Sig++ example sounds

---

• Descriptions of example programs that made these example soundfiles.
• Source Code of example programs that made these example soundfiles
• Binaries for programs which created these soundfile examples.
• Scripts that I use to generate these sounds and to make sure everything is still working.

Filtering

---

1. Resonator
   • Generation program source: resonator.cpp [script]
   • Utilized signal class: Resonator
   • Sounds: Using whitenoise as an input:
     • resonator1.wav Constant bandwidth, sweeping frequency linearly.
     • resonator2.wav Constant resonator frequency, widening and then narrowing bandwidth.
     • resonator3.wav Changing both bandwidth and center frequency simultaneously
   
   
   
2. Simplest lowpass filter
   • Generation program source: simplestLowpass.cpp [script]
   • Utilized signal class: Add
   • Sounds: Averages neighboring samples together Using whitenoise as an input:
     • lowpass1.wav
   
   
   
3. Allpass filtering
   • Generation program source: allpass.cpp [script]
   • Utilized signal class: Allpass
   • Sounds: Using whitenoise as an input:
     • allpass1.wav: Loop frequency: 440 Hz, Feedback gain: 0.8
     • allpass2.wav: Loop frequency: 2000 Hz, Feedback gain: 0.4
   
   
   
4. Comb filtering
   • Generation program source: comb.cpp [script]
   • Utilized signal class: Comb
   • Sounds:
     • comb1.wav: 440 Hz fundamental frequency, feedback gain 0.95.
     • comb2.wav: 100 Hz fundamental frequency, feedback gain 0.99.
   
   
   
5. Group Delay: frequency-dependent time delay
   • Generation program source: dispersion.cpp [script]
   • Utilized signal class: Allpass1
   • Sounds: The feedback gain of the allpass1 filter controls the delay of frequencies. See a graph of the group delay, where g is the feedback gain parameter for the filter.
     • dispersion1.wav: feedback gain of 0.7, using 800 allpass1 filters. Filtered 0.25 seconds of whitenoise.
     • dispersion2.wav: feedback gain of 0.5, using 3000 allpass1 filters. Filtered 0.25 seconds of whitenoise.
     • dispersionFrench.wav: feedback gain of 0.99, using 1000 allpass1 filters. Compare to original.
     • dispersionFrench2.wav: feedback gain of 0.5, using 2000 allpass1 filters. Compare to original again.
   
   
   
6. Flanging
   • Generation program source: flanger.cpp, flanger2.cpp, flanger3.cpp [script]
   • Utilized signal class: Delayline, Function
   • Sounds: Filtering 5 seconds of whitenoise as an input:
     • flanger1-1.wav: A slow flange effect.
     • flanger2-1.wav: Several flangers going on at the same time with different oscillation rates.
     • flanger3-1.wav: A flanger with a very fast rate.
   
   
   
7. Onepole filter
   • Generation program source: onepole.cpp [script]
   • Utilized signal class: Onepole
   • Sounds: Using whitenoise as an input:
     • onepole1.wav: feedback gain = -0.999.
     • onepole2.wav: feedback gain = 0.999.
     • onepole3.wav: feedback gain = (0, -0.999; 50, -0.999; 150, -0.8; 180, 0; 200, -0.9999; 250, -0.9)
   
   
   

Effects

---

1. Ring modulation
   • Generation program source: ring.cpp , am-mod.cpp , hybrid-ring.cpp [ring script] [am script] [hybrid script]
   • Utilized signal class: Multiply, Add, Envelope
   • Sounds:
     • ring1.wav: voice modulated by a sinewave with frequency envelope (0 400; 1 800). original.
     • Amplitude modulation is the same as ring modulation: two signals are multiplied together. In classical "ring modulation" one of the signals is usually a single sinewave. In amplitude modulation, there are two sinewaves: one as in ring modulation, and the other is at 0 Hz. Here is the french paperjam voice with amplitude modulation gradually added: am1.wav.
     • If you want to control the amplitude of both the carrier (input sound file) and the modulation (output of ring program), then you can use hybrid-ring which adds an optional envelope to control the amplitude of the input sound in the resulting ring modulated output soundfile: hybrid-ring1.wav. This soundfile crossfades between purely the original sound and purely ring modulation (which doesn't have any of the original frequencies in it).
   
