ- no stuck notes, no MIDI timing issues.

It's when I dump sequences from either of these units/programs and start overdubbing MIDI tracks in DP that the (lack of) fun begins....

Tobor

[Shooshie]

I'm coming into this discussion a little late, but maybe this will offer another "data point," as qo described above. When I installed MOTU's latest Firewire/Universal update about a month ago (I don't remember exactly when this happened, as the holiday season seems to have wiped my memory clean), I found that Clockworks no longer worked properly, and while my 424 card worked, CueMix no longer could find a "connected audio interface." I use a 1296 and 2408 mkII, which are legacy devices, and I assumed that MOTU was moving on, distancing itself from the legacy devices. I also use an 896, and it worked fine, but CueMix has not been an option on it, anyway, since switching to OSX. Clockworks, in that situation, would no longer honor my connections and routing, and I could not use the graphic interface to connect. I never got as far as testing the timing.

So, I just did an uninstall and put the old drivers back in. Now everything works again. While up to now I have not had the timing problem that everyone else is reporting, I have just updated OSX to 10.4.4, so we'll see if that introduces any new problems. I did not mean to install it, but accidentally left it selected in Software Update, only to come back and find it installed. Oooops. But I was ok in 10.4.3, so maybe .4 won't be a problem.

Anyway, I wanted to report the driver problem with this latest Universal Update from MOTU. Scared the crap out of me. I thought I was going to have to do another complete system re-install, but downgrading back to the previous set of drivers got all my functionality back.

Shooshie


____________________________
Dual 1GHz G4, OSX 10.4.4, 1296, 896, 2408 mkII, MTP-AV USB + MTP AV Serial, Kurzweil 2600 controller, assorted rack gear, Mackie 1402 VLZ Pro, Mach Five, Ivory, Akoustik Piano, MOTU Symphonic Instrument, Waves Platinum, Wireless Apple Bluetooth Keyboard, Wireless Kensington Turbo Mouse Pro, Tannoy Monitors, Conrad Johnson power amp, GE 60 Watt light bulb...

[markwayne]

Well, I felt I should chime back in as I just experienced DP MIDI nonsense. No stuck notes that I could hear (although I was working with some percussion sounds so I really would not have known if they had stuck). What I got was something I haven't had in months. It sounds like the MIDI buffer filled up and DP just stopped sending MIDI altogether. Then, when I stop the sequence, DP spits out all the notes at once. Network cable unplugged as per my normal setup. Really stupid problem to have in this day and age. I restarted and it seems to have gone away. I'm starting to use Sonar on a PC more and more for sequencing and audio work. It's a sad thing for me to say, but I've got work to do.

As to the post about the Atari: yes, I had one and yes, the MIDI timing was rock solid. That thing was as tight as a Linn MPC60. (The tightest, hardware MIDI box I've ever worked with. Hell, my Mac plus with a Studio 3 and Performer 3.1 ran rings around DP 4.6 on my G4. Nothing in my rack has changed beyond the new interface. I do a huge amount of MIDI sequencing with rarely more than a few of tracks of audio.

Oh, I should add that this was after coming back to the studio after a four hour break. Everything had been working fine up until the break.

oh well,
Wayne

[Tobor]

Ah yes, Wayne.... the dreaded machine gun syndrome. For me this happened every time I would hit a locate button on my Roland VS when DP was slaved via midiclock sync. If I would play back the sequence from the beginning, no problem. Like you, I use a lot of MIDI tracks and use my Roland for the audio. This system has worked flawlessly over the years with every Roland VS machine and every version of DP- until version 4.X.

For me DP4 has been virtually useless. It's like having a lot of great toppings on burnt toast- looks great but you can't eat it. If I can't depend on the MIDI, I can't get into all the exciting new features in the audio department.

The one good thing about using my old iMac is that I can boot into DP3.1.1 to to get MIDI work done without feeling I'm walking around land mines. But this computer is really too slow to run Reason 3.0, for example, so I've been itching to upgrade a new Mac. I have a sinking feeling, however, that even with a super fast intel machine and DP 5.0 these MIDI problems haven't been addressed.

