Log sustained T330W Low and 10-fps parking scan

docs/card-profiles/viofo-t330w.md

@@ -10,10 +10,32 @@ T330 and T330W share the same main-camera family; the attached rear camera is a
 
 ## Card evidence
 
-Four sanitized learning submissions from one owner confirmed `DCIM/Movie`, `DCIM/Movie/Parking`, `DCIM/Movie/RO`, and `DCIM/Photo` with three-channel F/I/R and PF/PI/PR suffixes. The first completed scan (84 files) was originally labeled normal-bitrate driving by the owner, who later corrected the setting name to **Medium**; it also included low-bitrate parking. The second (36 files) was owner-labeled high-bitrate driving and 1-fps Night Vision time-lapse parking. The third (60 files) retained all earlier 36 and added Maximum driving and Auto Event Detection test clips. The fourth (108 files) retained those 60 and added Low driving and 10-fps time-lapse parking clips. These settings can coexist on a card; sample clips by folder/channel and do not extrapolate a setting to all files.
+Five sanitized learning submissions from one owner confirmed `DCIM/Movie`, `DCIM/Movie/Parking`, `DCIM/Movie/RO`, and `DCIM/Photo` with three-channel F/I/R and PF/PI/PR suffixes. The first completed scan (84 files) was originally labeled normal-bitrate driving by the owner, who later corrected the setting name to **Medium**; it also included low-bitrate parking. The second (36 files) was owner-labeled high-bitrate driving and 1-fps Night Vision time-lapse parking. The third (60 files) retained all earlier 36 and added Maximum driving and Auto Event Detection test clips. The fourth (108 files) retained those 60 and added Low driving and 10-fps time-lapse parking clips. The fifth (849 files) adds longer Low/10-fps evidence, but its 120-file media sample cannot establish complete retention of prior files. These settings can coexist on a card; sample clips by folder/channel and do not extrapolate a setting to all files.
 
 The second submission's `RO` parking clips have 30-fps playback, with two recording starts 30 minutes apart and one 60-second clip per channel; this is consistent with 1-fps capture over 30 minutes. The owner reports no audio on those clips, but the sanitized scan does not include audio-track presence. Do **not** label all RO clips as impact events or all RO clips as time lapse: use model-specific media timing/audio context or a confirmed setting. The profile folder mapping stays protected/neutral pending that classification fix.
 
+## Longer Low / 10-fps card scan reviewed 2026-10-11
+
+The owner submitted another **complete 0.1.29 T330W scan** after extended Low driving and 10-fps time-lapse parking. Exact-stamp OCR again matched T330W. The card scan measured **849 copyable files totaling 108,609,992,495 bytes (108.61 decimal GB)**: 774 `DCIM/Movie` driving MP4, 24 `DCIM/Movie/Parking` MP4, 30 `DCIM/Movie/RO` protected MP4, and 21 photos. The owner reports that some photos were taken from impacts while parked; the sanitized scan does not establish the trigger for every photo or `RO` video. The app displays the Parking-folder clips as **“Parking Motion Detection”**, but this is scanner inference and does **not** supersede the owner-selected **10-fps time-lapse** setting.
+
+The preceding completed scan had 108 files. The new scan has a net increase of 741, but **the media-file sample is capped at 120**: 81 of the prior sampled filenames still appear at unchanged sizes; 27 previously sampled driving filenames are absent from the new bounded sample. The aggregate folder counts and capped samples cannot establish that all 108 old files remain. Do not equate a sampled-name omission with deletion, or claim full retention without a complete manifest.
+
+The new scan contains 49 paired video metadata samples, including older setting generations. After excluding already-retained identical tuples and those older generations, **24 distinct additional Low/10-fps pairs** are recorded in both public evidence databases (anonymous numerical digest `eeb4c5bc3b6ce3f8f0e44f5e84fdfbc18ffaa6262b9f9d7b6e939147017f46ac`). They are additional sample records, not necessarily 24 newly recorded unique clips. The private derived submission retains the folder summaries and 120 bounded media samples. All sampled videos report H.264 and nominal 30-fps playback.
+
+| Owner-selected cohort / folder | Channel | New paired samples | Encoded-video bitrate (Mbps) | Paired whole-file sizes (bytes) | Playback durations (s) |
+|---|---|---:|---:|---:|---:|
+| Low driving, `Movie` | front 4K | 3 | 27.032774–27.088368 | 205,520,896 (two full); 27,394,048 (short) | 60 (two); 7.77 |
+| Low driving, `Movie` | interior 2K | 3 | 11.878273–11.882247 | 92,274,688 (two full); 12,713,984 (short) | 60 (two); 7.77 |
+| Low driving, `Movie` | rear 2K | 3 | 11.878314–11.884869 | 92,274,688 (two full); 12,713,984 (short) | 60 (two); 7.77 |
+| 10-fps setting, `Parking` | front 4K | 3 | 27.032976–27.072760 | 205,520,896 (two full); 6,422,528 (short) | 60 (two); 1.57 |
+| 10-fps setting, `Parking` | interior 2K | 3 | 11.878523–11.888993 | 92,274,688 (two full); 4,325,376 (short) | 60 (two); 1.83 |
+| 10-fps setting, `Parking` | rear 2K | 3 | 11.878198–11.953873 | 92,274,688 (two full); 4,325,376 (short) | 60 (two); 1.70 |
+| 10-fps setting, `RO`, trigger unknown | front 4K | 2 | 27.030056–27.037234 | 205,520,896; 159,383,552 | 60; 46.83 |
+| 10-fps setting, `RO`, trigger unknown | interior 2K | 2 | 11.877875–11.890379 | 92,274,688; 71,303,168 | 60; 47.13 |
+| 10-fps setting, `RO`, trigger unknown | rear 2K | 2 | 11.876321–11.877510 | 92,274,688; 71,303,168 | 60; 47.00 |
+
+Complete 60-second triplets still total **390,070,272 bytes**, as in the first Low/10-fps scan. The shortened files are not representative minute rates. No new current or voltage measurement was included; **588 mA average** remains the owner's earlier 10-fps parking observation. Exact capture cadence, `RO` trigger and elapsed parked-time retention are not established by encoded playback metadata alone.
+
 ## Low driving and 10-fps time-lapse scan reviewed 2026-10-11
 
