| 00:37 | rton | left the channel |
| 01:01 | futarisIRCcloud | joined the channel |
| 01:26 | davidak | joined the channel |
| 01:51 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 03:06 | illwieckz | left the channel |
| 03:07 | illwieckz | joined the channel |
| 03:22 | BAndiT1983 | changed nick to: BAndiT1983|away
|
| 05:36 | davidak | left the channel |
| 05:36 | davidak | joined the channel |
| 05:38 | davidak | left the channel |
| 05:38 | davidak | joined the channel |
| 06:18 | davidak | left the channel |
| 06:21 | futarisIRCcloud | left the channel |
| 06:43 | seku | left the channel |
| 07:01 | pusle | left the channel |
| 07:02 | pusle | joined the channel |
| 07:23 | XDjackieXD | left the channel |
| 07:29 | XDjackieXD | joined the channel |
| 08:11 | tushar1997 | left the channel |
| 08:31 | Anant | joined the channel |
| 08:31 | tushar1997 | joined the channel |
| 08:36 | Anant | Hello My name is Anant I am Gsoc 2018 aspirant. I am interested in the project real time peak focusing. I have some knowledge on VHDL and doing a steep learning of image processing. I am new to this, please tell how should I follow I wen
|
| 08:36 | Bertl_oO | welcome Anant!
|
| 08:37 | Bertl_oO | you already managed the first step by showing up here :)
|
| 08:37 | Bertl_oO | you also want to do the VHDL challenge
|
| 08:37 | Anant | I went through some papers and same to know that CMOS camera sensor is being used. My task in vhdl would be
|
| 08:38 | Anant | To implement line buffer that carries image data right?
|
| 08:38 | tushar1997 | left the channel |
| 08:39 | Bertl_oO | yes, line buffer, a kernel to detect the gradient and some logic to make the peaking visible
|
| 08:40 | Anant | So should I use peak detection algorithm to detect peaks...?
|
| 08:41 | Bertl_oO | yes?
|
| 08:42 | Anant | And about the sobel pipeline, does that means to detect edges through sobel edge detection
|
| 08:43 | Bertl_oO | that is the idea, if you have a better one, no problem
|
| 08:46 | Anant | left the channel |
| 08:48 | tushar1997 | joined the channel |
| 08:49 | tushar1997 | left the channel |
| 08:50 | tushar1997 | joined the channel |
| 08:50 | tushar1997 | left the channel |
| 08:50 | tushar1997 | joined the channel |
| 08:54 | Anant | joined the channel |
| 08:55 | Anant | I am just assuring if I am correct. As I am very new to this. please tell m
|
| 08:55 | tushar1997 | left the channel |
| 08:55 | tushar1997 | joined the channel |
| 08:55 | Bertl_oO | most likely you are correct :)
|
| 08:56 | Anant | Please tell me from where to start learning it's implementation I have already started looking into the qualification task
|
| 08:57 | Bertl_oO | well, there are a bunch of good VHDL 'courses' online
|
| 08:57 | Bertl_oO | also check the coding style used in the existing AXIOM Beta VHDL codebase
|
| 08:59 | Bertl_oO | for the image processing part I wouldn't worry too much right now, that is something we will discuss in detail during GSoC
|
| 08:59 | Anant | Okay thank you very much
|
| 08:59 | Anant | left the channel |
| 09:00 | rton | joined the channel |
| 09:03 | tushar1997 | left the channel |
| 09:03 | tushar1997 | joined the channel |
| 09:06 | Anant | joined the channel |
| 09:08 | Anant | Sorry Bertl_oO I am unable to find the vhdl codebase for Axiom beta can you please send me the link
|
| 09:09 | Bertl_oO | https://github.com/apertus-open-source-cinema/beta-software
|
| 09:12 | Anant | Thanks again : )
|
| 09:12 | Anant | left the channel |
| 09:49 | se6astian|away | changed nick to: se6astian
|
| 10:13 | futarisIRCcloud | joined the channel |
| 10:44 | LordVan | joined the channel |
| 11:25 | niemand | joined the channel |
| 11:25 | niemand | left the channel |
| 11:25 | niemand | joined the channel |
| 12:55 | se6astian | changed nick to: se6astian|away
|
| 13:28 | LordVan | left the channel |
| 13:59 | se6astian|away | changed nick to: se6astian
|
| 14:41 | pusle | left the channel |
| 14:43 | pusle | joined the channel |
| 15:47 | se6astian | changed nick to: se6astian|away
|
| 16:14 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 18:04 | BAndiT1983 | changed nick to: BAndiT1983|away
|
| 18:04 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 18:20 | davidak | joined the channel |
| 18:36 | davidak | left the channel |
| 18:36 | BAndiT1983 | changed nick to: BAndiT1983|away
|
| 19:14 | davidak | joined the channel |
| 19:32 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 21:21 | TofuLynx | joined the channel |
| 21:21 | arpu | joined the channel |
| 21:28 | TofuLynx | Hello! I am an Informatics Engineering student which heard about GSoC yesterday and I saw on the list Apertus. Due to my interests in image/video related programming and making films as a hobby, I loved your project! I wanted to start here to have a general idea about what to do first and how to do it, and to ask some questions.
|
| 21:29 | Bertl_oO | welcome! and ask away!
|
| 21:30 | TofuLynx | Thank you :) My strengths are mainly C++ and I think that's where I am more suited to collaborate, and I saw the C++ challenge and the C++ related ideas (T720, T763 and T722)
|
| 21:32 | TofuLynx | I think my approach should be reading the related documentations about AXIOM Beta, the Collected Information
|
| 21:33 | Bertl_oO | that's a good start, yes
|
| 21:33 | TofuLynx | Was there some thing like AXIOM alpha? or other project?
