Repository navigation
hisilicon-ev200 / goke-gk7205v200: IMX335's day AE back to the vendor's highlight correction, so the sun in frame no longer halves the exposure - #2694
Merged
Conversation
…'s highlight correction, so the sun in frame no longer halves the exposure The day half's [static_ae] ran HistRatioSlope 168 / MaxHistOffset 32, tighter than the vendor's defaults (128 / 16 in HiISP 开发参考, ISP_AE_ATTR_S), which also say not to exceed 0x100: "a small overexposed area makes AE think the frame is far brighter than the target, so it keeps lowering exposure and the picture goes dark". The night half already runs 128. Measured on a west-facing gk7205v300 + IMX335 in the field, sun in frame, highlight priority, a few minutes apart (widgetii/majestic#1122): exposure AE average shaded ground blown 168 / 32 (current) 67 us 27 36 / 44 3.0% 128 / 16 (this) 133 us 41 50 / 61 5.8% and again an hour later: 128 / 16 at 370 us, AE average 42. With 168 / 32 the AE settles about 1 EV under its own target of 56 as soon as the sun enters the frame. Without a bright light source the pair does not engage, and the picture does not change. Lab hi3516ev300 + IMX335 (behind glass) and gk7205v300 + IMX335, this profile against the current one with day DRC, alternated, tone controller on: mean 120.8-121.8 against 119.8-123.3 and 123.4-123.5 against 122.9-123.7, blown 0.02-0.04% either way, AE Offset 0-3. The night half is unchanged.
widgetii
requested review from
cronyx,
flyrouter and
viktorxda
as code owners
October 26, 2026 14:02
PR Summary by QodoRestore vendor highlight correction for IMX335 daytime AE
AI Description
Diagram
High-Level Assessment
Files changed (1)
|
Code Review by Qodo
1. Shared IMX335 day exposure default changes for deployed cameras
|
![qodo-free-for-open-source-projects[bot]](https://avatars.githubusercontent.com/in/1068000?s=60&v=4){kind=small}
Member
Author
|
Follow-up from the same field camera (west gk7205v300 + IMX335), 2026-10-26 14:45Z. This was dusk: every arm ran at the 30 ms exposure ceiling and AE varied gain instead. The run waited for any anti-flicker hold to clear before measuring and found none.
With #2691's day DRC in place, 128/16 is neutral at dusk: the same picture within noise and no extra blown area. With the sun in frame it doubles the exposure (67 → 133 µs, figures in the description). So the change helps when a bright source is in frame and costs nothing otherwise. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The day half's [static_ae] ran HistRatioSlope 168 / MaxHistOffset 32,
tighter than the vendor's defaults (128 / 16 in HiISP 开发参考,
ISP_AE_ATTR_S), which also say not to exceed 0x100: "a small overexposed
area makes AE think the frame is far brighter than the target, so it keeps
lowering exposure and the picture goes dark". The night half already runs
128.
Measured on a west-facing gk7205v300 + IMX335 in the field, sun in frame,
highlight priority, a few minutes apart (widgetii/majestic#1122):
168 / 32 (current) 67 us 27 36 / 44 3.0%
128 / 16 (this) 133 us 41 50 / 61 5.8%
and again an hour later: 128 / 16 at 370 us, AE average 42. With 168 / 32
the AE settles about 1 EV under its own target of 56 as soon as the sun
enters the frame.
Without a bright light source the pair does not engage, and the picture
does not change. Lab hi3516ev300 + IMX335 (behind glass) and gk7205v300 +
IMX335, this profile against the current one with day DRC, alternated,
tone controller on: mean 120.8-121.8 against 119.8-123.3 and 123.4-123.5
against 122.9-123.7, blown 0.02-0.04% either way, AE Offset 0-3.
The night half is unchanged.