It's sad because for me the bottom line is that DP isn't fun anymore. It used to be the first thing I'd turn on and I was always playing around with some little silly sequence or another. Now it's the last thing I turn on after I get a standard MIDI file from another program. I've remained loyal to MOTU- I just keep waiting and hoping that they'll FIX THE MIDI.

I hate to post in the new DP5.0 thread and be a wet blanket, but I think that a large percentage of their original core loyalists who work mainly in MIDI are going to start looking elsewhere if these basic stability issues aren't dealt with.

And unfortunately the only thing we can do is to squeak the wheel.

Tobor

[Shooshie]

Ah yes, Wayne.... the dreaded machine gun syndrome. For me this happened every time I would hit a locate button on my Roland VS when DP was slaved via midiclock sync. If I would play back the sequence from the beginning, no problem. Like you, I use a lot of MIDI tracks and use my Roland for the audio. This system has worked flawlessly over the years with every Roland VS machine and every version of DP- until version 4.X.

For me DP4 has been virtually useless. It's like having a lot of great toppings on burnt toast- looks great but you can't eat it. If I can't depend on the MIDI, I can't get into all the exciting new features in the audio department.

The one good thing about using my old iMac is that I can boot into DP3.1.1 to to get MIDI work done without feeling I'm walking around land mines. But this computer is really too slow to run Reason 3.0, for example, so I've been itching to upgrade a new Mac. I have a sinking feeling, however, that even with a super fast intel machine and DP 5.0 these MIDI problems haven't been addressed.

It's sad because for me the bottom line is that DP isn't fun anymore. It used to be the first thing I'd turn on and I was always playing around with some little silly sequence or another. Now it's the last thing I turn on after I get a standard MIDI file from another program. I've remained loyal to MOTU- I just keep waiting and hoping that they'll FIX THE MIDI.

I hate to post in the new DP5.0 thread and be a wet blanket, but I think that a large percentage of their original core loyalists who work mainly in MIDI are going to start looking elsewhere if these basic stability issues aren't dealt with.

And unfortunately the only thing we can do is to squeak the wheel.

Tobor

Tap tempo doesn't work right, and hasn't worked consistently since DP4.0. I can sometimes get it to work if I play with it long enough, but only maybe for one take. Then it still doesn't work anymore. I used to use tap tempo practically every day in OS9. I keep waiting for them to fix it, but it's like they aren't even trying. Do they think people don't do MIDI anymore? I'm a MIDI arranger who uses audio recording, not a recording engineer who uses MIDI. DP was always my first choice because they led the pack regarding MIDI. They owned MIDI, as far as I was concerned. Sad to see that go.

Shooshie

[Tobor]

Tap tempo doesn't work right, and hasn't worked consistently since DP4.0. I can sometimes get it to work if I play with it long enough, but only maybe for one take. Then it still doesn't work anymore. I used to use tap tempo practically every day in OS9. I keep waiting for them to fix it, but it's like they aren't even trying. Do they think people don't do MIDI anymore? I'm a MIDI arranger who uses audio recording, not a recording engineer who uses MIDI. DP was always my first choice because they led the pack regarding MIDI. They owned MIDI, as far as I was concerned. Sad to see that go.

Shooshie

Well put. After giving up on midiclock sync, I started using MTC sync (not half as useful because DP doesn't 'snap' to the locate point unless you hit play), and even then I would sometimes have to hit stop and start three times to get everything playing back correctly.

Are we the 'lost generation' because we still use MIDI arranging as opposed to loops?

If I buy a new Intel WonderMac and DP 5.0, do I still have to boot up my lowly 800mz G4 in 9.2.2 to order to actually play music instead of fight it?

Tobor

[markwayne]

Well, I can confirm that this is a DP issue on my rig. Things continued to get worse and worse here to the point where I would start to hear flamming and general MIDI timing weirdness after just about a minute into recording on any MIDI track. I've tried everything possible from a hardware standpoint. In desparation I started up Live to try and just get back to work. While Live is far less elegant than DP, it has been glitch free. It still doesn't feel as solid as pre-USB MIDI, but it works.

