Public Version
www.ti.com
Camera ISP Functional Description
NOTE:
When A-Law compression in CCDC is enabled during dark-frame write, it must be enabled
when the stored dark frame is used.
6.4.7.1.6 Camera ISP VPBE Preview Dark-Frame Subtract or Shading Compensation
The preview engine can fetch a dark frame containing 8-bit values from memory (8 bits in input are
converted internally to 10 bits by adding two zeros to the left) and subtracting it, pixel by pixel, from the
incoming input frame (
[6] DRKFEN). This function removes pattern noise in the sensor. The
output of the dark-frame subtract operation is 10 bits wide (U10Q0).
There must be adequate memory bandwidth if this feature is enabled. If the data fetched from memory
arrives late, the
[31] DRK_FAIL status bit is set to indicate a fail.
Instead of performing the dark-frame subtract, the preview engine can perform lens-shading compensation
(if
[21] SCOMP_EN is set along with
[6] DRKFEN). In this case, the 8-bit unsigned
value fetched from memory is multiplied by the incoming pixel, and the result is right-shifted by the number
of bits specified by the
[24:22] SCOMP_SFT parameter (0-7 bits).
The SBL data read port is shared between the CCDC and preview module. The read port must be
affected to the preview module by writing 0 into the
[28] SBL_SHARED_RPORTB.
Programmers must ensure that the CCDC module does not use this port before switching to the preview
module.
6.4.7.1.7 Camera ISP VPBE Preview Horizontal Median Filter
The preview engine contains a horizontal median filter that can help reduce temperature-induced noise
effects. The input and output of the horizontal median filter are 10 bits wide (U10Q0).
NOTE:
Line-width reduction: If the horizontal median filter is enabled, the preview engine reduces
the length of the output line of this stage by 4 pixels (2 starting pixels -left edge and 2 ending
pixels -right edge). For example, if the input size is 656 × 490 pixels, the output is 652 × 490
pixels. There is no truncation of input data line if this block is disabled.
6.4.7.1.8 Camera ISP VPBE Preview Noise Filter and Faulty Pixel Correction
The noise filter and couplet defect correction (CDC) operate on the same 33 matrix of same color pixels.
Both functions can be enabled separately using the
[27] DCOREN and
[9] NFEN
bits.
The faulty pixel correction function replaces the central pixel (x0) with one of the neighbors (x1-x8) in the
following cases:
•
When it has been identified as faulty.
•
When the feature is enabled.
The noise filter modifies the central pixel when it is enabled.
NOTE:
Some details of these features are not available in public domain
6.4.7.1.9 Camera ISP VPBE Preview White Balance
The white-balance module has a digital gain adjuster and a white-balance adjuster. In the digital gain
adjuster (
[9:0] DGAIN), RAW data is multiplied by a fixed-value gain, regardless of the
color of the pixel to be processed.
In the white-balance gain adjuster (
), RAW data is multiplied by a selected gain
corresponding to the color of the processed pixel.
NOTE:
Some details of this feature are not available in public domain
1209
SWPU177N – December 2009 – Revised November 2010
Camera Image Signal Processor
Copyright © 2009–2010, Texas Instruments Incorporated
