Click for SMOKING FAST SATA 6G SSDs!
Click for SMOKING FAST SATA 6G SSDs!


A Click on this Banner shows your site support to my Sponsors

Accelerate Your Mac!
Bring in the Noise
Mac Audio Column
by Thad Brown

4/24/2000

Musikmesse Coverage

Well, I made a brave return to the German speaking world. As some who read this may remember, I spent a rather horrifying year playing in Vienna a while back, leaving me with deep fears of all things German. So, it was not without trepidation that I went to the land of sausage and beer.

Another source of trepidation was how it came about. Through this web page I get to meet lots of interesting people. One of them is Stephan M. Sprenger, DSP guru at Prosoniq. Stephan and Prosoniq asked me to attend the show (largely on their dime) as a native speaker of English who is very familiar with their software. That way they could have someone on hand to demo Prosoniq stuff to people if none of the Prosoniq English speakers were available. It really didn't know exactly how to handle this, but decided to take the offer. I fully understand that people could consider everything I write from here on out as tainted, and I certainly wouldn't say that interpretation is completely off base. But, before you decide that is the case, read on for a bit.

In some ways this site is like any other site or magazine, in some ways it is not at all like other music information outlets. First, unlike many such places, Mike takes no advertising from audio companies. He also has no special relationship to any company or retailer that deals in music equipment of software. [For the record, I have never traveled to any show on anything but my own funds, although I did give Thad $500 to help a bit with his NAMM show travel expenses recently.-Mike] He never tells me what I have to say. In that way I'm more independent than most places. On the other hand, since this site generates no income for me, I am completely at the mercy or software and hardware vendors for review material. I would say that in general my relationship with the industry is if anything less cozy than average, maybe because I'm so small scale that nobody cares.

So in this context I get an offer basically to work for Prosoniq. I've gone on record as liking Prosoniq stuff for a while, so there's no surprise in that. I didn't take any profit from the trip, in fact, I paid for a few hundred dollars of it out of my pocket. I decided that it was worth it to get coverage of the show and the information that I would get from meeting developers at the show. I hope that makes sense to those who read this column. I'll include a similar disclaimer in the future when I review any Prosoniq software. Now, on to Musikmesse coverage proper . . .

The Death of MIDI

To paraphrase Nietzsche "MIDI is something that must be overcome." Long before I ever got into making music on my computer, MIDI was the standard that allowed all of our keyboards to talk to each other. According to the Great History of Music Making on Computers, getting MIDI together was no simple task. Manufacturers of keyboards understandably wanted to shift more units by leveraging the ability of their tools to work together. This was not great for the consumer, since, in essence, once you got a Brand X keyboard, it would only work with other keyboards from Brand X. Heroic efforts by a few individuals resulted in the open, published, MIDI spec which has continued largely unchanged now for something like twenty years.

Not surprisingly, those intervening twenty years have put some stress on that spec. MIDI itself helped generate the idea of a sequencer, which originally was a software tool that recorded and played back MIDI events. Still, it wasn't all that horribly difficult to get a serial port on a computer to talk MIDI, and MIDI is so low bandwidth that many many events can go out those serial ports to many many synths.

The sophistication of current sequencer, and the integration of digital audio into these sequencers, however, has begun to run up against serious problems with MIDI itself. My guess is that the fundamental problem is that MIDI comes from a strong "keyboard" mindset, it was designed originally so that one keyboard could control multiple synths. So, the model used was one that would allow playback to happen as quickly as possible. Once a key is struck, the point is to shoot that event out the port on the back of the synth as fast as possible so it can get to whatever synth it is supposed to control. Events are spit out the back as fast as possible in the order they are created. That's the only model possible for dealing with live performance, but it leaves a good bit to be desired as a way to work with data in a sequencer.