   
   
2. Echos
   • Generation program source: decayEcho.cpp [script]
   • Utilized signal class: Delayline
   • Sounds: as an input:
     • echo1.wav: 3 dB decrease between each echo amplitude, using Ping.wav.
   
   
   
3. Weird echo
   • Generation program source: weirdEcho.cpp [script]
   • Utilized signal class: Delayline
   • Sounds:
     • weirdEcho1.wav
     • weirdEcho2.wav original
   
   
   
4. Amplitude Panning
   • Generation program source: pan.cpp [script]
   • Utilized signal class: Pan
   • Sounds:
     • panLinear.wav Linear interpolation of amplitudes.
     • panPower.wav Constant power as signal is panned between channels.
   
   
   
5. Precedence Panning
   • Generation program source: panprec.cpp [script]
   • Utilized signal class: Delayline Resonator
   • Sounds:
     • panPrec1.wav
     • panPrec2.wav
   
   
   
6. Amplitude and Time Panning
   • Generation program source: pan2.cpp [script]
   • Utilized signal class: Pan2
   • Sounds: Combination of amplitude and timing differences for a more fancy pan, traveling in a line from left side to right.
     • movingRooster.wav
     • movingSine.wav
   
   
   

Synthesis

---

1. White noise
   • Generation program source: whitenoise.cpp [script]
   • Utilized signal class: WhiteNoise
   • Sounds:
     • whitenoise.wav uniformly random samples, amplitude = 0.2 .
     • noise10.wav: random samples which get updated every 10 samples -- causes noise to become lowpassed. amplitude = 0.2 .
   
   
   
2. Additive synthesis
   • Generation program source: additive.cpp [script] [script2]
   • Utilized signal class: Osc
   • Sounds:
     • additive1.wav: Three sinewave oscillators changing pitch.
     • xen.wav: a la Xenakis. 11 oscillators were used to generate a pattern which can be seen visually in this picture.
   
   
   
3. Graphics to Sound conversion
   • Generation program source: line2sine.cpp
   • Utilized signal class: Osc
   • Sounds: The line2sine program converts lines in a Diagram.app file into sinewaves. In the table below are the Diagram pictures and their corresponding sounds.

     example1.gif	example1.wav	example2.gif	example2.wav
     example3.gif	example3.wav	example4.gif	example4.wav
     example5.gif	example5.wav	example6.gif	example6.wav
     example7.gif	example7.wav	example8.gif	example8.wav
     example9.gif	example9.wav	example10.gif	example10.wav
     example11.gif	example11.wav	example12.gif	example12.wav

   
   
   
4. Karplus-Strong algorithm
   • Generation program source: karplus.cpp [script]
   • Utilized signal class: Delayline WhiteNoise
   • Sounds: Using whitenoise as an input:
     • karplus1.wav: 100 sample delayline length. (high pitch)
     • karplus2.wav: 1700 sample delayline length. (low pitch)
   
   
   
5. Fixed Sinewave
   • Generation program source: asinewave.cpp [script]
   • Utilized signal class: FixedSine
   • Sounds: Using whitenoise as an input:
     • sinewave.wav: 800 Hz
   
   
   
6. Silence
   • Generation program source: silence.cpp [script]
   • Utilized signal class: Constant
   • Sounds:
     • silence1.wav: one second of silence.
   