|
| 21:33 | Bertl_oO | also it's good that you found your way here and hanging around here might be beneficial as well
|
| 21:33 | Bertl_oO | yes, the AXIOM Alpha was the proof of concept prototype we did
|
| 21:34 | TofuLynx | I am happy I am off to a good start
|
| 21:35 | TofuLynx | And was the proof of concept before or after you were kickstarted?
|
| 21:35 | Bertl_oO | before
|
| 21:35 | Bertl_oO | and it was actually indiegogo
|
| 21:36 | TofuLynx | Oh I see
|
| 21:37 | TofuLynx | That's a great accomplishment!
|
| 21:37 | TofuLynx | and AXIOM Beta is based on arch linux
|
| 21:37 | TofuLynx | is it important that its software use low level instructions?
|
| 21:38 | Bertl_oO | hmm?
|
| 21:38 | TofuLynx | I mean, processor specific code
|
| 21:39 | Bertl_oO | you mean native code or what's the question here?
|
| 21:40 | TofuLynx | for example, SIMD
|
| 21:40 | Bertl_oO | ah, we don't currently use the NEON instructions
|
| 21:41 | Bertl_oO | the ARM cores are mostly used for simple tasks like configuration and control
|
| 21:41 | TofuLynx | control, like, image previewing and stuff?
|
| 21:42 | Bertl_oO | mostly adjusting (sensor) registers and similar
|
| 21:42 | TofuLynx | I see, and where does image processing occurs?
|
| 21:43 | Bertl_oO | the image pipeline is implemented in the FPGA
|
| 21:45 | TofuLynx | Thanks for the information! I am going to read about the ideas and the challenge
|
| 21:47 | Bertl_oO | you're welcome!
|
| 22:05 | TofuLynx | How many hours of continuous filming is a single AXIOM's battery expected to take?
|
| 22:06 | Bertl_oO | depends on the battery :)
|
| 22:06 | TofuLynx | Ah xD
|
| 22:07 | TofuLynx | But isn't there any reference or something regarding this topic?
|
| 22:07 | TofuLynx | like a duration with a specified battery
|
| 22:07 | Bertl_oO | not yet
|
| 22:11 | TofuLynx | Are there any special programming practices related to the development board?
|
| 22:12 | Bertl_oO | like?
|
| 22:13 | TofuLynx | well, just making sure if there arent some advises related with it
|
| 22:17 | Bertl_oO | nothing special
|
| 22:17 | niemand | left the channel |
| 22:18 | TofuLynx | "Bonus points if you use an existing debayering method to recover a color image.", what do you mean by existing? Already used before in anything, not related to apertus?
|
| 22:19 | Bertl_oO | any good debayering will do :)
|
| 22:47 | TofuLynx | the debayering process should be programmed by me or is it preferable to use some foss library?
|
| 22:51 | BAndiT1983 | usually by you, to show us that you understand the stuff
|
| 22:51 | TofuLynx | Yeah I see
|
| 22:51 | BAndiT1983 | it can be asimple one which interpolates between 2 or 4 values, depends on the case
|
| 22:52 | TofuLynx | but I could use a more advanced one with pattern recognizing and stuff?
|
| 22:52 | TofuLynx | or is it overkill?
|
| 22:53 | BAndiT1983 | it is not the goal of the challenge
|
| 22:53 | TofuLynx | Ok I see
|
| 22:53 | BAndiT1983 | the challenge is to see if candidates for gsoc can develop an application up to the task
|
| 22:54 | TofuLynx | Ok Ok
|
| 22:54 | BAndiT1983 | no fancy stuff should be done, just plain and simple conversion from 12 bit stream and output of the image, possibly in colour, but liniearizing not required
|
| 22:54 | TofuLynx | what do you mean by liniearizing?
|
| 22:56 | BAndiT1983 | usually raw images have to be upscaled in the color range from 12 bit to 16 bit, it depends on the colorspace they were shoot in, but this stuff is too advanced for now
|
| 22:58 | TofuLynx | so, for example, if the raw image green pixel has a value of 4096 (2^12), it should be 65536 in the final image?
|
| 22:58 | Bertl_oO | off to bed now ... have a good one everyone!
|
| 22:58 | Bertl_oO | changed nick to: Bertl_zZ
|
| 22:58 | TofuLynx | Good Night!
|
| 23:01 | BAndiT1983 | a sort of, there are several methods, and one should be aware of clipping if the values rise too high
|
| 23:01 | TofuLynx | Ok! :)
|
| 23:01 | BAndiT1983 | https://photo.stackexchange.com/questions/72505/what-does-linearization-mean-in-raw-processing
|
| 23:05 | TofuLynx | I am understanding it now! makes sense
|
| 23:05 | TofuLynx | how do I measure the sensor response?
|
| 23:06 | BAndiT1983 | can't answer on the fly, my knowledge of that stuff is also not that deep
|
| 23:07 | TofuLynx | Thanks!
|
| 23:07 | BAndiT1983 | off for today
|
| 23:07 | TofuLynx | See you next time!
|
| 23:23 | BAndiT1983 | changed nick to: BAndiT1983|away
|
| 23:31 | rton | left the channel |
| 23:42 | anuejn | TofuLynx: you can play a bit with debayering algorithms with https://anuejn.github.io/batic/
|
| 23:42 | TofuLynx | Hey!
|
| 23:42 | TofuLynx | Thanks!
|