gave Thanks, Gene! I think the strongest effort to replace MIDI-1, is OSC. OSC-over-Ethernet seems the common transport, and it doesn't seem to have a maximum transport speed or rigid timing like MIDI does. There is an HD-MIDI out there but it still appears to be mostly in the process of seeking to be used. But just now I was not able to find a single 88-key weighted controller with OSC. Hopefully in the near future. J.E.B.

H'm. Some questions arise: 1. Is MIDI-over-USB as constrained in timing, as MIDI-over-DIN? 2. I had not heard of MIDI-over-OSC. And although I had heard of MIDI-over-Ethernet I had not thought of this context. Is it possible to send MIDI events, either encapsulated in OSC or by USB or Ethernet, without that low timing resolution of the DIN setup? If I set up something like USB-MIDI-to-Ethernet, is it possible to see real advantage? 3. Is JACK MIDI constrained in the low timing resolution of original MIDI, or is it more flexible? If/when more than 16 patches are needed in a setup based on my current, I'll probably do some sort of bank switching based on program change commands, but the above timing issue appears to be worth contemplating! -- Jonathan E. Brickman | jeb(a)ponderworthy.com | (785)233-9977 Ponderworthy | http://ponderworthy.com Music of compassion; fire, and life!!!

There are two things to keep in mind here: 1. Are events handled sequentially in time? 2. How many events can be transmitted in a unit of time? If events are handled sequentially in time, then it is not possible to play a chord of multiple notes at _exactly_ the same time. On Mon, Feb 16, 2015 at 5:11 PM, Jonathan E. Brickman <jeb(a)ponderworthy.com> wrote: No, the MIDI "cable" baud rate is 31250, while MIDI over USB runs at 112500. Its faster - but there are still limitations. Regarding the above: -Sequentially handled: yes -Bandwidth: lots better than DIN-cable MIDI, but limited Yes - almost everything is possible with OSC - any form of data can be passed around as a "blob" of binary data, timestamps are available to be used for timing, and "bundles" can be used to have multiple events be processed at exactly the same time. Sequential handling? Yes - unless in a bundle. Bandwidth: unlimited. JACK MIDI is sample accurate - and multiple events can be written on a single audio frame (internally in JACK graph). Sequentially: No - except on playback to hardware devices Bandwidth: configurable (at compile time IIRC), but "lots". ipMIDI: Not much techie stats online available - http://www.nerds.de/en/ipmidi.html rtpMIDI: No details easily found on how it actually works - http://www.tobias-erichsen.de/software/rtpmidi.html Another list of MIDI over IP projects: http://openmuse.org/transport/mip_oview.html To sum up - JACK MIDI is awesome - its widely used, it is sample-accurate within the JACK graph, and events can occur at exactly the same times. With the QMidiNet project, broadcasting JACK MIDI over the network is possible, http://qmidinet.sourceforge.net/qmidinet-index.html I've not extensively used / or concidered-using the others - there's too much choice, and the advantage over JACK MIDI is minimal to non-existant. Cheers, -Harry -- http://www.openavproductions.com

Hanspeter Portner wrote: USB MIDI must use bulk transfers, which are not allowed for low-speed devices. 512 is the _only_ allowed value for the maximum high-speed bulk packet size, so every high-speed device must be able to handle such packets. While MIDI over DIN is asynchronous, USB bulk transfers have an implicit throttling mechanism, i.e., a packet transfer is not complete until the device has chosen to accept it. This means that the bandwidth is just as high as the device can process the data. USB/MIDI interfaces cannot run faster than DIN cable speed, and there might be devices where the USB interface has been duct-taped onto an existing MIDI interface. (And the really stupid MFM0860 chip is _slower_ than 3125 bytes/s.) Regards, Clemens

