Understand the failure.
Verify the repair.
Small, reproducible software failures. The broken implementation, the fix that didn’t work, and the one that passed—preserved together.
Explore the cases ↓How results are verified ↗WHAT THE ARCHIVE CONTAINS
100840 executable cases. 20168 are open.
Every case records the implementation that fails, the fix that did not work, and the repair that passed its checks—with recorded outputs and source hashes. This release adds 100840 cases across 20168 failure mechanisms and 254 domains.
The open tier gives you the failure and the unsuccessful fix for one case in every mechanism. The remaining 80672 cases, 5 variants per mechanism, are member-only: the verified repair, its recorded checks, and the full fixture suite are held in the member archive. Read the methodology ↗
A RECORD OF WHAT WENT WRONG
Browse the archive / 100840
Pcm error feedback quantizer: restored amplitude · case 01
Feedback treats output code as original-domain amplitude.
Pcm error feedback quantizer: restored amplitude · case 02
Feedback treats output code as original-domain amplitude.
Pcm error feedback quantizer: restored amplitude · case 03
Feedback treats output code as original-domain amplitude.
Pcm error feedback quantizer: restored amplitude · case 04
Feedback treats output code as original-domain amplitude.
Pcm error feedback quantizer: restored amplitude · case 05
Feedback treats output code as original-domain amplitude.
Pcm error feedback quantizer: reset timing · case 01
Packet reset clears noise-shaping error after its target sample.
Pcm error feedback quantizer: reset timing · case 02
Packet reset clears noise-shaping error after its target sample.
Pcm error feedback quantizer: reset timing · case 03
Packet reset clears noise-shaping error after its target sample.
Pcm error feedback quantizer: reset timing · case 04
Packet reset clears noise-shaping error after its target sample.
Pcm error feedback quantizer: reset timing · case 05
Packet reset clears noise-shaping error after its target sample.
Pcm error feedback quantizer: quantizer midpoint · case 01
Error-feedback conversion uses floor instead of nearest midpoint bins.
Pcm error feedback quantizer: quantizer midpoint · case 02
Error-feedback conversion uses floor instead of nearest midpoint bins.
Pcm error feedback quantizer: quantizer midpoint · case 03
Error-feedback conversion uses floor instead of nearest midpoint bins.
Pcm error feedback quantizer: quantizer midpoint · case 04
Error-feedback conversion uses floor instead of nearest midpoint bins.
Pcm error feedback quantizer: quantizer midpoint · case 05
Error-feedback conversion uses floor instead of nearest midpoint bins.
Pcm subtractive dither: noise phase · case 01
A deterministic PCM dither stream starts at the wrong sample phase.
Pcm subtractive dither: noise phase · case 02
A deterministic PCM dither stream starts at the wrong sample phase.
Pcm subtractive dither: noise phase · case 03
A deterministic PCM dither stream starts at the wrong sample phase.
Pcm subtractive dither: noise phase · case 04
A deterministic PCM dither stream starts at the wrong sample phase.
Pcm subtractive dither: noise phase · case 05
A deterministic PCM dither stream starts at the wrong sample phase.
Pcm subtractive dither: prequantizer noise · case 01
Dither is omitted before PCM precision reduction.
Pcm subtractive dither: prequantizer noise · case 02
Dither is omitted before PCM precision reduction.
Pcm subtractive dither: prequantizer noise · case 03
Dither is omitted before PCM precision reduction.
Pcm subtractive dither: prequantizer noise · case 04
Dither is omitted before PCM precision reduction.
Pcm subtractive dither: prequantizer noise · case 05
Dither is omitted before PCM precision reduction.
Pcm subtractive dither: subtractive restoration · case 01
Decoded PCM retains deterministic dither instead of subtracting it.
Pcm subtractive dither: subtractive restoration · case 02
Decoded PCM retains deterministic dither instead of subtracting it.
Pcm subtractive dither: subtractive restoration · case 03
Decoded PCM retains deterministic dither instead of subtracting it.