   
   

Psychoacoustics

---

1. JND of Frequency
   • Generation program source: ab-test.cpp [script]
   • Utilized signal class: Osc
   • Sounds: Can you tell which of the sounds below has a higher frequency? Use ab-test to do a blind test on yourself. I correctly identified the correct match 14 out of 15 times (93%) using headphones.
     • first sound
     • second sound
   
   
   
2. Equal-Loudness Curves
   • Generation program source: osc.cpp [script]
   • Utilized signal class: Osc
   • Sounds: Equal loudness curves at various loudness levels from measurement of Craig Sapp's ears, January 1995. Note that loudness is not equal to amplitude. Note that you would have to be in the exactly same sound setup for these to be true to what I heard, and it is difficult to measure the equal loudness curve anyway. Frequency range goes from 20 to 10000 Hz.
     • craigEqualLoud.wav: oscillator sweeping up in frequency for an equal-loudness curve measured at a relatively loud intensity level.
     • craigEqualMid.wav: oscillator sweeping up in frequency for an equal-loudness curve measured at a medium-loud intensity level.
     • craigEqualLow.wav: oscillator sweeping up in frequency for an equal-loudness curve measured near the threshold of hearing intensity level. On headphones for me this sounds pretty constant in loudness, the low end in my speakers is a bit loud and the last bit of high frequencies gets very slightly louder.
     • equalAmp.wav a comparison of an oscillator sweeping up in frequency with a constant amplitude.
   
   
   
3. Masking
   • Generation program source: additive.cpp [script]
   • Utilized signal class: Osc
   • Sounds:
     • masking1.wav: two oscillators. First remains fixed at 440 Hz at an amplitude of 0.7, the other oscillator sweeps from 660 Hz to 880 Hz going from an amplitude of 0.01 to 0.001. In decibels, the sweeping oscillator starts at -37 dB below the 440 Hz masker and goes to -57 dB by 880 Hz. I cannot hear the sweeping oscillator at all.
     • masking2.wav: This soundfile contains only a single oscillator at 440 Hz at 0.7 amplitude. Compare to masking1.wav -- I cannot tell the difference between them.
     • maskingDiff12.wav: Subtracting masking2.wav from masking1.wav removes the 440 tone and leaves the masked sound audible. Try ab-test with the two sounds to see if you can tell the difference between them.
     • masking3.wav: making the masking sound barely audible (with headphones), by sweeping at a constant amplitude of 0.01. I can hear beating especially at the end of the sound.
     • maskingDiff23.wav: subtracting masking2.wav from masking3.wav to leave only the masked oscillator.
   
   
   
4. Echo Experiment
   • Generation program source: echos.cpp [script]
   • Utilized signal class: Delayline
   • Sounds: Determine the minimum timing for perceiving echos. (around 20ms, but depends on the sound, reverb, etc.).
     • echo10ms.wav: 10 millisecond time separation between sounds. The two separate attacks should not be distinguishable.
     • echo30ms.wav: 30 millisecond time separation between sounds. The two separate attacks should be barely distinct.
     • echo50ms.wav: 50 millisecond time separation between sounds. The two separate attacks should be distinct.
   
   
   
5. Minimum periods for pitch
   • Generation program source: periods.cpp [script]
   • Utilized signal class: FixedSine, Multiply
   • Sounds: Here is a test to see how many periods of a sinewave are necessary in order to detect its frequency:
     • Periods: 1, 2, 3, 4, 5, 6, 7, 8, 9, Each sound is windows by a cos*2 to minimize AM noise. You need about 7 periods of a sinewave to detected the correct pitch.
   
   
   
6. Precedence Effect
   • Generation program source: precedence.cpp
   • Utilized signal class: Delayline Impulse
   • Sounds:
     • example 1: clicks (20 samples of white noise) gradually shift from the center to the left, in increments of 23 microsecond delays. By 30 samples the click is clearly in the left ear. Amplitude remains the same in all clicks: precedence1.wav. [script].
     • example 2: This example creates precedence effect soundfiles using two different random noise clicks. For myself, the effect is still there, but the movement to the right ear is at first fast, then at about 45 degrees becomes more gradual, not getting fully into the right ear until about 25 or 30 sample delays. precedence2.wav. [script].
     • example 3: Testing effect of amplitude on the precedence effect. precedence3a.wav reinforces the precedence effect by having the louder sound come first in the left channel. precedence3b.wav tries to counteract the precedence effect by placing the louder sound in the delayed right channel.

