4. Audio files and inputs

New objects: readsf~, key, adc~, env~, dbtorms, sigmund~, delwrite~, delread~

Resource: Freesound

Media: bells-cluster.wav - [source]

Review previous patch

• arpeggiator-2.pd
• Floats are weird to wrap your brain around at first, but they’re very powerful.
  • remember: hot inlets send outputs immediately, cold inlets take in data and wait
• expr is handy for stringing arithmetic operations together

Loading samples and playing them back

• find the sound file you want (pick something small)
  • record your own
  • download something from freesound
  • example above (bells-cluster.wav)
• put the file in the same folder as your patch
• readsf~
  • simplest way to play back audio (note: very limited!)
  • argument sets number of channels (default is mono!)
  • output to dac~ (can go through *~ to adjust volume or create an envelope with line~ or vline~)
  • takes in messages for controls
    • open [filename] (if giving a file name alone, it must be in the same folder as the patch, otherwise a relative or absolute path can work)
    • give a second or two to load the file, then send a 1 message to start playback and a 0 to stop it
    • IMPORTANT! The sound file is cleared from memory after it is done or stopped. It must be reloaded each time.
  • outputs a bang from the rightmost outlet (after the audio channels) when playback completes
    • demo: handling “done” signals
      • hook up a bang to the rightmost outlet for debugging
      • use “done” outlet to reload the soundfile so it is ready the next time you need it
• more powerful audio playback can be achieved with soundfiler and tabread4~. You can read about them in Help.

Triggering events from the keyboard

This one is pretty simple.

• key watches keyboard inputs and sends a number for each “printing” key to the outlet each time the key is pressed.
  • use a number box to figure out which is which
  • use select to grab the numbers you want to have them do stuff like play back a sample or change notes in our arpeggiator from earlier
• great for simple triggers (like “start the next section’s audio track”, “start a counter”, “release a sustained sound”, “toggle an audio effect on or off”)
• limitations
  • holding a key can create different results based on system settings (recall the result of holding down a letter key while typing in a text document)
  • operating system settings might change what the code numbers are, and people in other countries may have different keyboard layouts and numbers
  • Pd must be the frontmost application to get these messages.
• more powerful alternatives:
  • MIDI
    • go to Help > Pure Data > 5.reference > midi-help.pd for lots of useful examples of the midi objects in Pd
  • OSC (oscparse receives and oscformat sends)
    • more powerful, but a little trickier to use
    • TouchOSC (iOS and Android app) is a great way to build a touch interface for your patches. Possibly the best cheap ($5) tool in computer music.

Microphone inputs

Meet adc~, dac~’s backward cousin.

• check your audio settings first
  • Media > Test Audio and MIDI…
• demo: follow incoming audio level with adc~ and env~
  • make an adc~, note that like dac~, it will use your default audio device and assume you want it to be stereo (can be altered with arguments like dac~)
  • incoming stream is audio data
  • convert to volume levels with env~
    • takes an incoming audio stream (could be a file or mic)
    • outputs the envelope level in decibels (show with number box)
    • decibels are tricky to work with mathematically, so we want to convert the range to our usual 0-1 scale with dbtorms
    • show RMS volume with number box
  • What is this number good for?
    • controlling the envelope of a synth
    • controlling the pitch or timbre of a synth
      • do some math to map an appropriate range for pitch, amplitude, modulation index, etc.
  • other info from the incoming signal: sigmund~
    • audio inlet, number outlets
    • left outlet: pitch
      • MIDI number, with decimals for pitch variation
      • only runs if there is enough volume and pitch clarity
      • filter out unpitched things with select -1500
    • right outlet: identical to env~

Intro to live DSP

Digital signal processing. Have a quick trigger finger on cmd-. or ctrl-. when playing with these!

• Use delwrite~ and delread~ to create a basic digital delay (there are more powerful options, but this is a good start)
• open the helpfile for delread~ as an example
• delwrite~ stores incoming audio to a buffer (temporary quick-access storage)
  • name of the delay buffer (could be whatever makes sense to you)
  • number of ms to store in the buffer (This is the longest delay possible with this particular buffer. It is not the actual delay time.)
• delread~ plays back from the buffer
  • name of delay buffer
  • number of ms to delay (This is the actual delay we hear!), delay time can be changed dynamically by sending a new number to the inlet!
• multiple delwrite~s to the same buffer causes problems, but multiple delread~s does not.
• demo: simple delay “pedal”
  • copy the delread~ helpfile to a new patch
  • clean up the comments
  • change the delay buffer name! (important, as the helpfile is still trying to write to that same buffer, or you could just close the helpfile)
  • remove sig~ and replace with input from adc~
  • before the next step: Find the mute button for your speakers and/or get ready to kill DSP (cmd-.)
  • remove snapshot and replace with output to dac~
  • How can we control this better so we can more elegantly (and programmatically) turn the delay line off or on?
    • scale adc~ and/or dac~ with *~ to 1 or 0
    • ramp up or down with line~
• related objects:
  • delread4~ is more complicated but more powerful
  • delay~ is powerful and a little simpler than delread4~, but requires the cyclone extension (Help > Find externals and search for cyclone)
  • freeverb~ works similarly (actually, it’s much simpler) to process audio live