Marble / Peridot / Garnet / Gold / rodin Photography | Repository: post #72 — TG.ME

Supported gcams: 9.6
Devices: Poco F7 | Onyx
By @BdrcnAYYDIN and @LibsLabs Team
Version: V2Beta
Usage Status: Should work well overall and be of high quality

Features:
9.6 Based on Google lib

~ Special denoise profile for sensor, optimization according to F7
~ AWB suitable for device model optimized, more realistic colors
~ More detail in low light, however low light power boost ( Attempting to adapt the exposure calculation according to the F7 light sensor)
~ Sharpness (IMX882 0.8µm için optimized: 1.18x attempt )
~ Color detail protection
~ Sensor-based Luma Denoise optimized to preserve sensor detail
~ Highlight Fix
~ Sensor-based controlled sharpness
~ Exposure -0.2 EV counters the sensor's tendency to highlight clipping
" According to my research on the Poco f7 awb model, the cold / green color shift in the interior is therefore feature (1-2-3-4)"
1:) Pseudo-CT AWB Enable
2:) Spatial RGB
3:) Highlight saturation
4:) Output Color Space
" Poco f7 Exposure & HDR Behavior based on my research feature (1-2-3)"
1:) HDR Ratio
2:) TET Tuning
3:) Bracketed TET
" Based on my researchs f7 for portrait - selfie issues (feature 1-2)"
1:) Portrait Matting
2:) Edge threshold improvement
~ Luma settings are better and more consistent
~ Improved contrast
~ Shadow balance is more controlled
~ HDR+ frame stacking stability
" F7 Sensor DNG RAW for some bugs fix: ( feature 1-2-3)

1:) RAW16 Fix
2;) DNG Lossless Compression
3:) JPG Quality %100

🧾 You should definitely read Description :
• Remember that it is in Beta 2 phase, however, if you are sharing in any way, please share the post with a quote or link.
• Remember that it is in the test phase and it is met with satisfaction in any feedback support group.
• If you are using a different device or if you have activated some things in the developer options, do not report
• Remember that there will not be the same results for all my gcam
• You can check the group notes for details.

👥 Don't forget to join us with common sense and tolerance for any error notification, suggestions and more:
💬 @LibsLabs
📣@LibsLabsRepository
[email protected]143.5 MB · SO
January 26, 2026 165 1