class ClockSyncPipelineTest

For a signal change occurring at frame T, how far BEFORE that frame will the effects of that

change be reflected in the output. We use no effects; this comes from SincSampler only.

For a signal change occurring at frame T, how far AFTER that frame will the output reflect some

effect of the previous signal. We use no effects; this comes from SincSampler only.

Maximum number of frames needed for a transition between two adjacent signals. At the beginning

of this interval, the output begins to reflect the new signal; only at the end of this interval

is the full effect shown. During this interval, the output is a cross-fading mixture of the

preceding signal and the new signal. We use no effects; this comes from SincSampler only.

These are SOURCE frames, but rates are so near unity that we safely use them interchangeably.

Offset of the first audio sample. This should be greater than TotalRampFrames() so that there

is silence and then transitional frames at the start of the output, following by the signal.

These are SOURCE frames, but rates are so near unity that we safely use them interchangeably.

Capture the ring buffer and rotate it leftward by the given offset, so the output starts at [0]

Return the index of the peak sample, relative to the first frame in the slice.

Verify that the clock for this renderer is running at the expected rate

Send two impulses separated by frames_between_impulses, using a reference clock with the given

slew. The output should contain two impulses separated by NumFramesOutput.

This test validates that time is correctly translated between the two clocks.

This test validates the following, with two 1-frame impulses during clock synchronization:

A, The impulses are peak-detected in the output, with expected magnitudes;

B. The impulse-to-impulse interval is the expected number of frames;

C. The renderer clock is running at the expected rate.

All measurements use tolerance ranges except where explicitly stated as exact.

Send a flat signal (step function) of size num_frames_input, using a reference clock with the

given slew. The output should contain an equivalent step function of size NumFramesOutput.

Note, the exact values are not important. The primary goal of this test is to ensure the output

does not have any dropped frames. A buggy mixer might drop frames if there is a gap between mix

calls, specifically when the destination clock is running faster than the source clock.

This test validates the following, rendering a step function during clock synchronization:

A. The output step signal starts at the expected frame;

B. The output step signal has the expected magnitude for its entirety (no dropouts);

C. The output step signal ends at the expected frame;

D. Subsequent output signal (after PostRampFrames) is precisely zero;

E. The renderer clock is running at the expected rate.

All measurements use tolerance ranges except where explicitly stated as exact.

Send a sine wave using a clock with given slew. The output should be a sine wave at slewed

frequency. Each sinusoidal period contains (num_frames_to_analyze / input_freq) frames.

This test validates the following, rendering a sinusoid during clock synchronization:

A. The output signal's magnitude is essentially unattenuated (within tolerance);

B. The output signal's center frequency is shifted by exactly the expected amount;

C. No other frequencies exceed the noise floor threshold (with a few exceptions);

D. The above-noise-floor frequencies are clustered around the primary output frequency;

E. The width of that cluster (from leftmost to rightmost) is below a certain "peak width";

F. The renderer clock is running at the expected rate (within a certain tolerance).