Akai MPC

[jones-y]

While that may be the case, it is only relevant should you be using a VSTi 'live' or if your sequencing a whole load of external gear. If your computer is triggering virtual instruments with its midi, the whole notion of MPC's having a 'tighter feel' or whatever you want to imply is irrelevant. If you record a 'take' of a couple of bars, any supposed 'usb latency' captured by your sequencer is easily rectified and a beginner will almost certainly quantize the data anyway.

That's a lotta 'if's'...

And latency isn't the problem. I can deal with latency. It's jitter. It's good that you acknowledge that there's a problem to be 'rectified' post-recording. For me that's unacceptable.

Me personally, I prefer my own 'swing' or 'groove' so having a sequencer that can capture it is a requirement. I could give two mistletoes about a feature set if the sequencer can't get it's most basic and fundamental function right.

I think people have this misconception that people who use hardware use hardware _only_. I have a DAW (don't know why I'd be on KVR if I didn't). I just don't (well rarely to be precise) use it for MIDI sequencing. I do, however, use my computer 'live' as you would say; I use it as a sound module (among other uses like recording, arranging, and mixing and of course checking KVR every now and then) so I still encounter its MIDI jitter and it's still off-putting. It's almost 15 years after the onset of the DAW era and things are getting worse, not better. Ataris still outclass the latest PCs in that regard... IMO they all should be ashamed of themselves that they pile on feature after feature year after year but still can't take the time to fix such a fundamental flaw. I guess it doesn't matter to featurewhores Just my opinion...

[afreshcupofjoe]

It's almost 15 years after the onset of the DAW era and things are getting worse, not better. Ataris still outclass the latest PCs in that regard... IMO they all should be ashamed of themselves that they pile on feature after feature year after year but still can't take the time to fix such a fundamental flaw. I guess it doesn't matter to featurewhores Just my opinion...

I generally agree with your point, but you can't really blame music hardware and software manufacturers for this. The problem is that the modern PC isn't designed specifically for music. Device and software manufacturers are limited by the constraints of existing system spec. The reason why an Atari has such good timing is because the midi interface was directly integrated into the hardware from the get go. The Atari ST was designed to be a music machine at the motherboard and processor level. A modern equivalent would be something along the lines of Foxcon deciding to implement midi ports hardwired into all of their motherboards. It's just not something that is going to happen. The only way you would see this is if someone like Alesis got into the motherboard business, which is equally unlikely. The result is that midi will continue to be an "add-on" component for modern computer systems and that means work-arounds will continue to have to be used to circumvent the jitter issue.

[jones-y]

But see here's the thing, a PC is capable of tighter MIDI timing (less latency, lower jitter) through busses other than USB (like PCI Express). The manufacturers just don't really make any devices to exploit that anymore. It isn't entirely on the operating environment as a whole (although I'll cede that Windows will never be a Real Time MIDI OS), its for the most part the architecture of the USB bus.

The manufacturers are treating us like consumers and not professionals, and we are acting like consumers and not professionals by continuing to patronize them.

[Uncle E]

MIDI timing was crappy with every PCI midi interface I've ever had. Here's info about Steinberg's old USB MIDI interfaces:

Steinberg's Midex 8, like Emagic's existing AMT8 and Unitor 8, is designed to provide much tighter MIDI timing than traditional Mac/PC MIDI interfaces, at least within their own sequencer, by using an 'intelligent' buffering system to clock data out (see the How LTB Works box). Data can be sent from multiple ports simultaneously, which is a big improvement over a standard serial or parallel-port MIDI interface where the more ports there are, the more the delays can accumulate.

<snip>

How LTB Improves MIDI Timing On Playback

Steinberg and Emagic made headlines at the 1998 Frankfurt Musik Messe, when they announced that they would in future share certain aspects of their expertise. Unfortunately this camaraderie didn't last very long, as Emagic later decided to develop their own 'improved' version of ASIO called EASI, and Steinberg took Emagic's existing AMT (Active MIDI Transmission) technology and created their own 'improved' Linear Time Base technology. LTB is used for the first time in the Midex 8.

Both require specially written code in their host sequencer application to benefit, which is why the Midex 8 will only deliver improved timing with Cubase VST 5.0 or Nuendo 1.5 (and Emagic's AMT with Logic Audio 4.0.1 or higher). In essence, both do exactly the same thing. Once MIDI data has been recorded, the sequencer sends it ahead of time to the MIDI interface, where it's stored in a temporary buffer. Then a high-resolution clock inside the MIDI interface sends the time-stamped data out at exactly the right time.

This overcomes the sometimes temperamental nature of both the Universal Serial Buss and the operating system, since the interface doesn't have to rely on smooth delivery of data in real time. As long as enough data is sent in advance (even in fits and starts) to keep the buffer topped up, the AMT8, Midex 8, or Unitor 8 will be able to carry on with perfect precision. It also means that MIDI data can emerge at precisely the same time from each of the eight MIDI ports when required, which again would be a problem for any standard multi-port USB MIDI interface.

Bear in mind that just as with VST Instruments, your MIDI timing won't be any more precise during recording -- it's only on playback that higher precision timing kicks in -- so you won't gain anything from LTB or AMT when attempting to capture the subtleties of a live performance. Nor can these technologies overcome the serial nature of MIDI itself: if you send 16 simultaneous notes to a single MIDI port, once on each channel, then they will still emerge as a serial stream lasting 16mS. That's MIDI for you.

[Uncle E]

Check it out:

http://earthvegaconnection.com/evc/prod ... sults.html

[afreshcupofjoe]

Those numbers are a lot more promising. I'll have to download the app and test my hardware.