Current Server Time: 16:55 (Central Europe)

#apertus IRC Channel Logs

2019/07/09

Timezone: UTC


00:01
leonardus
According to this page, it's still the most recent: https://github.com/wolfgangw/digital_cinema_tools/wiki/Frequently-asked-questions#smpte-where
00:03
Bertl_oO
https://ieeexplore.ieee.org/document/7864300
00:06
leonardus
ah, I've got it now
00:06
leonardus
by being an upstanding law-abiding citizen and paying full price, of course
00:09
Bertl_oO
the way it is supposed to be :)
00:13
RexOrCine
Bertl: Is Interface (Dummy) Board technically part of the stack or the SFE?
00:13
Bertl_oO
part of the stack, although it is not doing much
00:13
RexOrCine
Thanks.
00:16
Bertl_oO
np
00:31
illwieckz
left the channel
00:40
illwieckz
joined the channel
01:14
RexOrCine
Night all.
01:14
RexOrCine
changed nick to: RexOrCine|away
02:46
Bertl_oO
off to bed now ... have a good one everyone!
02:46
Bertl_oO
changed nick to: Bertl_zZ
05:15
BAndiT1983|away
changed nick to: BAndiT1983
06:21
illwieckz
left the channel
06:37
BAndiT1983
changed nick to: BAndiT1983|away
07:13
se6astian|away
changed nick to: se6astian
11:55
mithro
left the channel
11:56
uberardy
left the channel
12:01
mithro
joined the channel
12:01
uberardy
joined the channel
12:04
Umori
left the channel
12:06
Umori
joined the channel
12:08
se6astian
changed nick to: se6astian|away
12:17
Bertl_zZ
changed nick to: Bertl
12:17
Bertl
morning folks!
12:56
se6astian|away
changed nick to: se6astian
12:59
se6astian
left the channel
12:59
se6astian
joined the channel
13:06
apurvanandan[m]
Hi Bertl
13:07
apurvanandan[m]
Can we assume that once the data transfer has started, the delay in clock and data path won't change in between?
13:07
apurvanandan[m]
Do we have to realign at regular intervals?
13:09
Bertl
for a short period, you can assume that there won't be any changes
13:10
Bertl
for longer periods, changes in temperature or power might change the alignment, but usually not very much
13:10
Bertl
so the best approach here is to monitor the 'quality' of the link and if necessary, make some adjustments
13:13
apurvanandan[m]
Ok fine
14:42
se6astian
changed nick to: se6astian|away
14:50
se6astian|away
changed nick to: se6astian
14:50
se6astian
left the channel
14:50
se6astian
joined the channel
14:50
se6astian
changed nick to: se6astian|away
14:52
se6astian|away
changed nick to: se6astian
14:52
se6astian
left the channel
14:52
se6astian|away
joined the channel
15:04
se6astian|away
changed nick to: se6astian
15:04
se6astian
left the channel
15:04
se6astian|away
joined the channel
15:05
se6astian|away
changed nick to: se6astian
15:05
se6astian
left the channel
15:05
BAndiT1983|away
left the channel
15:05
philippej
left the channel
15:05
Nira|away
left the channel
15:05
RexOrCine|away
left the channel
15:45
dev__
joined the channel
15:50
supraraj
joined the channel
15:51
supraraj
Gfddj
15:51
supraraj
Good evening dev__
15:51
dev__
left the channel
15:52
supraraj
oh well :/ I thought only I had unreliable connection
15:59
dev__
joined the channel
15:59
supraraj
Greetings :)
15:59
dev__
Hello supragya_
15:59
dev__
I am here now
16:00
supraraj
Very well... we can start. Andrej will join later... he sent an email saying he's busy
16:00
dev__
okay,
16:00
supraraj
So... Could you brief me about the key things done this particular week
16:01
supraraj
I have a few questions on how you have conceptualized the solution and a few coding practices I see on github
16:01
dev__
Yes, First the was to correct the out of aviencode and then i added static allocators in it.
16:02
dev__
Had some problems in flipping the frame (inverted frame)
16:02
supraraj
okay...
16:03
supraraj
what problems were you getting?
