Current Server Time: 23:43 (Central Europe)

#apertus IRC Channel Logs

2019/03/29

Timezone: UTC


00:05
RexOrCine
changed nick to: RexOrCine|away
03:42
aSobhy
left the channel
05:43
rish
joined the channel
05:43
rish
left the channel
05:43
rish_
joined the channel
05:50
rish_
left the channel
06:01
BAndiT1983|away
changed nick to: BAndiT1983
06:50
BAndiT1983
changed nick to: BAndiT1983|away
06:55
Bertl_zZ
changed nick to: Bertl
06:55
Bertl
morning folks!
08:37
siddhantsahu
morning!
09:17
sebix
joined the channel
09:24
futarisIRCcloud
left the channel
10:51
mappie
joined the channel
10:54
mappie
hey guys and mentor i'm interested in white balancing algorithm for opencine. i have already done project "underwater images and video enhancement using fusion process in python" . can i still get in touch with mentor ? sorry for the late approach
10:54
mappie
**mentors
11:05
mappie
left the channel
11:05
saurabh_raj
joined the channel
11:05
mappie
joined the channel
11:06
saurabh_raj
Hi.. any mentors around?
11:13
Bertl
hey saurabh_raj! what's up?
11:26
saurabh_raj
I tried the C++ qualification task, I was hoping mentors could review my code any provide some suggestions on what should I improve upon
11:26
Bertl
BAndiT1983|away is the right person for C++
11:28
saurabh_raj
Ok.. Sure.. Thank you... I will reach out to him when he's around.
11:32
mappie
left the channel
11:45
apurvanandan[m]
Hello, Bertl
11:50
apurvanandan[m]
I have completed the Task 1 of vhdl challenge. Pll is the only primitive I am using for generating 300 Mhz clock. I implemented DDR at link ends and it is working fine. Only problem problem i am seeing is that the reset signal needs to be de-asserted only at the clock edges or else the FFs go into metstate.
11:51
apurvanandan[m]
If it is not a problem then I can start work on my proposal
11:51
Bertl
okay, sounds good, how did you test?
11:52
apurvanandan[m]
It is working on zynq xc7z020-3clg484 on post route simulation
11:52
apurvanandan[m]
With 10 ns clock input from test bench to the pll in top module
11:54
Bertl
where can I take a look at the code?
11:54
Bertl
and is it properly formatted? :)
11:54
apurvanandan[m]
Yeah , i will do that properly and then send you the link.
11:54
Bertl
excellent!
11:55
apurvanandan[m]
Reset is not a problem then?
11:55
Bertl
well, you can make it synchronous with a register, no?
11:56
Bertl
i.e. I do not see a reason for a ansynchronous reset
11:56
apurvanandan[m]
Ok fine, i will do that and send you the code.
11:57
apurvanandan[m]
Sorry for delaying (semester load)
11:57
Bertl
no problem, take your time!
12:15
apurvanandan[m]
In creating makefile, can we use xilinx Synthesis tool
12:16
Bertl
if you manage to get that working properly without special assumptions on the pathes then yes, why not
12:52
aSobhy
joined the channel
13:48
sebix
left the channel
14:25
se6astian|away
changed nick to: se6astian
15:28
araml
joined the channel
16:17
BAndiT1983|away
changed nick to: BAndiT1983
16:20
se6astian
changed nick to: se6astian|away
16:58
Bertl
off for now ... bbl
16:58
Bertl
changed nick to: Bertl_oO
17:48
se6astian|away
changed nick to: se6astian
18:17
Dev
joined the channel
18:24
saurabh_raj
BAndiT1983: Hi, This is saurabh here, I am interested in the frameserving capabilities of OpenCine
18:27
saurabh_raj
I have done the basic tasks of C++ challenge and one additional task of saving the pic as BMP file which is required for OpenCine proposal
18:27
Dev
left the channel
18:27
saurabh_raj
I am currently trying the second task of saving the pic as single frame avi
18:29
saurabh_raj
BAndiT1983: I was hoping you could review my current code and provide some suggestions on what should I improve upon
18:31
BAndiT1983
hi saurabh_raj, have you committed your code to github already?
18:35
saurabh_raj
I have the code in my local directory, if needed I can create a repo on github and share the link
18:36
BAndiT1983
yes, please create a repo, i don't accept any archives, as they can contain malicious software
18:37
saurabh_raj
ok.. sure.. I will create the repo then get back to you.
18:38
BAndiT1983
alright
18:53
mrohit[m]
Hi BAndiT1983 , please have a look at my code too. I have updated it as suggested
18:53
mrohit[m]
Hi BAndiT1983 , please have a look at my code too. I have updated it as suggested
18:53
saurabh_raj
BAndit1983: mailed you link to the github repo
18:54
mrohit[m]
Sorry, there was a glitch in my irc client ,apparently sent the message twice
19:00
BAndiT1983
saurabh_raj, i will respond by mail when i have taken a look
19:01
BAndiT1983
mrohit[m], will take a look when i have some time, but this time i expect much better results, if you have followed my advices, especially in OOP area
19:02
mrohit[m]
yes, sure
19:03
BAndiT1983
saurabh_raj, biggest complaint from the start is formatting, please adjust it, as i miss indentations and please place braces correctly, it's C++ not java or javascript, so in my books opening braces on same line is a big no
19:06
BAndiT1983
sorry mrohit[m], but i don't see proper OOP there yet, you are allocating a lot of memory in every step instead of using the image object
19:06
BAndiT1983
casting forth and back is not a proper way for this challenge task