Once that MIDI data is in a sequencer, and needs to be played back, it works just as above. In essence, the computer becomes the keyboard player, sending data from the sequencer to the OS or MIDI driver, which is then responsible for spitting it out of a serial (or now, USB) port on the back of the computer to a MIDI interface of some kind, which then relays it as fast as possible to a synth or another computer or a time code device or whatever. It is here that MIDI is showing its problems.

In even a modest computer studio of today, that same sequencer can be playing back many tracks of 24 bit digital audio, running a software synth, and doing DSP on the audio tracks, while the OS is in charge of driving two or three monitors, a bunch of PCI cards, and doing all of its necessary system tasks to keep those type 11 errors at bay. The internal timing of every sequencer is so fine as to be considered perfect, but once it starts to send time sensitive MIDI data over a busy PCI bus, with a lot of processes running at the same time, and the whole system under heavy load, things can go awry. For various reasons, Macs have it easier than PCs, but it's an issue even for Macs. This is made worse by the fact that audio events can be synchronized with very high precision (i.e. the sample rate) even between multiple machines. The milliseconds of timing inaccuracy that have been part of sequencing since it began may not have seemed so horrible before, but now people increasingly think of that slop as a real problem.

Needless to say, software vendors have noticed this and are trying to exploit it. VST 2.0, for instance, is a great example of a workaround. By using the sequencers internal timing reference for software synths, VST 2.0 applications can legitimately claim that they have "sample accurate" MIDI timing. This is possible, of course, because they can avoid the whole issue of playback problems, since the synth is nearly running inside of the sequencer. Emagic claims similar performance for their synths and samplers integrated into Logic.

All this is well and good. It's innovative, it's getting around the problem, and it's in software. However, the real death of MIDI is somewhere else. Let's go back to our situation with hardware synths. Our sequencer is spitting out MIDI data that goes to a serial port and then a MIDI interface and then to a synth as fast as possible. Wouldn't it be great if we could put a time stamp on that MIDI data, send it out early, and tell everything outside of the computer "Hey guys, here's a note, play it back 134 ticks after beat 4 of bar 21" or better yet, "Hey guys, here's a note, play it back 112432 milliseconds after 00:00:000" and then the interface would deliver that note exactly where it was supposed to be? It would get around all of our problems, the data could be buffered somewhere along the line so small timing slop involved with the drivers and the computer could be negated.

Of course that would be fantastic, and for computers with the horses under the hood that we have now, it's not much of a task, but there's one problem, MIDI doesn't support it. MIDI is a "send it down the pipeline as fast as possible and don't look back" kind of spec. So, the smart folks at some of these companies are working on ways of "extending" what MIDI does. Remember those Microsoft "extensions" of Java? If you do, then you'll know that one of the problems with them was that they mysteriously seemed to work best (or only) with Msoft tools, thereby leveraging what was supposed to be an open spec for a particular product. That's also what's happening with this MIDI stuff. Mark of the Unicorn and Emagic started the fight by using distinct technologies to increase timing resolution when using hardware interfaces made by the sequencer company. So, if you are a Performer user, and you use a recent MOTU MIDI interface, you get better timing because of their MIDI Time Stamping software. Not to be undone, Steinberg announced similar capability in Cubase 5.0. Once again, it is supposed to dramatically increase MIDI timing accuracy, as long as you are using the forthcoming Steinberg MIDI interface.

MIDI has barely changed since it was first put together. This is mostly because it was designed very well by very smart people who left room for innovation in the future. But the stability of the spec also owes something to the simple fear that if it were opened up for discussion again, the whole house of cards might come down because of conflicting interests and agendas. One need only look to other pieces of the music software world, like audio card drivers, to see what can happen in such situations.

So, I say that MIDI is dying. Not in the sense that MIDI data will cease to be hurled about in our studios, in fact more and more of it is getting hurled about in my studio all the time, and to more and more different kinds of software and hardware destinations. No, the original goal of MIDI is what is gasping for air. Any of you read "The Hitchhiker's Guide to the Galaxy?" The author Douglass Adams had the little problem of how our earthling hero could possibly communicate with the different kinds of aliens he would encounter. It wouldn't do for the first fifteen years of his hitchhiking to be spent in language school. His solution was the comically named babel fish. Drop this fish in your ear, and you could understand any language in the Galaxy. MIDI has been the babel fish of music making for some time, and it's the babel fishness that is in increasingly danger.

