This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TI2A problem?

I am running RDK2.8 on Appro DM8127IPNC.

I got rid of system_server, let ipnc_rdk_mcfw.out run stand alone. If I turn off AEWB, the video looks fine. If I enable TI2A, the video shows up very briefly, then goes to totally black. I dumped some debug output fro ae_ti.c. Is there any extra setup for AEWB that's done in system_server that I am missing?

[m3vpss ] ae: awb gains = 1024, 1024, 1024, numPixels = 60, 16, 32
 [m3vpss ] ae: ipipe gainss = 2048, 5000, 2
 [m3vpss ] min_y = 404
 [m3vpss ] R = 23, G = 33, B = 23
 [m3vpss ] curY before dgain = 30
 [m3vpss ] currentY = 30, avgY = 30
 [m3vpss ] ae is not converged
 [m3vpss ] ae frames = 0, curY = 30, avgy = 330, target = 40, thrld = 10
 [m3vpss ] sat_sum = 30, 6144000, un_sat_sum = 30, 6144000
 [m3vpss ] numHistory = 11, adjRatio = 1365
 [m3vpss ] ae: adjRatio here = 4
 [m3vpss ] we are here exposure time 0, 16667, 5000
 [m3vpss ] adjRatio, again = 4, 1000
 [m3vpss ] nextExposure: 5000, curExposure:5000
 [m3vpss ] nextGain: 1000, curGain:2
 [m3vpss ] nextIGain: 1024, curIGain:2048

[m3vpss ] ae: awb gains = 1024, 1024, 1024, numPixels = 60, 16, 32
 [m3vpss ] ae: ipipe gainss = 1024, 5000, 1000
 [m3vpss ] min_y = 126
 [m3vpss ] R = 7, G = 10, B = 7
 [m3vpss ] curY before dgain = 9
 [m3vpss ] currentY = 9, avgY = 28
 [m3vpss ] ae is not converged
 [m3vpss ] ae frames = 1, curY = 9, avgy = 309, target = 40, thrld = 10
 [m3vpss ] sat_sum = 9, 6144000, un_sat_sum = 9, 6144000
 [m3vpss ] numHistory = 11, adjRatio = 4551
 [m3vpss ] ae: adjRatio here = 4110
 [m3vpss ] we are here exposure time 20000, 16667, 5000
 [m3vpss ] adjRatio, again = 1024, 1209
 [m3vpss ] nextExposure: 16600, curExposure:5000
 [m3vpss ] nextGain: 1209, curGain:1000
 [m3vpss ] nextIGain: 1024, curIGain:1024


[m3vpss ] ae: awb gains = 1024, 1024, 1024, numPixels = 60, 16, 32
 [m3vpss ] ae: ipipe gainss = 1024, 16600, 1209
 [m3vpss ] min_y = 383
 [m3vpss ] R = 22, G = 31, B = 22
 [m3vpss ] curY before dgain = 28
 [m3vpss ] currentY = 28, avgY = 27
 [m3vpss ] ae is not converged
 [m3vpss ] ae frames = 2, curY = 28, avgy = 307, target = 40, thrld = 10
 [m3vpss ] sat_sum = 28, 6144000, un_sat_sum = 28, 6144000
 [m3vpss ] numHistory = 11, adjRatio = 1462
 [m3vpss ] ae: adjRatio here = 1701
 [m3vpss ] we are here exposure time 27400, 16667, 5000
 [m3vpss ] adjRatio, again = 1024, 1661
 [m3vpss ] nextExposure: 16600, curExposure:16600
 [m3vpss ] nextGain: 1661, curGain:1209
 [m3vpss ] nextIGain: 1024, curIGain:1024


  • When further investigating the problem, I found that the filter_flag in IssAlg_capt2AProcessTI is set as IPIPE_RGB_TO_YUV_FLAG, which doesn't make sense because following code was setting rgb_rgb1 coefficients. After changing the flag to IPIPE_RGB_RGB1_FLAG, I can see live video now. Color is a little bit off, but almost there.