Pcm subtractive dither: subtractive restoration · case 04
Decoded PCM retains deterministic dither instead of subtracting it.
Pcm subtractive dither: subtractive restoration · case 05
Decoded PCM retains deterministic dither instead of subtracting it.
Pcm subtractive dither: quantizer step · case 01
Dithered amplitude is emitted without division into PCM code bins.
Pcm subtractive dither: quantizer step · case 02
Dithered amplitude is emitted without division into PCM code bins.
Pcm subtractive dither: quantizer step · case 03
Dithered amplitude is emitted without division into PCM code bins.
Pcm subtractive dither: quantizer step · case 04
Dithered amplitude is emitted without division into PCM code bins.
Pcm subtractive dither: quantizer step · case 05
Dithered amplitude is emitted without division into PCM code bins.
Pcm subtractive dither: wire and restored · case 01
Wire code and restored amplitude fields are swapped.
Pcm subtractive dither: wire and restored · case 02
Wire code and restored amplitude fields are swapped.
Pcm subtractive dither: wire and restored · case 03
Wire code and restored amplitude fields are swapped.
Pcm subtractive dither: wire and restored · case 04
Wire code and restored amplitude fields are swapped.
Pcm subtractive dither: wire and restored · case 05
Wire code and restored amplitude fields are swapped.
Pcm triangular dither: independent difference · case 01
Two supplied uniform draws are added rather than differenced for centered dither.
Pcm triangular dither: independent difference · case 02
Two supplied uniform draws are added rather than differenced for centered dither.
Pcm triangular dither: independent difference · case 03
Two supplied uniform draws are added rather than differenced for centered dither.
Pcm triangular dither: independent difference · case 04
Two supplied uniform draws are added rather than differenced for centered dither.
Pcm triangular dither: independent difference · case 05
Two supplied uniform draws are added rather than differenced for centered dither.
Pcm triangular dither: noise amplitude · case 01
Dither amplitude is used as a DC bias instead of gain.
Pcm triangular dither: noise amplitude · case 02
Dither amplitude is used as a DC bias instead of gain.
Pcm triangular dither: noise amplitude · case 03
Dither amplitude is used as a DC bias instead of gain.
Pcm triangular dither: noise amplitude · case 04
Dither amplitude is used as a DC bias instead of gain.
Pcm triangular dither: noise amplitude · case 05
Dither amplitude is used as a DC bias instead of gain.
Pcm triangular dither: highpass sign · case 01
Highpass dither adds consecutive noise samples.
Pcm triangular dither: highpass sign · case 02
Highpass dither adds consecutive noise samples.
Pcm triangular dither: highpass sign · case 03
Highpass dither adds consecutive noise samples.
Pcm triangular dither: highpass sign · case 04
Highpass dither adds consecutive noise samples.
Pcm triangular dither: highpass sign · case 05
Highpass dither adds consecutive noise samples.
Pcm triangular dither: delay feedback · case 01
Dither delay stores filtered rather than raw noise.
Pcm triangular dither: delay feedback · case 02
Dither delay stores filtered rather than raw noise.
Pcm triangular dither: delay feedback · case 03
Dither delay stores filtered rather than raw noise.
Pcm triangular dither: delay feedback · case 04
Dither delay stores filtered rather than raw noise.
Pcm triangular dither: delay feedback · case 05
Dither delay stores filtered rather than raw noise.
Pcm triangular dither: continuation noise · case 01
Packet continuation resets the dither delay.
Pcm triangular dither: continuation noise · case 02
Packet continuation resets the dither delay.
Pcm triangular dither: continuation noise · case 03
Packet continuation resets the dither delay.
Pcm triangular dither: continuation noise · case 04
Packet continuation resets the dither delay.
Pcm triangular dither: continuation noise · case 05
Packet continuation resets the dither delay.
Pcm feedback saturation: clipping detection · case 01
Representable endpoint codes are incorrectly flagged as clipping.