On Mon, 16 Feb 2015, Harry van Haaren wrote: I had not heard of it either, but while reading the documentation on liblo (an OSC lib) I found there is a function just for sending MIDI events. Note that this is a 4 byte event only, not sysex. I don't know if there are size limits to OSC blobs... but when one wants to do data over ip of any size, timing becomes secondary because both ends will need time to deal with 1000 patches... why use OSC? there are lots of ways to send a config file of any size over IP already. It has always been up to the designer how they use sysex anyway and how transparent they decide to make it. Some have been very transparent and well documented so that others could make things like patch editors. Others have been just a binary blob, storable, but not easy to edit. netjack can also extend jack MIDI across the net. For 99% plus music performance use, MIDI is not really out of date. That is why all keyboard/synths still come with it. The connector has been suplemented with USB for higher speed/throughput, but not much else has changed. Pitch bending still produces more data than is easy to handle... do we really want higher resolution? More channels? How many keyboards will I play at a time? Even with two kb they are probably USB anymore and so show up as two separate MIDI devices (32 channels). Maybe some of the things we want to do with MIDI should be done with something else. MMA is working a thing called HD (not MIDI they say) that has bigger everything. I am kind of shy of anything called HD any more as being hype. I think there are better ways of dealing with control surfaces than repurposing MIDI (or expanding MIDI). In my mind, the Mackie control surface shows how really bad MIDI is for that use. Almost the whole thing is on channel one... execpt 9 faders take 7 channels of "Pitch bend?" The whole thing is a kludge to make MIDI fit where it really doesn't. MCP is a one horse game, it was made for one DAW... it is just the fact that people can see what the data is that it sends for each move (there is documentation) that has allowed other DAWs to use it. If anyone was to build a control surface for a well known top selling DAW that used OSC, those OSC commands would have become "standard" in the same way. The difference being that any OSC set of commands I can think of would have been expandable. All this to say (I think) that to make OSC something people use requires a great piece of hardware that works well with the most used DAW. Get that DAW to suggest this is "THE" control surface to use to get the most out of your SW. I think OSC is missing parts though (just like AES67?). I should be able to add a control surface/keyboard to a network and the unit should show me every other unit (SW or HW) that it can talk to. From that list I should be able to connect my output to a number of inputs and connect my input to whatever output. Then if I play notes on a keyboard I should get the right sounds, or move a control the right thing should change. That is a lot of things to happen. My controller has to be able to set it's networking up and then discover what other OSC stuff is around. OSC right now seems to be totally manual. The user has to know the target IP and port (99% musicians just blanked out) and have the server running and enter those things just to have a connection. Even with a standard set of commands this is already a fail. I think this is the reason MIDI is still going strong, plug a keyboard/controller into MIDI/USB, the SW shows the available MIDI ports, the user selects the one they want... done ready to role. Going beyond MIDI requires something at least as good. -- Len Ovens www.ovenwerks.net

On Tue, 17 Feb 2015 10:25:23 +0100, Raphaël Mouneyres wrote: Len isn't just right with this. Len also explained why old faithful MIDI, even with it's drawbacks, is a very useful thing. I'm a MIDI guy from the beginning and I disliked MIDI much, but nowadays we seldom suffer from MIDI issues, usually we suffer from odd MIDI implementations. The most annoying issue IMO is that we don't get hard real-time nowadays. There's a big difference between computers from the 80s and 90s, providing hard real-time and the crap we get today. It always was possible to use more than just one MIDI interface. The first computers, such as the C64 suffered from bad resolution, but later computers, such as the Atari ST didn't suffer from this issue. The baud rate and other limitations are an issue, but the real big issue nowadays is the design of the hardware and software. Fortunately we are still free to use more than just a single MIDI interface, we still could use separated MIDI interfaces for each channel. SysEx data isn't real time data, but it can be treated as real-time data, assumed we use several MIDI interfaces and the rest of the hardware should be ok and the software should be ok too. For example, there's no need to send checksums with SysEx. SysEx "commands" often cause less traffic than Pitch Bend or After Touch does. When using MIDI, the KISS principle is the most important approach, but unfortunately it seems to be out of style.

On Mon, 16 Feb 2015 10:11:50 -0700 (MST), Jonathan E. Brickman wrote: As Paul already said, no, it isn't. You can use several "virtual" MIDI outs and you also can use several hardware outputs. [rocketmouse@archlinux ~]$ amidi -l Dir Device Name IO hw:0,0 HDSPMx579bcc MIDI 1 IO hw:1,0 TerraTec EWX24/96 MIDI IO hw:2,0 TerraTec EWX24/96 MIDI For audio and MIDI I use the PCIe HDSP card and for MIDI in addition I use at least one of the TerraTec cards. IOW I've got 3 * 16 channels. Sometimes it's also good to have separated interfaces, even if you're using less than 16 channels. I'm using USB MIDI to dump data from external synth, but for short SysEx data in "quasi real time" I'm using the PCIe and PCI devices. -- "Aber beklecker nicht das Sofa, Sofa!" - Frank Zappa

On Sun, 15 Feb 2015 16:12:16 -0800 (PST) Len Ovens <len(a)ovenwerks.net> wrote: This is the thing that made MIDI take off. For once, a group of people got together and designed a standard right at the onset. I very much like the idea of OSC in principle, but this apparent lack of any kind of standard will make it very hard to achieve mainstream acceptance :( -- Will J Godfrey http://www.musically.me.uk Say you have a poem and I have a tune. Exchange them and we can both have a poem, a tune, and a song.