Saddly Live's audio quality isn't quite there so I'm probably going to work out a way to rewire Live to DP and see if that can get me to the end of this project. Then I'm going to have to make a decision. I'm thinking about moving my old 8600 back into the studio rig and going back to serial MIDI and OS9 at this point. It's more than a little frustrating. I have no other serious problems with DP besides MIDI hell. DP has not quit or crashed on me in over a month of heavy, daily use. I can stack up to a couple of dozen audio tracks with plugins without any problems. However, as soon as I add even a single MIDI track, things start to go downhill.

On a happy note, I'm actually seeing an increase in my productivity after only two days using Live as my main sequencer. Wow! I have not worked this quickly since SVP on a Mac 8600.

Wayne

[misentropic]

After reading everything here and troubleshooting with posted suggestions an MOTU's (Set up your USB like this: Keyboard to the computer - mouse to keyboard - MIDI Express 128 to keyboard. No other USB should be connected for this test. Do you have the same problems with this USB configuration?) which actually made it worse the only reliable way I can control this is by making sure the "wait for MIDI note" is disabled. If it is on the MIDI notes stick like dust on a candy apple - off and it seems tolerable.

Still: $$$ for Tiger Ugrade + $$$ for DP upgrade = stuck MIDI notes and a sequencer that works worse than it did 6 moths ago = #$%#@#$%!!!!!

Apple and MOTU both better get their pencils sharpened for the next go around of upgrades and or updates. This is bloody ridiculous. I would have never guessed in my wildest nightmares that there would be this many people with this profound a problem. I'm shock and pissed.

Thad

[qo]

The first MIDI spec was published in October 1983. Let's see, that's like, oh, 22 years and change to get basic stuff right. If it ain't right by now, it ain't ever gonna be right (joking).

Yeah, I agree with others that it WAS right for a while. I think what killed it from a technical perspective was USB/OSX, and probably sync compromises that had to be made when VI's came along (e.g. trying to sync hardware synths, which are reasonably deterministic, with software instruments, which are not very deterministic, perhaps led to decisions being made that were to the detriment of hardware synths since, at some point, all this MIDI data needs to be multiplexed/demultiplexed coherently in DP).

OSX is more a unixy timeshared system than OS9 and so things that require strict timing are going to suffer, despite thread prioritization, since when you have several things that need equal priority e.g. all the threads that deal with audio/MIDI i/o and processing, the best the OS can do it dole out processor time fairly among them. In OS9, you also had multiple processes that required realtime, but the way things are doled out is different and leaner. And, the requirements on the processor are greater in OSX because more background processes exist. I'm guessing that this part of the problem will get better as our machines get faster (multiple processors to handle interrupts from our audio and MIDI interfaces, and generally faster processors at that, combined with better multiprocessing support in the kernel).

USB doesn't have isocronous support (not sure about USB 2.0, but the MTP/AV isn't USB 2.0 anyway). A dedicated serial port, while slower, will trounce USB since it's possible to give it one interrupt for one dedicated device (as opposed to it sharing an interrupt, though this can be done as well). Who knows how Apple implemented USB internally? My guess is that, even though there are multiple ports, they are sharing an interrupt). It might be worth buying a PCI USB card and plug your interface into that to see if that improves timing. I might try that today...

qo

[Tobor]

Thanks, qo, for the info and insight.

It's very frustrating when you know there's a problem and whenever you bring it up people get defensive or treat you like you're an anti-DP crackpot or novice or what have you (check out the 'DP5.0 is AWSOME' (sic) thread

I've been handcuffed by this issue for two years now, looking back. I don't want to give up on DP as my main sequencer. Recently I had DP slaved to my VS2480 via MTC overdubbing one note horn lines on my Triton. Before listening back to them I would have to shift the track forward 20 ticks! Makes working slow, tedious, and very unmusical.

You bring up good points about hardware/OSX/USB. Indeed it seems like the introduction of midisync or MTC into the sequencing proceedings seems to be the rally-killer. It would be great if a hardware or software workaround could be found to keep the train on the track. Otherwise it looks like we're all going to be forced into 'the box'.