 The owner changed to **Low driving bitrate and 10-fps time-lapse parking** at approximately 07:52 local, without formatting. The completed 0.1.29 scan selected `viofo-t330w` by matched model-stamp OCR and measured **108 files / 108 copyable items**: 72 driving videos, six Parking-folder videos, 18 `RO` protected videos, and 12 photos. **All 60 prior files remain with identical filenames and sizes**; 48 new items comprise 27 driving MP4, six Parking MP4, nine `RO` MP4, and six JPG. The 07:51:16 triplet is still Maximum-sized; the 07:52:16 triplet is a short boundary clip with Maximum-like bitrate. Low-sized triplets begin at 07:53:14. Later Parking and `RO` triplets begin between 08:00:07 and 08:12:07. Do not assign boundary clips to the new Low setting.
@@ -38,7 +60,7 @@ A complete three-channel Low driving triplet uses **390,070,272 bytes per 60 sec
 
 Source: owner-labeled sanitized 0.1.29 scan received 2026-10-11 and retained in the private submission store (anonymous numerical evidence digest `70d8b5b34c973767872de30349d986c1700c2983c809374cbcf2563b05a03419`). It selected `viofo-t330w` through the exact on-video model-stamp OCR on an actual card. The 60-file scan contains 45 driving MP4, nine `RO` MP4, and six JPG. **All 36 files** in the preceding complete scan remain present with unchanged sizes; 24 items were added (18 driving, three `RO`, three JPG). The owner changed High driving to Maximum and 1-fps Night Vision time-lapse to Auto Event Detection at approximately 19:25, without formatting. The last earlier driving triplet starts at 19:24:41 and the first later triplet at 19:26:23. Settings attribution is based on that owner report, clip times, and separated bitrate bands, not on the RO folder alone.
 
-The structured evidence databases `web/recording-time/storage-evidence.json` here and `docs/data/storage-evidence.json` in the camera-profile repository each retain **84 anonymous, paired size/duration/video-bitrate samples**, separated by `settingsLabel` and source digest: 33 from the earlier Medium/low-rate scan, 30 from this High/Maximum/1-fps/Auto Event mixed-card scan, and 21 from the new Low/10-fps scan above. The private submissions retain per-file sizes, and their video summary groups preserve the full sampled bitrate ranges. Numbers below are decimal Mbps and exact whole-file bytes; video-track rates exclude audio, padding, and container overhead. Counts are *sampled clips* per channel, not all 54 videos in that 60-file scan.
+The structured evidence databases `web/recording-time/storage-evidence.json` here and `docs/data/storage-evidence.json` in the camera-profile repository each retain **108 anonymous, paired size/duration/video-bitrate samples**, separated by `settingsLabel` and source digest: 33 from the earlier Medium/low-rate scan, 30 from this High/Maximum/1-fps/Auto Event mixed-card scan, and 21 from the first Low/10-fps scan and 24 additional distinct pairs from the longer Low/10-fps scan above. The private submissions retain per-file sizes, and their video summary groups preserve the full sampled bitrate ranges. Numbers below are decimal Mbps and exact whole-file bytes; video-track rates exclude audio, padding, and container overhead. Counts are *sampled clips* per channel, not all 54 videos in that 60-file scan.
 
