Difference between revisions of "256a-fall-2010/hw1"

From CCRMA Wiki
Jump to: navigation, search
(Specification (part 1 of 3): Real-time Audio)
Line 81: Line 81:
*** lists any collaborators
*** lists any collaborators
* 3) '''email Ge and Nick''' with the link to your web page, as a confirmation that you are submitting the assignment
* 3) '''email Ge and Nick''' with the link to your web page, as a confirmation that you are submitting the assignment
** ge /at/ ccrma.stanford.edu  and njb /at/ ccrma.stanford.edu

Revision as of 17:46, 21 September 2010

Homework #1: Real-time Audio, Buffers, and Waveforms

Due date: 2009.9.29 11:59:59pm (or thereabout), Wednesday.


Let's get cookin'.

Specification (part 1 of 3): Real-time Audio

  • create a program that is capable of real-time audio input/output
  • give it name (e.g., sig-gen; creative names are always welcome)
  • start with a blank C++ program
  • create a minimally compilable program (e.g., something like HelloWorld):
    • if you'd like, you can start with this very basic makefile (will need to make changes). For a simple introduction to make files, check out this link
  • next, add real-time audio support, using the RtAudio Library (version 4.0.4)
    • download it from here
    • even though it's useful to briefly look through the package, the only files you'll need are:
      • RtAudio.h (the header file for RtAudio, it contains the class definitions)
      • RtAudio.cpp (the implementation)
      • RtError.h (header containing various error handling constructs for RtAudio)
  • this is similar to (but not identical to) the example we did in class (see HelloSine):
    • your program is using an updated RtAudio interface, which is different from the one we used in class
    • it may be also useful to browse the RtAudio documentation and the example programs in the RtAudio distribution
    • NOTE: even though the code is nearly all there in the example, it's infinitely more useful to actually write the code from scratch - even if you copy/type it in line by line!
  • implement the callback function to generate the expect number of samples per call for a sine wave at 440Hz
  • the overall behavior when you run the program should be a continuous sine tone at 440hz...
  • to quit: press ctrl-c

Specification (part 2 of 3): Waveforms

  • modify your program to take command line arguments and generate different signals, depending the command line flag you specify:
    sig-gen [type] [frequency] [width]
        [type]: --sine | --saw | --pulse | --noise | --impulse
        [frequency]: (a number > 0, only applicable to some signal types)
        [width]: pulse width (only applicable to some signal types)
  • where the flags correspond to the following signals:
    • --sine : sine wave
    • --saw : saw tooth, the width is a number between 0.0 and 1.0 the determines the shape of the wave (e.g., width=.5 should result in a triangle wave)
    • --pulse : rectangular pulse wave, the width ([0.0-1.0]) controls the pulse width (e.g., width=.5 should result in a square wave)
    • --noise : white noise
    • --impulse : impluse train, the frequency should determine the impulse train's fundamental period
  • it might be a good idea to output the usage (as show above), if insufficient or incorrect parameters are given
  • you'll need to implement a simple command line parser, with basic error checking (e.g., what to do when invalid/irrelevant parameters are provided?)
  • you'll also need to organize your code a bit, to selectively generate the request signal

Specification (part 3 of 3): One Ring to Modulate Them All

  • Lastly, add another command line flag:
    • --input : mic/line input (make sure to enable it in the code when initializing RtAudio)
  • if specified, this flag tells the program to take the mic/line input and and multiply it against the signal being generated, and output the result!


  • have fun with it!!!
  • your code should compile and run on the CCRMA machines
  • comment your code!
  • choose your own coding conventions - but be consistent
  • you are welcome to work together, but you must do/turn in your own work (you'll likely get more out of it this way)
  • some considerations:
    • how to organize the code for the various types of signals?
    • how much error-checking and error-reporting on the command line arguments?

Testing Audio Options

  • Use Chuck as a reference signal generator
  • For an easy way to test the generated audio output and avoid writing a wavefile using something like Libsndfile, just write a binary output file. The Audacity audio editor can "Import Raw" audio. Select the correct format and import. Audacity is installed on all CCRMA machines or can be downloaded for free at audacity.sourceforge.net. You will automatically view the time domain signal. To see the frequency response, you can select a portion of audio and click Analyze->Plot Spectrum.


turn in all files by putting them in your Library/Web/256/hw1/ directory, and concise online documentation + readme

  • 1) source code to the project (*.h, *.cpp, *.c makefile, etc.)
  • 2) online page for your project (should be viewable at http://ccrma.stanford.edu/~YOURID/256a/hw1/). It should include:
    • links to your files of various kinds
    • instructions on building the project (for example, anyone in the class should be able to download
    • a short README text section that:
      • conveys your ideas/comments in constructing each program
      • describes any difficulties you encountered in the process
      • lists any collaborators
  • 3) email Ge and Nick with the link to your web page, as a confirmation that you are submitting the assignment
    • ge /at/ ccrma.stanford.edu and njb /at/ ccrma.stanford.edu