At any rate it seems important to keep this subject current and find a cure or at least a viable workaround. Either that or come flat out and tell us 'DP's MIDI functions no longer work properly in OSX when synced to certain hardware. Deal with it.'


Tobor

[markwayne]

Yeah, I agree with others that it WAS right for a while. I think what killed it from a technical perspective was USB/OSX

Well, I would say MIDI has been right pretty much from the start. I lived by MIDI for over twelve years on the road doing live shows with everything from a Commodore 64, Atari ST and ending with a Mac SE30 around the early 90s. I never worried about these issues then. Back then we mostly worried about 3.5 floppies and drives going bad.

I'm with you on OSX and USB. I also noticed a huge drop in "tightness" with MIDI that started with USB and got worse with OSX. However, I must stress that I can use the same hardware and OSX with Live, Logic or Cubase and record hours of dense MIDI traffic. I've never had a glitch with MIDI in the above programs. I'm not saying it's the tight MIDI of my youth, but it works.

I don't know what DP is doing different with MIDI buffers and MTC and what not, but it ain't working in my case.

Wayne

[grimepoch]

I'm just trying to understand the differences in our setups to see where the problems come from. This is probably a stupid question, but why use midiclock sync on an external device? (I'm mainly trying to find out if I am doing something wrong even though I hear nothing wrong) I use 9 different HW synths with DP, all fed their own MIDI with my ExpressXT and their own audio channel on two 896s and an analog to ADAT converter.

I thought I was just using the internal timing of the MIDI express. I've never noticed notes being off from were I sequenced them, other than I believe the more notes you get going at the same time out 8 channels, you will get a tiny bit of delay just because of limitations of MIDI.

Is it the syncing to external devices with the MIDI? Does that give you more deterministic timing. Again, I apologize for such a basic question.

I also run my ExpressXT on one input into my box. Checking the system configuration, it's the only thing seen on that hub. Not sure if that helps.

[billf]

So far I've been MIDI problem free, but of late I haven't been really pushing any of my outboard gear either.

Anyway, I was reading in a developer forum, someone talking about iLok's on the USB ports causing problems with USB MIDI. I don't know if that is the case with DP or not. But since other programs that interface with DP use iLok (such as VI's) it could be one of the possible points of failure. So any test of USB MIDI probably should also be done with and without iLoks.

Is there anyone who is iLok free and having the MIDI issue?

[Tobor]

grimepoch:

For years I've used the Roland VS recorders for my audio (now the 2480). Farming out the audio let me get by on my 800 iMac as long as I have, and I just got used to editing quickly on it. I would slave DP and my hardware synths to the VS via midiclock. That way I could hit a locate point on my VS and DP would snap to the exact point via song position pointer for editing in DP. The VS thus functions as a hardware controller to some degree. At some point I might dump my DP tracks into the VS, record more vocals and sax, etc. If I needed to I could always call up the original DP sequence, re-edit or remix the sequence, and dump it back in and things would line up flawlessly.

In another scenario, I might compose a tune on a Yamaha PSR keyboards' internal sequencer, which only has midiclock. I could then slave DP to the PSR (or vice versa) via midiclock and record all the MIDI parts into DP so I could edit, overdub other keyboards, etc., and get the whole thing in shape. Then dump that sequence into the VS and away we go.

All that went out the window with DP4.X. After hitting a locate point on the VS, DP would now snap to the correct locate point, but then would wait a couple seconds, spit out 100 MIDI notes, and then lag a measure or two behind the correct location for the duration. Stuck notes.

Working with MOTU, we discovered a few months back that there are two different specs of midiclock. Some devices like those from Roland and Yamaha send out continuous midiclock whether they're in 'play' mode or not, and that was affecting DP adversely. MOTU said they would work on it and it would be included in the next update. Judging from the reappearance of the stuck note thread by those who are using 4.6, I am wondering if that issue is indeed fixed.

Gotta run, but at least here's some more specific info for your perusal and the reason why I depend on midiclock for many projects. Maybe the Beat Detective can come up with a solution.

Tobor

[Tobor]

I don't use iLoks.

Tobor