I think another piece of this puzzle is that Opcode has disappeared. They made very good MIDI interfaces, and they were committed to maintaining OMS as a cross sequencer standard. The vacuum in their absence means that nobody is around on the Mac to make the case for an industry standard solution to these problems. Maybe Apple has something in the works, if they do, nobody who knows is talking to me. I certainly don't want to claim that these companies are doing anything wrong, they are just trying to make money and improve their products. But it's easy to forget how much something like MIDI means to us as music makers, and I advise everyone to keep an eye on this to keep your booty covered if the fragmentation I see comes to pass.

New Musikmesse Toys

Well, not surprisingly, I'm most up on what Prosoniq had. The main new toy was sonicWORX 2.0. The LONG awaited update to their flagship audio editor. The biggest news in 2.0 is the new real time page in sonicWORX Studio and Powerbundle, which allows VST plug-ins and Prosoniq real time modules to be used together on an audio file. The modules themselves, and the design of the interface, are fantastic. Artist and Powerbundle also have a number of new offline processing options, and there has been a major reworking on the entire interface. Look for a review here sometime before 2010. I also got to play for hours with the Prosoniq Orange Vocoder. It kicks booty, I gotta get it soon.

Creamware had a huge area for their four new products. Most interesting for Mac users is the Powersampler. It's a hardware accelerated sampler PCI card that will read Akai CDs and work as an ASIO card. It's finished for the PC, and it seems pretty cool. Creamware wasn't saying exactly when the Mac version would be finished, but they seem confident it won't be terribly long. The interface looks nice and it sounded good, though assessments of sound quality in a trade show floor are always pretty sketchy.

Another piece of hardware comes from RME. I was treated to a horrifying story of what it took to get Mac drivers for their Hammerfall card done. Eventually, they decided to bring the Mac driver development in house, and they had a working demo on the floor in a G3. The Hammerfall, in case you don't know, is a rocking good card with a unique touch. As they put it, it includes "ASIO in hardware." A chip on the card holds the ASIO driver, resulting in both very low latency and dramatically reduced CPU overhead for ASIO, even with all the I/O turned on. I'm going to try to get a review card to check it out, but it's a very reasonably priced way to get LOTS of digital I/O in your system. The super low latency numbers also make it an excellent option for software synth users, where latency on playback can be such a problem.

Native Instruments continues to do stuff that shouldn't be possible. The big new toy from NI was the B4, a software version of a Hammond B-3 organ. My days as a blues musician playing weedeldy weedeldy guitar solos that were too long taught me a great deal of appreciation for B3s. Nothing fills out a sound quite like a B3, and they are also just a gas to play with. Coupled with an actual Leslie speaker cab, they add a special bit of demonic possession to your band. They are almost never work quietly, and the various rotating devices in them don't usually completely stop, so even just sitting still between songs, they pant and wheeze like an armoire from the Exorcist. And, remember, that's just sitting still. Native Instruments was working on development until right before the show, so what they had wasn't a completed product, but the sound designing and modeling are finished, most of what's left I was told is cosmetic. The sound is truly extraordinary. I heard a couple of real B3 guys playing it and I was floored. I don't know if anyone can get the armoire from hell aspect in a plug-in, but they could NAIL the sound from "Green Onions" and a few other B3 signatures. I think it will be on the "must have" list pretty much the day it's released.

