ORIGINAL: pianodano
Regardless of any difficulties in making it work to at least achieve the old benchmarks under the current os, the real and central issue appears to be (imho) that midi has simply been relegated to the back burner and the resources are directed towards including the latest wizbang vitural audio thingamajig or new widget to make the (your choice of user) buy in at the continued expense of the intensive midi users who have, for far to long, been thrown only occasional little tidbits.

It's not MIDI that's been on the backburner, but dedicated clocks within the computer and OS for audio / MIDI timing. Because of the way DAW's have become universal / native applications that boast the feature of widespread compatibility across a variety of consumer and pro audio and MIDI devices, there is a requirement of certain basic support from the hardware and OS - and dedicated clocking for MIDI communications hasn't been included. The solution *could* be as simple as designating that MIDI interfaces (be they standalone boxes or the USB controller built into a keyboard) have their *own* timing source, or that there be a truly universal time-stamping protocol for the next generation of MIDI spec that takes advantage of the current stability of audio clocks, since most people using a DAW have some sort of audio device with its own clock. That way, your MIDI would be as stable timing-wise as your audio. But it would require that all of the MIDI manufacturers adopt this approach and incorporate the protocol into their drivers, and that the DAW programmers take advantage of the timestamps. Not impossible; just a standards hurdle like all the rest.

What do you (all) make of this, wrt jitter, accuracy, a computer's ability to translate a performance..... that kind of stuff....

Glenn Gould in Re-performance
Bach’s Goldberg Variations
http://zenph.com/sept25.html

"it was locked forever in monaural sound"

I'm dismayed that anyone would think that there is some downside to enjoying a fine classical recording in the mono format.

Most of my favorite Classical recordings are in mono and I use a mono playback system to enjoy them to their fullest.

best regards,
mike

If the computer is translating data in as uninterupted a state as possible then timing issues become almost nil. The speed and bandwidth issues pretty much exced what is needed for high quality reproduction. What we have a problem dealing with is Microsoft's one size fits all operating system (bloatware). The next great thing in computers will likely be an operating system that is clean and lean as the users tasks require and nothing more. Microsoft has always lead in exactly the opposite direction. Maybe in a few years Unix or something completely new will find a way to smack MS aside.

If the computer is translating data in as uninterupted a state as possible then timing issues become almost nil. The speed and bandwidth issues pretty much exced what is needed for high quality reproduction.

Actually, that's not quite the case. Modern transports involve assembling data (MIDI or audio) into packets, which always introduces some amount of delay. This is a serious issue when packing/unpacking MIDI data to/from USB or Firewire packets, and puts a lower bound on what kind of latency can be achieved. If you want to use Ethernet/Firewire/whatever for MIDI transport, _and_ you want to keep jitter to imperceptible levels, then some level of buffering delay is unavoidable (and quite acceptable, IMHO).

That said - it's technically quite possible to get MIDI latencies on the order of 1-2 millliseconds or so, and jitter of well under a millisecond, if the driver stack is designed properly and the operating system and other drivers don't get in the way . Unfortunately, those are pretty big "ifs" with Windows...

Russell T. Graham is on the right track, btw...

- Jim

ORIGINAL: mike_mccue

"it was locked forever in monaural sound"

I'm dismayed that anyone would think that there is some downside to enjoying a fine classical recording in the mono format.

Most of my favorite Classical recordings are in mono and I use a mono playback system to enjoy them to their fullest.

best regards,
mike

I shouldn't care, but I do. I can't handle mono when stereo's available. I.e. I'll purchase the Callas re-records in stereo just to get the strings in stereo. Shouldn't matter but it does (to me).

ORIGINAL: Marah Mag

What do you (all) make of this, wrt jitter, accuracy, a computer's ability to translate a performance..... that kind of stuff....

Glenn Gould in Re-performance
Bach’s Goldberg Variations
http://zenph.com/sept25.html

Interesting. I haven't had time to actually listen to the clips on the site yet, but I poked around for some info on what exactly they're doing in terms of their "translation" from classic classical recording (say that 5 times fast) to "reperformance." Maybe there's something I'm missing on the site, but it seems like they're keeping their process pretty close to the vest. I can't find info on whether they're using audio-to-MIDI triggering, what OS they're running on, what sequencing software (if any), whether it's a proprietary system or platform, etc. Without knowing whether they're running some combination of publicly available products, it's difficult to conjecture as to how much the typical MIDI timing issues might affect the performance.

That said, it's certainly an interesting concept. If they *are* in fact using one of the more popular sequencers / DAWs, then I'd be curious to know what they have to say about the typical MIDI jitter issues, and to what extent it may or may not affect the "nuances" or "anomalies" that they talk about as being signatures or trademarks of these classic classical recordings that they are "reperforming."

ORIGINAL: Jim Wright

Russell T. Graham is on the right track, btw...

Well, thanks. Doesn't seem to help much though, unless some MIDI interface manufacturers want me on their boards, or are reading this forum.
We'll have to wait and see what happens.

Here's some more info. They're using a Yamaha Disklavier Piano.... at least for playback. But no, there doesn't seem to be much on how they're capturing the re-performance data.

Fascinating though. Is the reperformance really "by" Glenn Gould? Does he get performance royalties on CD sales?

http://www.musiciansnews.com/ya/yamaha/yamaha_disklavier_piano_helps_restore_glenn_gould_recording.shtml

Would be interesting to run play the original recordings on top of the reperformance. They should line up exactly Presumably they'd've done that as part of their proof...?

Me... I'd be happy if I could get V-Vocal's audio-to-MIDI to work an a bass part!

ORIGINAL: Marah Mag

Fascinating though. Is the reperformance really "by" Glenn Gould? Does he get performance royalties on CD sales?

...
...

Would be interesting to run play the original recordings on top of the reperformance. They should line up exactly Presumably they'd've done that as part of their proof...?

Good questions.
It reminds me also of a CD that I have, that I was very excited about when I got it, of a "reperformance" (though I don't think that's the word they used) of Gershwin playing some of his own compositions, including Rhapsody In Blue. If I'm remembering it correctly (the actual CD and liner notes are in storage, having been converted to MP3), the technicians and producers rigged up a player piano to control robotic arms that then played a concert grand piano, and thereby "played" Gershwin's own piano rolls on a higher-quality instrument. I don't recall whether there was any MIDI capture involved, but I recall wondering at the time just how closely the performance was to the way Gershwin actually played it - having used player pianos in the past, I'm aware of how the various levers and settings on the instrument can mechanically affect the playback speed, dynamics, etc.

EDIT:

Hey - what happens when you reach 1000 posts here?

ORIGINAL: RTGraham

Hey - what happens when you reach 1000 posts here?

I'm not sure. When I hit 999, I created a new name and started again at 0.