MIDI and Latency (was Re: About Teleo)


    Feb 02 2006 | 2:21 pm
    Here's my 2 cents. Before you make a big deal over latency, you
    should do a reality check and make sure it is an issue. In everything
    I have ever done for over 20 years in interactive performance, MIDI
    latencey has _never_ been a problem.
    I have sent signals from a force sensing resistor over a wireless
    link and out a MIDI interface, checked for a threshold crossing in
    Max after receiving the MIDI and used this to trigger a sample, and
    the result has been effectively instantaneous. I can tap the FSRs
    and the samples sound like they're triggering immediately. Consider
    the potential latencies here, and you'll see what I mean.
    If you're moving ten faders at once really fast controlling synched
    oscillators at high resolution, MIDI latency might be an issue. In
    practice, it almost never is.
    The threshold for distinguishing two notes as separate events is
    about 30 ms (yes, you can argue all different figures for this, but
    this is a decent benchmark). The threshold for hearing a note as
    being "late" is probably slightly longer. MIDI costs about 1 ms per
    message. Add a few more ms for the interface and the time for the
    software to respond, and you're still way under.
    On the upside for MIDI, it works with everything and is extremely
    robust as a transport mechanism. (A friend reported to me that he put
    MIDI connectors at either end of a kilometer of speaker cable,
    injected broadband noise onto the cable and it _still_ worked.)
    OK, speed freaks, now go ahead and slam me :-)
    Eric
    --
    ********************************************
    * MidiTron MIDI to Real World Interface *
    * Now available from http://eroktronix.com *
    * Only $149 *
    ********************************************

    • Feb 02 2006 | 3:48 pm
      Just curious, what type of wireless link(s) have you used to send MIDI
      signals from FSR 's?
    • Feb 12 2006 | 11:38 am
      Thank you Eric.
      In effects, 1 ms per message is not an issue.
      In a couple of days I have to decide :-) , so I would like to ask one thing:
      I will use 20 analog outputs in Miditron, connected to a drum machine with 20 samples, outputs are linked to 20 lamps (analog outputs for dimming) , so that the amplitude of any sample is reflected to lamps in real time.
      Is it possible for you? I mean, how many lamps could I link without having latency issues?
    • Feb 12 2006 | 1:51 pm
      keep in mind that the ear is much more sensitive to latency than the
      eye.
      in sound, 10 ms is at the critical treshold, in vision, 50 ms is.
      this is the basis of 25 fps as film standard.
      moreover, if you use candescent light (ordinary light bulbs), the
      response time
      of the lamp is much larger than 50 ms. If you switch off the lamp, it
      still glows
      for the better part of a second.
      HtH
      -jennek
    • Feb 12 2006 | 2:22 pm
      Thank you very much.
      And does it exist something better than candescent light?
      Something that have less latency and that is not "binary"...so that I will obtain 128 values of brightness...
      (like "neon" for example? )
    • Feb 12 2006 | 4:19 pm
      traditionally: light bulbs with mechanical shutters.
      recently: light bulbs with LCD type shutters (aka beamer)
      i have no experience in regulating neon.
      I do have in using TL, and that is not trivial in the low end
      (it is difficult to drive a TL from zero to say 10% as it takes
      more than 10% power to ignite the tube)
      -jennek
    • Feb 12 2006 | 5:21 pm
      strong LED modules, controlled by PWM (pulse width modulation). as
      long as i can remember miditron should have few PWM outputs. To some
      extent LEDs can be dimmed by applying lower voltage, but the curve is
      definitely not linear.
    • Feb 12 2006 | 8:03 pm
      LED modules for use in theatre etc you will find at http://
      www.lagotronics.com/
      -jennek
    • Feb 13 2006 | 3:47 am
      You can run all 20 lamps. However, you will trade off PWM speed
      and/or resolution. The manual (downloadable from
      http://eroktronix.com) has a chart of maximum rate/resolution versus
      number of PWM channels.
      Eric
    • Feb 13 2006 | 3:53 am
      At 6:21 PM +0100 2/12/06, klif wrote:
      >strong LED modules, controlled by PWM (pulse width modulation). as
      >long as i can remember miditron should have few PWM outputs.
      Up to 20, with maximum refresh rate going down as more PWMs are enabled.
      Eric
      --
      ********************************************
      * MidiTron MIDI to Real World Interface *
      * Now available from http://eroktronix.com *
      * Only $149 *
      ********************************************
    • Feb 13 2006 | 12:43 pm
      On 12 Feb, 2006, at 14:22, Bruno Zamborlin wrote:
      > Thank you very much.
      > And does it exist something better than candescent light?
      > Something that have less latency and that is not "binary"...so that I
      > will obtain 128 values of brightness...
      > (like "neon" for example? )
      >
      LED systems are probably best - you can now get LED replacements for
      bulbs in theatre lamps - at a price of course.
      L
      Lawrence Casserley - lawrence@lcasserley.co.uk
      Lawrence Electronic Operations - www.lcasserley.co.uk
      Colourscape Music Festivals - www.colourscape.org.uk
    • Feb 13 2006 | 4:27 pm
      Thank you all, I'll try theatre lamps (as soon as Miditron will arrive) :-)