Not all American companies made much of a showing at the Musikmesse. One that did is Digidesign. Digi had a monster booth with both high end and mid-line products being shown. I watched some guys playing around with the Pro Control surface, which I wish I could use. They were also making a big deal out of the Digi001 interface and the various software tools that can be integrated with it. Some DAW manufacturers seem hell bent on ignoring the market share and popularity of dance music and the tools that surround it. After all, most people with "golden ears" can get too excited about a being able to draw your own wave oscillator in a soft synth. Somehow it's just not as sexy as talking about different models of Neumanns and Avalons. Digi was always a bit stodgy about this kind of thing, and seemed to intentionally pursue the high end and pitch their products only as replacements for analog recording tools. Perhaps they change their pitch because they are in Europe (where dance music completely dominates pop radio), or maybe it's a global change within the company, but their demos included lots of stuff about the Access Virus plug-in for Direct Connect, and they talked up koblo as well. Native Instruments has also announced Direct Connect will be part of Reaktor. I can't imagine they won't sell a ton of those 001 boxes.

As mentioned before, Steinberg announced Cubase VST 5.0. Ship date is June (presumably this next one), and the new features are impressive looking. The high end version will support 32 bit floating point files, and the lower two will now support 24 bit files. A new set of plug-ins will be included, and the EQ actually looks like it might make sense now. They were also pushing their "True Tape" simulation to get saturation sounds onto tracks at the get go. They are using the sound from the well reputed Magneto plug-in to get that analog love everyone is always pursuing. Finally, the high end version will also include the UV22 noise shaping that Steinberg licensed from Apogee. I hope in their zeal to include these new toys, they didn't forget to get in there and fix the display bugs in the Mastertrack.

Directly across from the Prosoniq booth was the Edirol booth. I got to hear the demo song for their new USB audio/MIDI control box about fifteen thousand times. Thank goodness it wasn't too bad a song, and the musicians weren't half bad. They were showing the U-8. There are no Mac drivers for it yet, but it's still pretty cool. It's a little box that has some MIDI faders, a couple of mic pres, analog and digital I/O, and some effects built in. When it comes over to the Mac, it might be a nice entry level audio interface if your sequencer supports it.

Mysteriously, Adilla, the Portuguese beauty queen handing out flyers at the Edirol booth, was able to resist the laser charm exuded by your dedicated author. In fact, she seemed have absolutely no problem at all resisting said charm. I think maybe something was lost in translation . . .

The chorus of silence surrounding OS X was deafening. I didn't talk to a single developer who was very far along with their work supporting OS X. I know that everyone is planning on supporting it, but people are taking a real "wait and see" attitude about the next generation OS for the Mac. As the driver support gets worked out and the nature of the audio and MIDI features come forward, I think we'll get a better idea. Still, don't expect to install it after Mac World NYC and fire up all your audio apps and have your sound cards work. This is not to say that people aren't excited about some of what OS X has to offer. The real problem, of course, is one of resources, and even the top tier companies don't have lots of developer hours to spare. As of now, I'd say that OS X support will be very much incremental for us music makers.

I know I'm not really supposed to talk about non-Mac external hardware that much, since this site is all about getting "in the box." But I have to note one thing. I want every single synth Waldorf makes. In fact, I want two of a few of them. They all kick booty. They are also doing a bit of software themselves. Steinberg announced that they are distributing a VST plug-in version of the PPG 2.3, an early and fairly legendary Waldorf offering. I saw and heard it at the Steinberg booth, and if you liked the original, start putting your pennies aside.

All in all, it was a great experience, thanks to Prosoniq for making it possible. Next up, the final, long overdue, can't possibly be worth the wait, Reaktor for Mac review. Bis gleich . . .


Have an Audio question? Check the Audio FAQ first, then the General FAQ Audio topic area.

Send Thad Feedback or new links at: tcb@caliban.grendelnet.com


Back Issues:

Back to XLR8YOURMAC.COM
Your Source for the best in CPU/SCSI/VIDEO card reviews, daily news, and more!

Disclaimer: The opinions/comments expressed here are the author's alone,
and do not necessarily represent those of the site publishers. Read the site Terms of Use.

xlr8yourmac, 1997-2000.
No part of this site's content is to be reproduced in any form without written permission.
All brand or product names mentioned here are properties of their respective companies.