16:03
dev__
https://docs.microsoft.com/en-us/windows/win32/api/wingdi/ns-wingdi-tagbitmapinfoheader
16:03
supraraj
?
16:03
dev__
The document says, we can flip the frame by using negative biheight
16:03
dev__
it worked
16:04
dev__
but fails in VLC ..
16:04
supraraj
How did it work and failed...?
16:04
dev__
it worked in video player of ubunto
16:05
supraraj
Hmm maybe it's not widely used
16:05
supraraj
Or there maybe some other metas who may code this... which is used by others and more popular...
16:05
supraraj
either way... we should not blame VLC, but correct our code :)
16:06
dev__
Yes, makes sense
16:06
dev__
I will look for it surely
16:07
supraraj
I don't understand though... if we are always reading from position 0,0 which is top left
16:07
dev__
I was aiming for frameservering
16:07
supraraj
and that's our reference for everything .... when we allocate and save frames and channels in mem
16:07
supraraj
why is the thing flipped if we write it just that way in the avi :/ weird...
16:08
supraraj
anyways, please continue
16:09
dev__
okay , so u mentioned somethings avi file map for Fuse
16:09
dev__
yesterday
16:09
supraraj
let's not jump there so quick
16:10
supraraj
A few questions regarding static allocator
16:10
dev__
yes
16:10
supraraj
1. Is it fully built? what happens if we query something which is not there in static alloc? how are you managing the static alloc... LRU etc?
16:10
supraraj
Have you touched all the edge cases etc...
16:11
dev__
We check if the frame is present in buffer (_mem) first, if its not present at this point of time.
16:12
supraraj
2. Why do I see some code commits done ON github.com ... not pushed TO github.com ... Nothing particular here... but some commits worry me... (not a major point... just curious about the workflow)
16:12
dev__
we will always have its position in the main file buffer. we will allocated this frames where we find space
16:12
supraraj
please continue
16:13
dev__
if the _mem is full, we will start allocating memory from starting of buffer
16:15
dev__
a circular queue feature, we can say
16:16
supraraj
sure...
16:16
supraraj
regarding 2?
16:17
dev__
some commits worry me : can u please point me there
16:17
supraraj
your IDE is not correctly set up, commits are unclean... just so you know
16:18
supraraj
anyways... how do you visualise frameserver would function since you have already built so much
16:18
supraraj
Please explain what would we need now in codebase where and why
16:20
dev__
I would say , we need fuse implementation so we can request the data and supply back to video player.
16:20
dev__
The read callback implemention of Fuse
16:21
dev__
We are doing this currently using a loop (which loops from 1 to frameCount)
16:21
supraraj
A few concerns are certainly there: for example, how do you handle the case where downscaler is on (half size) vs off
16:22
supraraj
a dirty patch is in place where you /2 currently, but when working with full frameserver, that parameter is very vital for frame number computation
16:25
dev__
AVIEncode should know the parameters of downscaler
16:25
dev__
which parameters we were taking about that time, I am confused here a little
16:26
supraraj
https://github.com/kakashi-Of-Saringan/opencine/commit/eab7fa1448654fbd8d39b7612122609794589171
16:28
dev__
How I should replace /2 ?
16:29
supraraj
AVIEncode should know the parameters of downscaler <- this... you answered it yourself
16:29
BAndiT1983
joined the channel
16:29
supraraj
Hello BAndiT1983
16:29
dev__
hello BAndiT1983
16:30
BAndiT1983
Hi supraraj and dev__
16:30
BAndiT1983
hi Bertl, BIP crashed hard again, could not join earlier
16:30
dev__
I need to replace / 2, u Mean
16:30
BAndiT1983
how to solve /2 dilemma?
16:30
BAndiT1983
i gave you some hints a couple of weeks ago
16:30
BAndiT1983
remember the video clip class?
16:31
dev__
Yes , The order of memory ..
16:31
dev__
allocation
16:32
BAndiT1983
??
16:32
dev__
https://trello.com/c/ANcx80qY/9-order-of-memory-allocation
16:32
BAndiT1983
and what is your solution?
16:33
dev__
We should know beforehand , we are going to use downscaler , then we will only allocate /2 memory
16:34
dev__
that*
16:34
BAndiT1983
and if downscaler makes 1/3, 1/4 etc.?