19:07
BAndiT1983
also formatting is a no-go, either you have too many blank lines or no blank lines, which is making reading through the code rather difficult
19:10
mrohit[m]
But I have allocated memory only in the Image object itself, and passed all the data through the object itself
19:11
mrohit[m]
and getter functions are used to send the pointers to this memory
19:11
BAndiT1983
why is it casting forth and back? don't see a point to do it, as we know the data format here
19:12
saurabh_raj
Ok... I will modify my code according to these suggestions..
19:13
Dev
joined the channel
19:13
saurabh_raj
what about modularity, should I reduce the number of files?
19:15
BAndiT1983
it's not a competition of lowest number of files, the application has to follow some sort of processing pipeline for images, which should include necessary steps without blowing up with too many classes or unnecessary stuff
19:15
BAndiT1983
also please check your code against C++ naming conventions, as you use more C style there, instead of C++, this means that variable should use camel-case starting with lower-case and methods starting with upper case
19:16
BAndiT1983
Dev, please use main channel, every discussion has to happen here, as GSoC should be an open event
19:18
Dev
Okay BAndiT1983, I was looking for libraw subtask, for providing support to OC , we can use C++ API of libraw,
19:18
Dev
What are the specific things we need to implement, can u please specify ?
19:18
Dev
so i can focus on that
19:19
BAndiT1983
libraw should be used to load raw image data
19:19
BAndiT1983
what else do you want to know there?
19:19
saurabh_raj
BAndiT1983: Sure, I will adjust my code according to this.. What about the green channel? Is there some issue there in the logic?
19:19
BAndiT1983
saurabh_raj, let me check, if i see something
19:21
BAndiT1983
i don't see any byte swapping
19:21
BAndiT1983
can it be that you get wrong results because of that? also the code looks mostly not very performant, as you use a lot of loops, will omit the fact that maintenance is even more difficult
19:22
Dev
Okay, i will look more into it
19:24
Dev
left the channel
19:24
mrohit[m]
Okay BAndiT1983 , I see your point there. This can be done without casting too. I did it because the Image object can be used to store any kind of format if required(by using void*) and we won't need to change the implementation of Image for that.
19:24
mrohit[m]
Have made the changes for casting
19:26
saurabh_raj
if byte swapping was an issue then it should be for red and blue channels too.. right? regarding the loops.. I will try to figure out something else and adjust my code according to expected formatting
19:27
BAndiT1983
mrohit[m], what about formatting, splitting long methods etc.? i will not review every small change over and over, please adjust your code and supply to review
19:28
BAndiT1983
saurabh_raj, check your code whre you split the bytes into sensels, you should get 2 sensels from 3 bytes, 12bit each -> 3 * 8 = 24 / 2 = 12bit
19:32
saurabh_raj
yeah! I am getting two sensels from 3 bytes. Almost the same thing is done for all three channels but the issue is with green channel only so I think the issue should be in the debayering process of the green channel, but I have checked that too.. I think that the only thing that explain black boxes are if some positions the array of green channel is skipped
19:33
mrohit[m]
okay BAndiT1983 I will get it updated soon
19:35
saurabh_raj
but I have checked for that.. logically it should'nt happen but also checked by printing some values of the array (10x10 tiles) from various positions but it was'nt the case
19:35
BAndiT1983
saurabh_raj, could be, but i can't get the flow in your source code, as there is a little bit too much of initialization of things going on, which is not really required as the data can be extracted much simpler
19:36
BAndiT1983
as the data is linear, but don'T forget about swapping bytes, then you can just iterate through it and extract the sensels
19:36
saurabh_raj
Yeah.. I get your point.. I can try reading all of the bytes in a single char array and maintain two variables for row and col location
19:37
BAndiT1983
this sounds better instead of reading 3 bytes every time
19:39
BAndiT1983
off for a moment, if there is something oyu want to know, just write it, will read the logs later
19:39
BAndiT1983
changed nick to: BAndiT1983|away
19:39
saurabh_raj
Sure.. I will make all these adjustments which you have suggested and check it with you tomorrow
19:43
comradekingu
left the channel
19:44
mrohit[m]
is swapping bytes necessary for this data? because after swapping, the images look incorrect ,where as without swapping they are fine
19:45
mrohit[m]
has anyone of you guys faced something similar?
20:00
BAndiT1983|away
changed nick to: BAndiT1983
20:03
Bertl_oO
off to bed now ... have a good one everyone!
20:03
Bertl_oO
changed nick to: Bertl_zZ
20:04
saurabh_raj
left the channel
20:28
comradekingu
joined the channel
20:39
BAndiT1983
changed nick to: BAndiT1983|away
20:48
se6astian
changed nick to: se6astian|away
22:39
BAndiT1983|away
changed nick to: BAndiT1983
23:42
BAndiT1983
changed nick to: BAndiT1983|away