Sample report · synthetic runs

Four changes in. One unedited report out.

Two 30-second synthetic MCAP recordings: four designed changes in Target and one unchanged control stream. Prepared with the published Adapter Kit, compared by VeluneOS Compare. Check the report against what we changed.

Ground truth vs report

What we changed, and what Compare reported.

Stream in the synthetic runsWhat we changed in TargetWhat the report saysMatch
/planner/statusRemovedStream 004 (PLANNING_LOCAL) — present in Reference onlymatches
/diagnostics/thermalAdded, 1 HzStream 005 (SYSTEM_DIAGNOSTICS) — present in Target onlymatches
/lidar/points10 Hz → 5 Hz from 10 s to 20 sStream 003 (PERCEPTION_LIDAR) — message interval longer in Target, T+10s → T+20s, returnedmatches
/camera/front/imageSilent from 12 s to 17 sStream 001 (PERCEPTION_CAMERA) — Target recorded no messages, T+12s → T+17s, returnedmatches
/imuUnchanged, 100 Hz (control)Stream 002 (LOCALIZATION_IMU) — no material difference observedmatches

Topic names in the first column come from our generator; they are not in the uploaded file or the report, which name streams by neutral key. Categories in brackets are the ones we assigned in mapping.json. Every common stream kept its topic name, so no pairing needed confirmation. This small, clean case shows the product; it is not a measure of accuracy on real recordings.

The report · unedited service output

Exactly as the service returned it.

Files: report.html · report.md · comparison_report.json · comparison_summary.md · stream_categories.json

VeluneOS Structural Comparison — Engineering Change Review

Presentation mode: SHAREABLE (neutral aliases only)

Result

Result
Structural differences observed for review.
Supported differences
4 (limited evidence: 0)
Inconclusive observations
0
First supported difference
T+0s — Stream 004
Streams
Structural difference observed: 2; No material difference observed: 1; Present in Reference only: 1; Present in Target only: 1
Compared range
29 complete common one-second buckets (T+0s → T+29s)

Structural differences observed

Stream present in Reference only — Stream 004

What
Stream present in Reference only
Stream
Stream 004
When
T+0s → T+29.8s (observed messages)
Magnitude
150 messages in Reference
After the window
Not applicable
Evidence
Comparable: the recorded evidence on both sides supports this observation.
ID
candidate:STREAM_PRESENCE_CHANGE:stream.unmapped.004:0-29

Stream present in Target only — Stream 005

What
Stream present in Target only
Stream
Stream 005
When
T+0s → T+29s (observed messages)
Magnitude
30 messages in Target
After the window
Not applicable
Evidence
Comparable: the recorded evidence on both sides supports this observation.
ID
candidate:STREAM_PRESENCE_CHANGE:stream.unmapped.005:0-29

Recorded message interval longer in Target — Stream 003

What
Recorded message interval longer in Target
Stream
Stream 003
When
T+10s → T+20s (buckets 10–19)
Magnitude
Median interval Reference 100.0 ms → Target 200.0 ms (×2.00); messages in the window 100 → 50
After the window
Returned to the Reference pattern at T+20s
Evidence
Comparable: the recorded evidence on both sides supports this observation.
ID
candidate:CADENCE_DIVERGENCE:INTERVAL_INCREASE:stream.unmapped.003:10-19

Target recorded no messages while Reference did — Stream 001

What
Target recorded no messages while Reference did
Stream
Stream 001
When
T+12s → T+17s (buckets 12–16)
Magnitude
Reference messages in the window: 100; Target: 0
After the window
Returned to the Reference pattern at T+17s
Evidence
Comparable: the recorded evidence on both sides supports this observation.
ID
candidate:ACTIVITY_LOSS:stream.unmapped.001:12-16

Streams

StreamObserved statusFindingsCadence assessableRecorded-time comparabilitySmall count differences (max |Δ|)
Stream 001Structural difference observed1yesCOMPARABLE5 (20)
Stream 002No material difference observed0yesCOMPARABLE0 (0)
Stream 003Structural difference observed1yesCOMPARABLE10 (5)
Stream 004Present in Reference only1non/a0 (0)
Stream 005Present in Target only1non/a0 (0)

"No material difference" is bounded to the listed streams and the compared buckets. It is not a statement about other streams, other periods, or system health.

What was not compared

Run durations: Reference 29.99 s, Target 29.99 s (difference +0.0 ms).

Messages after the last complete common bucket were not compared on 3 stream(s). Differences there, including at the end of a recording, cannot be established from this comparison (per-stream counts in Level 3).

Level 3 — Evidence trace

Complete evidence with stable IDs. Sections expand on click.

