Property tests by davidirvine · Pull Request #43 · davidirvine/deepnote

davidirvine · 2025-07-23T23:20:27Z

No description provided.

- Implement real-time processing validation * Single voice processing speed verification (sub-100ms for 1s audio) * Multiple voice scaling tests (1-16 voices with CPU usage monitoring) * Real-time constraint validation for different voice counts - Add memory allocation stability tests * Long-duration processing without dynamic allocation * Consistent memory usage across multiple transitions * Amplitude and frequency bounds verification during extended processing - Implement stress testing * Rapid parameter changes with fast animation speeds * Extreme oscillator count testing (up to 16 oscillators) * High-complexity scenarios with detuning and rapid transitions - Add performance regression detection * Baseline performance measurement with detailed timing statistics * Per-sample processing time analysis (average, median, 95th percentile, max) * Performance consistency across different voice states - Fix duplicate symbol issue with inline specifier for init_voice function - Update CMakeLists.txt to include performance_tests.cpp - All tests pass with 606,406 assertions validating performance characteristics - Tests maintain C++14 compatibility and realistic performance expectations

- Implement frequency transition properties validation * Random parameter testing with 50 test cases across musical frequency range * Monotonicity property verification for ascending frequency transitions * Adaptive timeouts based on animation speed for reliable completion - Add extreme parameter value testing * Very low frequencies (20-40Hz sub-bass range) stability validation * Very high frequencies (8-18kHz treble range) without aliasing issues * Extreme frequency ratio testing (32x increases/decreases across 5+ octaves) * Error tolerance validation within 10% for extreme transitions - Implement Bezier curve parameter edge cases * Boundary value testing for control points (0.0-1.0 range) * Extreme control point combinations (linear, reverse, near-extremes) * Animation multiplier testing (0.5x to 10x speed variations) * Graceful handling of edge curve configurations - Add detuning edge case validation * Extreme detune values (0Hz to ±100Hz) stability testing * Negative detuning support verification * Detuning behavior during frequency transitions with amplitude variation * Beat pattern validation with reasonable amplitude bounds - Implement sample rate variation testing * Multi-rate support (8kHz to 192kHz) stability validation * Sample rate and frequency relationship verification * Proportional timing behavior across different sample rates - Update CMakeLists.txt to include property_edge_tests.cpp - All tests pass with 1,958,075 assertions validating comprehensive edge coverage - Property-based testing with random parameter generation for robust validation - Realistic bounds and tolerances for practical audio applications

Copilot

Pull Request Overview

This PR adds comprehensive property-based testing and performance validation for the deepnote voice system. The tests focus on validating voice behavior across the entire parameter space and ensuring real-time performance requirements are met.

Property-based testing with random parameter generation for frequency transitions, edge cases, and extreme parameter values
Performance testing to validate real-time processing requirements and memory stability
Edge case validation for extreme frequencies, Bezier control points, detuning values, and sample rates

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 6 comments.

File	Description
test/property_edge_tests.cpp	New comprehensive property-based tests covering frequency transitions, extreme parameters, Bezier curves, detuning, and sample rate variations
test/performance_tests.cpp	New performance validation tests for real-time processing, memory stability, and stress testing
test/CMakeLists.txt	Updated to include the new test files in the build
src/voice/deepnotevoice.hpp	Made init_voice function inline for better test performance

Comments suppressed due to low confidence (1)

test/property_edge_tests.cpp:34

[nitpick] The variable name 'test' is ambiguous. Consider renaming to 'test_iteration' or 'test_case' to be more descriptive.

        for(int test = 0; test < num_tests; ++test) {

Copilot · 2025-07-23T23:21:42Z

+        std::uniform_int_distribution<int> osc_count_dist(1, 8);
+        std::uniform_real_distribution<float> speed_dist(0.5f, 5.0f);
+
+        const int num_tests = 50; // Property-based testing with multiple random cases


The magic number 50 should be defined as a named constant to improve maintainability and make it easier to adjust test coverage.

Suggested change

const int num_tests = 50; // Property-based testing with multiple random cases

const int num_tests = NUM_TESTS; // Property-based testing with multiple random cases

Copilot · 2025-07-23T23:21:42Z

+            // Property: Final frequency should be close to target
+            float final_freq = voice.get_current_frequency().get();
+            float freq_error = std::abs(final_freq - target_freq);
+            REQUIRE(freq_error < 5.0f); // Should be within 5Hz


The magic number 5.0f should be defined as a named constant (e.g., MAX_FREQUENCY_ERROR_HZ) to improve maintainability and make the tolerance configurable.

Suggested change

REQUIRE(freq_error < 5.0f); // Should be within 5Hz

REQUIRE(freq_error < MAX_FREQUENCY_ERROR_HZ); // Should be within 5Hz

Copilot · 2025-07-23T23:21:43Z

+
+            // Allow some tolerance for numerical precision, but should be mostly monotonic
+            float non_monotonic_ratio = static_cast<float>(non_monotonic_count) / frequency_progression.size();
+            REQUIRE(non_monotonic_ratio < 0.1f); // Less than 10% non-monotonic steps


The magic number 0.1f should be defined as a named constant (e.g., MAX_NON_MONOTONIC_RATIO) to improve maintainability and make the tolerance configurable.

Suggested change

REQUIRE(non_monotonic_ratio < 0.1f); // Less than 10% non-monotonic steps

REQUIRE(non_monotonic_ratio < MAX_NON_MONOTONIC_RATIO); // Less than 10% non-monotonic steps

Copilot · 2025-07-23T23:21:43Z

+        INFO("Processing time for 1 second of audio: " << duration.count() << "ms");
+
+        // Should process 1 second of audio in much less than 1 second (real-time requirement)
+        REQUIRE(duration.count() < 100); // Less than 100ms for 1 second of audio


The magic number 100 should be defined as a named constant (e.g., MAX_PROCESSING_TIME_MS) to improve maintainability and make the performance threshold configurable.