Pcm feedback saturation: clipping detection · case 02
Representable endpoint codes are incorrectly flagged as clipping.
Pcm feedback saturation: clipping detection · case 03
Representable endpoint codes are incorrectly flagged as clipping.
Pcm feedback saturation: clipping detection · case 04
Representable endpoint codes are incorrectly flagged as clipping.
Pcm feedback saturation: clipping detection · case 05
Representable endpoint codes are incorrectly flagged as clipping.
Pcm feedback saturation: error reset on clip · case 01
Saturated PCM retains unbounded overload error that contaminates following samples.
Pcm feedback saturation: error reset on clip · case 02
Saturated PCM retains unbounded overload error that contaminates following samples.
Pcm feedback saturation: error reset on clip · case 03
Saturated PCM retains unbounded overload error that contaminates following samples.
Pcm feedback saturation: error reset on clip · case 04
Saturated PCM retains unbounded overload error that contaminates following samples.
Pcm feedback saturation: error reset on clip · case 05
Saturated PCM retains unbounded overload error that contaminates following samples.
Pcm feedback saturation: candidate feedback · case 01
Current PCM conversion ignores accumulated error.
Pcm feedback saturation: candidate feedback · case 02
Current PCM conversion ignores accumulated error.
Pcm feedback saturation: candidate feedback · case 03
Current PCM conversion ignores accumulated error.
Pcm feedback saturation: candidate feedback · case 04
Current PCM conversion ignores accumulated error.
Pcm feedback saturation: candidate feedback · case 05
Current PCM conversion ignores accumulated error.
Pcm feedback saturation: clamp order · case 01
Input-domain amplitude is clamped directly as output code.
Pcm feedback saturation: clamp order · case 02
Input-domain amplitude is clamped directly as output code.
Pcm feedback saturation: clamp order · case 03
Input-domain amplitude is clamped directly as output code.
Pcm feedback saturation: clamp order · case 04
Input-domain amplitude is clamped directly as output code.
Pcm feedback saturation: clamp order · case 05
Input-domain amplitude is clamped directly as output code.
Pcm feedback saturation: error source · case 01
Unclipped error calculation discards prior feedback contribution.
Pcm feedback saturation: error source · case 02
Unclipped error calculation discards prior feedback contribution.
Pcm feedback saturation: error source · case 03
Unclipped error calculation discards prior feedback contribution.
Pcm feedback saturation: error source · case 04
Unclipped error calculation discards prior feedback contribution.
Pcm feedback saturation: error source · case 05
Unclipped error calculation discards prior feedback contribution.
Pcm onebit pulse density: accumulate level · case 01
A bounded PCM-to-one-bit density adapter discards fractional accumulated level each sample.
Pcm onebit pulse density: accumulate level · case 02
A bounded PCM-to-one-bit density adapter discards fractional accumulated level each sample.
Pcm onebit pulse density: accumulate level · case 03
A bounded PCM-to-one-bit density adapter discards fractional accumulated level each sample.
Pcm onebit pulse density: accumulate level · case 04
A bounded PCM-to-one-bit density adapter discards fractional accumulated level each sample.
Pcm onebit pulse density: accumulate level · case 05
A bounded PCM-to-one-bit density adapter discards fractional accumulated level each sample.
Pcm onebit pulse density: threshold inclusive · case 01
An exact pulse threshold is delayed by one sample.
Pcm onebit pulse density: threshold inclusive · case 02
An exact pulse threshold is delayed by one sample.
Pcm onebit pulse density: threshold inclusive · case 03
An exact pulse threshold is delayed by one sample.
Pcm onebit pulse density: threshold inclusive · case 04
An exact pulse threshold is delayed by one sample.
Pcm onebit pulse density: threshold inclusive · case 05
An exact pulse threshold is delayed by one sample.
INSPECTABLE BY DESIGN
Every result has a runnable source.
Runnable implementations with recorded outputs, source hashes, and explicit contracts. Related variants share a failure mechanism and belong together in evaluation splits.
Read the methodology ↗