Evidence trace: structural events
Event IDTypeStream keyBucketsPatternEvidence qualitySupporting evidence IDs
event:STREAM_PRESENCE_CHANGE:stream.unmapped.004:0-29STREAM_PRESENCE_CHANGEstream.unmapped.0040–29NOT_APPLICABLECOMPARABLEpresence:REFERENCE_ONLY:stream.unmapped.004
event:STREAM_PRESENCE_CHANGE:stream.unmapped.005:0-29STREAM_PRESENCE_CHANGEstream.unmapped.0050–29NOT_APPLICABLECOMPARABLEpresence:TARGET_ONLY:stream.unmapped.005
event:CADENCE_DIVERGENCE:INTERVAL_INCREASE:stream.unmapped.003:10-19CADENCE_DIVERGENCEstream.unmapped.00310–19RECOVEREDCOMPARABLEcadence:INTERVAL_INCREASE:stream.unmapped.003:10-19
event:ACTIVITY_LOSS:stream.unmapped.001:12-16ACTIVITY_LOSSstream.unmapped.00112–16RECOVEREDCOMPARABLEreview:stream.unmapped.001:12, review:stream.unmapped.001:13, review:stream.unmapped.001:14, review:stream.unmapped.001:15, review:stream.unmapped.001:16
Cadence evidence (both directions)
Evidence IDStream keyDirectionBucketsTriggersBaseline side / medianWindow median ReferenceWindow median TargetMessages Ref → TgtRecovered at bucket
cadence:INTERVAL_INCREASE:stream.unmapped.003:10-19stream.unmapped.003INTERVAL_INCREASE10–19mean_intervalreference 100.0 ms100.0 ms200.0 ms100 → 5020
Cadence windows kept but not asserted

None.

Recorded-time constraints

None.

Recorded-time comparability per stream
Stream keyStateReasonsShared-stamp buckets Ref / TgtInterval grid Ref / TgtMedian interval Ref / Tgt
stream.unmapped.001COMPARABLEnone0/29 / 0/241 ms / 1 ms50.0 ms / 50.0 ms
stream.unmapped.002COMPARABLEnone0/29 / 0/291 ms / 1 ms10.0 ms / 10.0 ms
stream.unmapped.003COMPARABLEnone0/29 / 0/291 ms / 1 ms100.0 ms / 100.0 ms
Coverage per stream
Stream keyMessages compared Ref / TgtMessages after range Ref / TgtFirst message Ref / TgtLast message Ref / TgtTail
stream.unmapped.001580 / 48020 / 20T+0s / T+0sT+29.9s / T+29.9sINSUFFICIENT_COVERAGE
stream.unmapped.0022900 / 2900100 / 100T+0s / T+0sT+30.0s / T+30.0sINSUFFICIENT_COVERAGE
stream.unmapped.003290 / 24010 / 10T+0s / T+0sT+29.9s / T+29.9sINSUFFICIENT_COVERAGE
Bucket-level count differences (15; largest first, not promoted)

Display order only; the report JSON keeps the contract order. Count-only differences are not promoted to findings.

Evidence IDStream keyBucketReferenceTargetΔType
review:stream.unmapped.001:12stream.unmapped.00112200-20reference_nonempty_target_empty
review:stream.unmapped.001:13stream.unmapped.00113200-20reference_nonempty_target_empty
review:stream.unmapped.001:14stream.unmapped.00114200-20reference_nonempty_target_empty
review:stream.unmapped.001:15stream.unmapped.00115200-20reference_nonempty_target_empty
review:stream.unmapped.001:16stream.unmapped.00116200-20reference_nonempty_target_empty
review:stream.unmapped.003:10stream.unmapped.00310105-5message_count_difference
review:stream.unmapped.003:11stream.unmapped.00311105-5message_count_difference
review:stream.unmapped.003:12stream.unmapped.00312105-5message_count_difference
review:stream.unmapped.003:13stream.unmapped.00313105-5message_count_difference
review:stream.unmapped.003:14stream.unmapped.00314105-5message_count_difference
review:stream.unmapped.003:15stream.unmapped.00315105-5message_count_difference
review:stream.unmapped.003:16stream.unmapped.00316105-5message_count_difference
review:stream.unmapped.003:17stream.unmapped.00317105-5message_count_difference
review:stream.unmapped.003:18stream.unmapped.00318105-5message_count_difference
review:stream.unmapped.003:19stream.unmapped.00319105-5message_count_difference
Provenance and method
Report schema
velune.structural_comparison_report 0.5.0 (private_cloud_only)
Bundle ID
vcib_sha256_5173b991aecffc7d48af2b9fb83cc9a5ef1402fb963ad27c16652adf5fd034df
Bundle contract
velune.comparison_input_bundle 0.1.0
Mapping profile
generic.mcap.structural 0.1.0
Generated at
2026-10-16T02:54:15.179751+00:00
Comparison rule
complete common one-second buckets on each run's own time axis
Candidate review order
candidate:STREAM_PRESENCE_CHANGE:stream.unmapped.004:0-29, candidate:STREAM_PRESENCE_CHANGE:stream.unmapped.005:0-29, candidate:CADENCE_DIVERGENCE:INTERVAL_INCREASE:stream.unmapped.003:10-19, candidate:ACTIVITY_LOSS:stream.unmapped.001:12-16

This report records observed structural differences between two recordings. A matching run-relative time does not claim the same physical moment. Direction words describe the recordings, not improvement or degradation. No cause, fault, severity, or safety judgment is made.

Find the events. Engineers find the cause.

Embedded from samples/compare-sample-report.html without changes to its text; heading levels and the main landmark are adjusted to fit this page. Open the original file for the exact document.