 | Owner-reported setting | Channel | Sample count | Video bitrate range (Mbps) | Whole-file size range (bytes) | Encoded playback duration range (s) | Paired 60-s file (bytes) |
 |---|---|---:|---:|---:|---:|---:|

docs/video-quality-master-chart.md

@@ -4,6 +4,10 @@ Quick cross-camera reference for channel layout, measured file metadata, and par
 
 ## T330W mixed-setting bitrate and storage reference (reviewed 2026-10-11)
 
+### Longer 849-file confirmation scan
+
+A later completed T330W scan held 774 driving MP4, 24 Parking MP4, 30 protected `RO` MP4 and 21 photos, totaling **108.61 decimal GB**. It added 24 distinct Low/10-fps paired evidence records after deduplicating repeated numerical samples, for 108 T330W evidence pairs across all setting generations. The complete Low and 10-fps parking triplet size remains **390,070,272 bytes per 60 seconds of playback**; interrupted files have different byte sizes and must not be treated as full loops. The app's “Parking Motion Detection” label conflicts with the owner's 10-fps time-lapse setting and is an inference, not a verified trigger. The bounded 120-file sample cannot prove that every file from the preceding 108-file scan survived. [Detailed sizes, rates and provenance](card-profiles/viofo-t330w.md#longer-low--10-fps-card-scan-reviewed-2026-10-11).
+
 ### New Low and 10-fps time-lapse cohort
 
 A completed 108-file mixed-card scan retained all 60 files from the preceding scan. Owner-labeled **Low driving** clips (excluding the transition clip) measure ~27.03 Mb/s front and ~11.88 Mb/s on each add-on. A full three-channel 60-second file set is **390,070,272 bytes**, or **23.404 GB per encoded hour**. Owner-labeled **10-fps time-lapse parking** yielded two complete triplets in `Parking` and two complete triplets in `RO`, with the same full-file sizes and nominal 30-fps playback; start spacing of three minutes is consistent with 10-fps capture but does not independently verify capture cadence or `RO` trigger. The owner measured **588 mA average** in this parking setting, with voltage unknown. [Exact per-channel bitrates, paired sizes and durations](card-profiles/viofo-t330w.md#low-driving-and-10-fps-time-lapse-scan-reviewed-2026-10-11) and the 21 new numeric samples are retained separately from previous settings.

web/recording-time/COVERAGE.md

@@ -53,7 +53,7 @@ The new T340 four-channel Normal / Low Power scan adds eight encoded-video bitra
 
 ## T330W mixed-setting evidence reviewed 2026-10-11
 
-Retained 84 paired size/duration/video-bitrate samples in `storage-evidence.json` from three completed owner-labeled card scans: 33 Medium/low-rate, 30 mixed High/Maximum/1-fps time-lapse/Auto Event, and 21 new Low-driving/10-fps time-lapse samples. Each record has an explicit `settingsLabel` separating Low, Medium (originally submitted as Normal), High, and Maximum driving from low-bitrate parking, 1-fps Night Vision time-lapse, Auto Event Detection, and the separately unconfirmed low-rate `RO` clips. Both old and new files coexisted on an unformatted card. Representative complete 60-second three-channel driving files total 601,882,624 bytes High or 671,088,640 bytes Maximum. The time-lapse video's nominal 30-fps playback is not its capture cadence; an Auto Event Detection active clip cannot establish wall-clock parking retention. These raw measurements do not automatically add calculator presets. See [reviewed T330W evidence](../../docs/card-profiles/viofo-t330w.md#mixed-settings-card-measurements-reviewed-2026-10-11).
+Retained 108 paired size/duration/video-bitrate samples in `storage-evidence.json` from four completed owner-labeled card scans: 33 Medium/low-rate, 30 mixed High/Maximum/1-fps time-lapse/Auto Event, 21 Low-driving/10-fps time-lapse, and 24 additional distinct Low/10-fps pairs from a longer 849-file scan. These 24 are new numerical evidence records after deduplication, not necessarily 24 distinct newly recorded videos. Each record has an explicit `settingsLabel` separating Low, Medium (originally submitted as Normal), High, and Maximum driving from low-bitrate parking, 1-fps Night Vision time-lapse, Auto Event Detection, and the separately unconfirmed low-rate `RO` clips. Both old and new files coexisted on an unformatted card. Representative complete 60-second three-channel driving files total 601,882,624 bytes High or 671,088,640 bytes Maximum. The time-lapse video's nominal 30-fps playback is not its capture cadence; an Auto Event Detection active clip cannot establish wall-clock parking retention. These raw measurements do not automatically add calculator presets. See [reviewed T330W evidence](../../docs/card-profiles/viofo-t330w.md#mixed-settings-card-measurements-reviewed-2026-10-11).
 
 ## Correction made