#apertus IRC Channel Logs

the non-ascii names can probably be addressed with a simple sed script (will look into that soon) ... the bus problem needs to be investigated and addressed though

off to bed now ... have a good one everyone!

good day!

Meeting shortly.

I only host the RPGs.

Good evening everyone :D

Sebastian might be a little bit late because i showed him the Board Game Backpack on IndieGoGo and he went berserk.

"The most comfortable & convenient game bag ever created to protect your expensive board games on the go!"

haha..great!

I am doing great! What about you? :)

Yeah, however not the best semester I have had regarding marks

I've had similar experience last semester. You'll cover up :D

What about yours?

Uni is going to start from 2nd August , so I guess I'll have to cover up a lot of things. But things seems pretty exciting as this sem we'll have DSA and database management etc

Mine starts in october

We've 2 months summer breaks here, starts from May 26 and End on August 1.

AI*

Sorry, mistaken for IA in portuguese xD

interesting :D

Also here the summer breaks are from late june to september, inclusive

Yup! That's why I love summers! I tend to apply what I learned from the university

I won't start my project yet. Only after GSoC :P So I don't get lost in focus

What about you? Any project?

Sebastian is in emails with the makers of Boardgame Backpack. Can you PM me if you intend to speak during today's meeting please?

(That's the usual routine I think)

Okk Sure :P

supragya are you present?

not yet @supragya

yes I am RexOrCine

felix_ ?

Yes, sure!

although the formula I used is good for JPG based tools but it had some problems when it comes to raw data because while using gamma correction, white balance etc etc we are losing data and it’s quite probable to get overexposure.

a1ex and I talked about our approach yesterday and we are trying to use different formula to calculate chrominance and luminance (which isn’t using much color science).

So, I guess this week we’ll try to implement the discussed approach and it sound better than what I did earlier :)

Good stuff.

hi

Hi rahul__ ... OK TofuLynx next.

So, last week my goal was to finish the implementation of nearest neighbor in OpenCL

As I could pinpoint where the errors would come from

I made a wrong approach, I was thinking about a one kernel fits all the cases. But it's not the best approach.

although it increased the size of the kernels.cl file, it was a lot easier to maintain and to code on the .cpp side. And probably more efficient as well

I finished today the refactorization, I made another six different kernels for it, one for each case. It helped a lot. In fact, it was so easy to use that I could reuse the kernels for the border demosaicing in Bilinear Interpolation.

And that's all, folks!

Good work.

here now

hi

hi

No problem, se6astian :)

so should I pitch in....? if it's okay?

so, it's quite short.... we finally got a DNG running off of the MLV files we generated synthetically

it was just a single frame... but a few changes to the mlv structs made it possible to make it AXIOM specific

they can still be added still

today, I encoded multiple frames in that and started integrations with the emulator

se6astian: the last screenshot by g3gg0 was the best output so far

we still need a color matrix info and wbal info

for PLR blocks... they need changes in mlv_dump and mlvfs

https://github.com/apertus-open-source-cinema/misc-tools-utilities/blob/master/raw2dng/raw2dng.c#L46

you can take some bits like color matrix from raw2dng code probably

what does PLR mean? Is that tech used for HDR?

Ok :D

great, many thanks

I would love to tweet something about your progress, TofuLynx do you have a file produced by OpenCL or otherwise a screenshot of the OpenCL output?

supragya maybe a screenshot from a converted raw12 in the the MLV player?

se6astian: not really sure... I don't use MLV player directly.. I convert to DNG and see in darktable

yes but showing a DNG is darktable is not really doing all the stuff that went into it justice :)

great

thanks

let the integrations be stable a bit

I am still a bit confused on how the HDMI(TMDS ) block demands pixel stream from the Memory as there is not much documentation available and its not clear from the code, but Bertl helped me through the other stuffs as well so ill contact him soon regarding this as well.

Once I understand this concept then It would be easier for me to know whether extra delay blocks are required in parallel with my IP or not in the output pipeline.

Thats all from my side

thanks!

ArunM: you around?

nothing relevant from my side

Nothing really to report my side. Just documentation stuff. Presently building up the parts archive. Sebastian and i visited an equipment rental house last week and it's really good. They're 30 seconds walk from the office and it should mean that we're able to compare the camera against other cameras quickly and easily. Also should have access to a wide range of lenses. Tomorrow we're meeting with manfred to shoot some last clips for TT14.2 (the arti

TT14.3 it will be actually

many thanks everyone for attenting

and i joined too late :)

hi :D

newborn didnt want to sleep

@se6astian: do you have a color matrix for the sensor?

ah ok nm, supragya just sent me the link

well, maybe that the wisdom tooth extraction i had ~1,5 weeks ago went well

good to hear, still :)

good night

n8