Suggested change

REQUIRE(duration.count() < 100); // Less than 100ms for 1 second of audio

REQUIRE(duration.count() < MAX_PROCESSING_TIME_MS); // Less than 100ms for 1 second of audio

Copilot · 2025-07-23T23:21:43Z

+        REQUIRE(avg_time_ns < sample_period_ns * 0.1); // Use only 10% of available time
+        REQUIRE(p95_time_ns < sample_period_ns * 0.2); // 95% of samples under 20% of time
+        REQUIRE(max_time_ns < sample_period_ns * 3.0); // Even worst case should be reasonable (allow for OS jitter)


The magic numbers 0.1, 0.2, and 3.0 should be defined as named constants (e.g., CPU_USAGE_TARGET, CPU_USAGE_P95_LIMIT, CPU_USAGE_MAX_FACTOR) to improve maintainability.

Suggested change

REQUIRE(avg_time_ns < sample_period_ns * 0.1); // Use only 10% of available time

REQUIRE(p95_time_ns < sample_period_ns * 0.2); // 95% of samples under 20% of time

REQUIRE(max_time_ns < sample_period_ns * 3.0); // Even worst case should be reasonable (allow for OS jitter)

REQUIRE(avg_time_ns < sample_period_ns * CPU_USAGE_TARGET); // Use only 10% of available time

REQUIRE(p95_time_ns < sample_period_ns * CPU_USAGE_P95_LIMIT); // 95% of samples under 20% of time

REQUIRE(max_time_ns < sample_period_ns * CPU_USAGE_MAX_FACTOR); // Even worst case should be reasonable (allow for OS jitter)

Copilot · 2025-07-23T23:21:43Z

+
+            // Should not exceed reasonable CPU usage even with many voices
+            // Scale expectations based on voice count
+            double max_cpu_percent = voice_count <= 8 ? 50.0 : 80.0; // More lenient for many voices


The magic numbers 8, 50.0, and 80.0 should be defined as named constants to improve maintainability and make the thresholds configurable.

Suggested change

double max_cpu_percent = voice_count <= 8 ? 50.0 : 80.0; // More lenient for many voices

double max_cpu_percent = voice_count <= MAX_VOICE_COUNT_THRESHOLD

? LOW_CPU_USAGE_THRESHOLD

: HIGH_CPU_USAGE_THRESHOLD; // More lenient for many voices

+    std::mt19937 gen(rd());
+}
+
+TEST_CASE("Frequency transition properties") {


+    }
+}
+
+TEST_CASE("Extreme parameter values") {


+    }
+}
+
+TEST_CASE("Bezier curve parameter edge cases") {


+    }
+}
+
+TEST_CASE("Detuning edge cases") {


+    }
+}
+
+TEST_CASE("Sample rate variations") {


davidirvine added 2 commits July 23, 2025 14:27

Copilot AI review requested due to automatic review settings July 23, 2025 23:20

Copilot AI reviewed Jul 23, 2025

View reviewed changes

Merge branch 'develop' into property-tests

f7ceae3

github-advanced-security AI found potential problems Jul 23, 2025

View reviewed changes

davidirvine closed this Jul 23, 2025

davidirvine deleted the property-tests branch July 23, 2025 23:35

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Property tests#43

Property tests#43
davidirvine wants to merge 3 commits intodevelopfrom
property-tests

davidirvine commented Jul 23, 2025

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Jul 23, 2025

Uh oh!

Copilot AI Jul 23, 2025

Uh oh!

Copilot AI Jul 23, 2025

Uh oh!

Copilot AI Jul 23, 2025

Uh oh!

Copilot AI Jul 23, 2025

Uh oh!

Copilot AI Jul 23, 2025

Uh oh!

Check notice

Copilot Autofix

Check notice

Copilot Autofix

Check notice

Copilot Autofix

Check notice

Copilot Autofix

Check notice

Copilot Autofix

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

	const int num_tests = 50; // Property-based testing with multiple random cases
	const int num_tests = NUM_TESTS; // Property-based testing with multiple random cases

	REQUIRE(freq_error < 5.0f); // Should be within 5Hz
	REQUIRE(freq_error < MAX_FREQUENCY_ERROR_HZ); // Should be within 5Hz

	REQUIRE(non_monotonic_ratio < 0.1f); // Less than 10% non-monotonic steps
	REQUIRE(non_monotonic_ratio < MAX_NON_MONOTONIC_RATIO); // Less than 10% non-monotonic steps

	REQUIRE(duration.count() < 100); // Less than 100ms for 1 second of audio
	REQUIRE(duration.count() < MAX_PROCESSING_TIME_MS); // Less than 100ms for 1 second of audio

-            double max_cpu_percent = voice_count <= 8 ? 50.0 : 80.0; // More lenient for many voices
+            double max_cpu_percent = voice_count <= MAX_VOICE_COUNT_THRESHOLD
+                                     ? LOW_CPU_USAGE_THRESHOLD
+                                     : HIGH_CPU_USAGE_THRESHOLD; // More lenient for many voices

Conversation

davidirvine commented Jul 23, 2025

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Copilot AI Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jul 23, 2025

Choose a reason for hiding this comment

Uh oh!

Check notice

Uh oh!

Copilot Autofix

Check notice

Uh oh!

Copilot Autofix

Check notice

Uh oh!

Copilot Autofix

Check notice

Uh oh!

Copilot Autofix

Check notice

Uh oh!

Copilot Autofix

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants