To create intentional aliasing on a Mac, use Logic Pro’s Bitcrusher effect on a generated tone. Apple explicitly documents artificial aliasing through sample-rate division, making it a supported alternative when Audacity’s resampling removes the frequencies you wanted to fold back. Bitcrusher documentation.

This walkthrough is for an Audacity user who wants to generate a synthetic tone, deliberately produce frequency foldover, and save the result for further editing. It requires Logic Pro for Mac. The settings below are an illustrative experiment derived from official documentation; the predicted frequencies are calculations, not reported listening tests.

Why changing the sample rate in Audacity gives different results

Aliasing occurs when a sampling process encounters frequencies beyond its representable range. The Nyquist frequency is half the sample rate. Anti-alias filtering removes frequencies that would otherwise fold into the lower-frequency range, so a properly filtered conversion can eliminate the very material you wanted to turn into an effect. Audacity’s Digital Audio Fundamentals.

Audacity also distinguishes changing a track’s rate from resampling its audio. Its track-menu documentation explains that changing Rate spreads existing samples farther apart or squeezes them together, changing duration and pitch. Tracks > Resample changes the sample count while preserving duration and playback speed. Neither operation is documented there as a deliberate aliasing effect. Audio Track Dropdown Menu.

The converter-quality preferences serve another purpose: they control conversion quality during playback and when rendering or exporting. They do not document a switch that disables anti-alias filtering. Lowering a quality setting therefore provides no documented, predictable recipe for this task. Use a processor explicitly designed to generate aliasing instead. Quality Preferences.

Generate a controlled sine tone in Logic Pro

Start with one sine wave so that the experiment has a single input frequency. A complex waveform would introduce more components to track.

Set up the project and channel

In a new Logic project, choose File > Project Settings > Audio and set Sample Rate to 48 kHz. Keep that project rate throughout the experiment. Apple documents this setting separately from effects processing and advises against changing it after adding audio files. Set the sample rate of a project.

Create an audio track with Option-Command-N, choose Audio, then Mic or Line. Deselect Load Default Patch to start without effects, leave Input Monitoring and Record Enable off, choose the main stereo output, and click Create. The source will be generated internally. Create tracks in a Logic Pro for Mac project.

Add the generator

Click an empty Audio Effect slot on the track’s channel strip and choose Utility > Test Oscillator. Select Test Tone, choose the sine waveform, double-click Frequency, and enter 7000 for a 7,000 Hz tone. Set Level to an illustrative starting value of −18 dB. Apple documents direct frequency entry from 1 Hz to 22 kHz. Test Oscillator documentation.

The generator starts immediately in Test Tone mode; a MIDI note or recorded region is unnecessary for generating the signal. Bypassing Test Oscillator stops its tone. This behavior also means that stopping the transport is not the generator’s documented off switch. Test Oscillator documentation.

Apply sample-rate reduction that deliberately creates aliasing

In the next empty effect slot, choose Distortion > Bitcrusher. Keep it below Test Oscillator, so the generator feeds the processor. Logic’s insert effects operate in series, passing the output of one processor into the next. Bitcrusher; Route audio via insert effects.

Use these starting settings to concentrate on sample-rate reduction:

Control Starting setting Purpose
Drive 0 dB Avoid added input gain
Resolution 24 bits Minimize bit-depth distortion
Clip Level 0 dB Keep the chosen tone below the clipping threshold
Mix 100% Examine the processed signal alone
Downsampling Begin at 1x, then select 8x Compare the baseline with rate reduction

These are proposed experiment settings based on Apple’s documented controls. Bitcrusher parameters.

Apple defines 1x as unchanged, 2x as half the original sample rate, and higher factors as further division. Its statement that downsampling does not change playback speed or pitch describes the absence of a speed-change operation; the effect still intentionally creates aliased frequency content. Bitcrusher documentation.

Predict the foldover before changing settings

For an illustrative 48,000 Hz project, dividing the rate by eight gives an effective sampling rate of 6,000 Hz and a Nyquist boundary of 3,000 Hz. The 7,000 Hz input is above that boundary. Under ideal unfiltered sampling, its corresponding baseband alias is 1,000 Hz: subtract one effective sampling rate, giving 7000 − 6000 = 1000. This applies the sampling boundary described in Audacity’s fundamentals documentation.

The same calculation provides a useful settings map:

Downsampling factor Effective rate Nyquist boundary Ideal baseband frequency for a 7,000 Hz input
1x 48,000 Hz 24,000 Hz 7,000 Hz
2x 24,000 Hz 12,000 Hz 7,000 Hz
4x 12,000 Hz 6,000 Hz 5,000 Hz
8x 6,000 Hz 3,000 Hz 1,000 Hz
10x 4,800 Hz 2,400 Hz 2,200 Hz

These are calculated ideal-sampling predictions, not a specification of every spectral component or its amplitude in Apple’s implementation. For the 4x case, the input exceeds the 6,000 Hz boundary by 1,000 Hz, so it reflects to 5,000 Hz. For 10x, subtracting 4,800 Hz leaves 2,200 Hz, which lies below the new boundary.

The practical decision rule is to choose a source frequency above the processor’s effective Nyquist boundary while keeping it below the project’s original boundary. Here, 7,000 Hz is below the project’s 24,000 Hz boundary but above the 3,000 Hz boundary at 8x. That makes it a useful input for separating source generation from the subsequent aliasing operation.

Keep the source frequency fixed while changing the factor. Otherwise, two variables move at once, making the result harder to interpret. The table also shows why increasing the reduction factor need not produce a steadily descending alias: foldover depends on the relationship between the tone and the effective sampling rate.

Save the effect and inspect it in Audacity

Choose File > Bounce > Project or Section. Set an explicit, short range with Start and End; do not rely on the absence of recorded regions to define the duration. For this experiment, use Realtime, set Normalize to Off, and deselect Include Audio Tail. Apple specifically warns that the test oscillator can make a bounce excessively long when audio-tail inclusion is enabled. Bounce a project to an audio file.

In the current bounce options, select Uncompressed, choose WAVE, set Sample Rate to 48 kHz, choose 24-bit depth, and use Interleaved for a single stereo file. These are output-file settings; retain the project rate used for the calculations. Uncompressed bounce options.

Save one baseline file with Bitcrusher bypassed and another with the proposed processing enabled. Logic allows bypassing a plug-in from its slot or the plug-in window header, and project bouncing includes the active effects. Use distinct filenames so both versions remain available. Add, remove, move, and copy plug-ins; Project bouncing.

Open each file in Audacity, select a steady section away from its boundaries, and choose Analyze > Plot Spectrum. Use the Spectrum algorithm and the same settings for both files. A larger Size improves frequency resolution; if Audacity reports that the selection is too short, select more audio or reduce the size. Plot Spectrum.

Use the calculated table as a hypothesis: compare the baseline’s input-frequency peak with the processed file’s lower-frequency content. Additional peaks do not automatically invalidate the experiment, because the table predicts ideal baseband locations rather than the complete processor output. A spectrum comparison is more informative than judging aliasing solely from the waveform’s appearance.

What to check when the result is unclear

Work through the signal path before trying more extreme settings:

  • Confirm that Test Oscillator produces the baseline tone with the processor bypassed.
  • Confirm that the generator appears before the processor. Drag plug-ins between slots if necessary.
  • Bypass only the processor for the comparison; bypassing the generator removes the source.
  • Confirm that both files use the same bounce range and output rate.

Logic documents moving and bypassing effects in its plug-in management instructions.

For a separate waveform-generation experiment, bypass the processor, select Square Wave in Test Oscillator, and turn Anti Aliased off. Apple explicitly offers aliased and anti-aliased versions of Square Wave and Needle Pulse. That switch applies to those waveforms; it is not documented as a general anti-aliasing switch for sine generation or project resampling. Test Oscillator documentation.

Use the sine-through-processor workflow when you want a controllable foldover calculation and an exported effect. Use the oscillator’s waveform option when you want to explore aliased waveform generation directly. In either case, retain the baseline and the settings alongside the rendered file so the result can be reproduced and edited further.

Sources