       For myself, precedence3a.wav sounds as if the click is always coming from the left ear. Some ambiguousness on the first couple of clicks.

       For myself, precedence3b.wav sounds as if the click starts out in the right ear and gradually moves to the center. Then the clicks start to separate into two separate clicks which can be felt in both ears. [script].

     • Example 4: Using extended length noise in each channel. For this case the noise in each ear is different, but both are white noise. The right channel is gradually delayed from 1 to 30 samples. The sounds remain separate in each ear. precedence4.wav. [script].
     • Example 5: similar signal in each ear with lots of overlap This makes for very strong precedence. precedence5.wav. [script].
   
   
   

Soundfile Utilities

---

1. Concatenate -- appending soundfiles together into one filesoundfiles together into one file.
   • Generation program source: concatenate.cpp [script]
   • Utilized signal class: none
   • Sounds:
     • nextSounds.wav: Mixing most of the NeXT system sounds together into one file with 2000 samples of silence between each original soundfile.
   
   
   
2. Collapsing soundfile channels
   • Generation program source: collapse.cpp [script]
   • Utilized signal class: Add
   • Sounds:
     • Rooster.wav is stereo, and Rooster-mono.wav is the collapsed channel version which is monophonic.
   
   
   
3. Channel extraction
   • Generation program source: extract.cpp [script]
   • Utilized signal class: none
   • Sounds:
     • extract1.wav: the right channel of Rooster.wav.
   
   
   
4. Flipping channels
   • Generation program source: flipChannels.cpp [script]
   • Utilized signal class: none
   • Sounds:
     • original and the flipped version, where the left channels becomes the right channel and vice-versa.
   
   
   
5. Join Channels: Take two mono soundfile and make a stereo file from them.
   • Generation program source: joinMono.cpp [script]
   • Utilized signal class: Delayline
   • Sounds:
     • sine1 + sine2 = output
   
   
   
6. Soundfile mixing
   • Generation program source: mix.cpp [script]
   • Utilized signal class: Add
   • Sounds:
     • Cricket + Rooster (offset 0.1 sec) = Crooster
   
   
   

Miscellaneous

---

1. Frequency Envelopes
   • Generation program source: osc.cpp [script]
   • Utilized signal classes: Osc, Envelope
   • Sounds: single sweeping sinewaves in using both linear and geometric interpolation of the same frequency envelope
     • outLinear.wav: uses the frequency envelope (0 440; 25 440; 26 660; 50 660; 75 330; 90 660; 100 2000) which has data in time-frequency pairs that are interpolated linearly for intervening times.
     • outGeometric.wav: the same frequency envelope, but intervening times are interpolated geometrically.
   
   
   
2. Amplitude envelopes
   • Generation program source: envelope.cpp [script]
   • Utilized signal class: Envelope
   • Sounds:
     • ampEnvelope1.wav: A triangular linear amplitude envelope on a sinewave.
     • ampEnvelope2.wav: A triangular geometric amplitude envelope (exponential) on a sinewave.
     • ampEnvelope3.wav: A 1000 sample linear ramp up at the beginning and ending of a sinewave.
     • ampEnvelope4.wav: Test on a stereo file, envelope applied to each channel; original.
   
   
   

Unsorted

---

1. Binaural beats example: each channel of sound has exactly one frequency in it. The left channel stays constant at a constant frequency of 150 Hz, while the right channel goes from 150 Hz to 152 Hz. Notice that the two sounds start beating. Now listen only to one channel at a time, no beating is heard. The beating is being created inside the auditory cortex.
2. Reverb example: pure comb/allpass reverberation. Probably not working properly yet. out1.wav. [source].
3. Reverb mixed with original (direct signal). Definitely not working yet. out1.wav. out2.wav. [source].