Recent content by videodoctor

@0dan0 the mem read at the following location seems to have a mode table for the new sensor in the Type D units: mem r 0x80dbf440 0x80 dump addr=80dbf440 length=00000080 to console: 80DBF440 : 802CCBB8 802CCB20 00000790 00000434 ..,. .,.....4.. 80DBF450 : 00000790 00000434 00000790 00000434...

I looked around similar offsets to the range of buffers we saw in the ipl getimepath 0 results, and I found 16 buffers with live data, a few before the starting buffer shown in the 656x480 P2 pipeline, and several after. I tried filling all of them but still have a flicker. I'm not going to...

I think the buffers are correct as they all line up with the ipl getimepath buffer addresses---i think the issue is that AE function that you've previously hooked into has changed. After the frame counter is incremented in that native function, there's a modulus check for 3 which exits the AE...

Thank you, @0dan0 ! Also, for the expo_iso value, are there typical values it returns, such as 50, 100, 200, 400, or 800?

@0dan0 , i got the histogram overlay working, but it's flashing a lot, and I tried to add a for() loop in the draw section to make it more steady, but it didn't help much. Here's a recording of it. You'll see I've added little colored rects to the top right of the display to help me debug each...

Thanks for that, @0dan0 . It will be nice, i'm sure, to have a D unit in your hands! The address for frameno in your hist.c doesn't seem to be incrementing during preview or encode. In earlier Claude.ai research, it looked like 80f82464 might be a frame counter, but it's incrementing on boot...

@0dan0, what's your confidence so far with the (now) five Type D variables in your custom mods? imagebase = (uint8_t *)0xa2c7b9b0; //Type D expo_iso = (int *)0x80e5590c; //sensor ISO nvm_base = (int *)0x80E0ADA4; //exposure, sharpness, tint button = (uint32_t *)0xA0E8b7d0; Also, in your...

Thanks for the continued investigation, @0dan0 . I'll see if I can get hist.c working now---I did have it calling during preview (and I think capture), but with the imagebase pointing to the wrong address in Type D, that might be why I saw it crashing. I continued my investigation of sensor and...

You bet, @0dan0 ! Here's the output while recording, with the original stock D firmware: P1: SW:1 Num:6 H:1728 V:1296 LOFS:1728 FMT:102 BUF0 Y0xa2c7b9b0 U0xa2e9e5b0 V0x00000000 BUF1 Y0xa2fafbb0 U0xa31d27b0 V0x00000000 BUF2 Y0xa32e3db0 U0xa35069b0 V0x00000000...

Also, if I use the ipl setimepath command, I'm able to change the preview resolution to 320x240, and visibly changes the UI to show a smaller preview image: > ipl setimepath 0 2 0 1 0 IPL(0) ime path 2 0 1 0 > ipl getimepath 0 P1: SW:0 Num:0 H:-559038737 V:-559038737 LOFS:-559038737...

@0dan0, not sure I posted this before or we specifically discussed this, but this serial console command shows a 656x480 image size: ipl getimepath 0 P1: SW:0 Num:0 H:-559038737 V:-559038737 LOFS:-559038737 FMT:-559038737 SW:0 Cur 0xdeadbeef P2: SW:1 Num:6 H:656 V:480 LOFS:656 FMT:102...

@0dan0 , this is my latest version with just consoleD calls. No crash in this version, but still freezes. Which memory address did you use as a hook to call this code?

With one hand impaired for typing (i got a xiaflex injection in my right pinky finger yesterday and it's wrapped to my ring finger), I can't type as fast so work today was slow. I tried a few different ways to compile your res.c into your latest RBN, but I keep getting stalls or crashes. Here's...

Hmmm... well, here's the code I compiled: #include <stdint.h> //extern void console(const char *ctx, int a, int b); extern void consoleD(const char *ctx, int a, int b); void res_mod(void) { register uintptr_t s1val asm("s1"); uint32_t *scanSize = (uint32_t *)s1val; char*...

@0dan0 , where do i inject the compiled res.c and crop.c ? i presume i should be using your RBN file for these updates?