16:34
supraraj
How are links you give solution to anything someone asks? Please refrain from doing that without context.
16:35
dev__
I was typing at that time , sorry
16:35
supraraj
Another important thing: Do remember... when you are writing an AVI, which eventually you will frameserve, you will have to also tell the amount of frames in total that AVI file would have... this would mean that we need to know the number of files in the folder (seq)... dev__ do we already have a code for it or you have a solution?
16:36
dev__
if downscaler makes 1/3, 1/4 etc : - we should know about this also
16:36
BAndiT1983
But this would mean tight coupling
16:36
BAndiT1983
also what happens if we use some other method for debayering?
16:37
dev__
supraraj: we are suppling frameCount at this moment, that is remaining
16:39
dev__
So we need something which can know all these parameter like which processing method to be used, how much size we need to allocate
16:39
dev__
automatically .
16:39
BAndiT1983
hm, and where could it be?
16:39
dev__
That's what should a video clip class do , yes ?
16:40
BAndiT1983
yes, video clip class should hold general infos, which are valid for every frame
16:40
BAndiT1983
but how would it get the right height of the data, e.g. if it's smaller because of downscaler
16:42
sraj
joined the channel
16:43
dev__
i was thinking , the user will give these parameters
16:43
dev__
the UI structure, we planned for it
16:43
dev__
left the channel
16:43
dev__
joined the channel
16:44
supraraj
left the channel
16:44
BAndiT1983
??
16:44
sraj
Doesn't really sound right... Means that user will always have to tell what's to happen?
16:44
BAndiT1983
why should user give this parameters, if they should be got from the file and pipeline
16:45
dev__
MlV sample gives height and width, we want
16:45
dev__
are we considering other codecs here
16:46
BAndiT1983
i don't get the logic at the moment, please explain
16:47
sraj
Tomorrow if we have a smart way to detect whether we need downscaling on a certain machine through some internal system, oc may let's say automatically set it's variables... Why would u need user to provide it to the system through manual input
16:47
dev__
yes , makes sense.
16:48
dev__
But , I have to think how we can do this.
16:49
BAndiT1983
ehm, dev__, it's your task, remember?
16:49
dev__
yes BAndiT1983,
16:51
dev__
I am clear about the need of such system which can automatically know the height , width of data
16:51
BAndiT1983
???
16:51
BAndiT1983
we know width and height, it comes from the loader
16:51
BAndiT1983
you have to find a way how to tell the video data and allocator how much space is required for frames etc.
16:53
dev__
The space need to be allocated, will depend on parameters like how we are downscaling it etc
16:54
BAndiT1983
yes, and where could we get them?
16:56
dev__
it depends on what we what to do 1/3, 1/4 etc, as u said
16:57
sraj
:/
16:57
BAndiT1983
this is not really the answer i'Ve expected
16:59
dev__
I don't have solution yet ,sorry
17:00
dev__
Can i take time
17:00
Bertl
aSobhy, apurvanandan[m]: you around?
17:00
BAndiT1983
how long do you need?
17:00
apurvanandan[m]
Yes
17:01
sraj
Hello Bertl apurva
17:01
apurvanandan[m]
Hi Supragya
17:02
dev__
till tommorow , I will try to think about it
17:02
Bertl
apurvanandan[m]: did you manage to work with aSobhy on the BER testing?
17:02
BAndiT1983
ok, let us do a meeting day after tomorrow, as i also need time because of a very important project at work
17:03
BAndiT1983
will inspect your code, so i know what's going on there
17:03
dev__
I will mail u asap about the process
17:03
apurvanandan[m]
Actually I was working alone, afaik he was busy with his graduation project.
17:04
Bertl
thought so, but good to know :)
17:04
Bertl
so, what's the status and what are the current challenges?
17:05
dev__
sraj, BAndiT1983 , i will be working on it now.
17:06
BAndiT1983
very good
17:07
dev__
okay , update u , Good night!
17:07
apurvanandan[m]
Bertl: But it is at its final stage, all gearwork is working. I am making that 8:1 to 10:1 converter after that I will test on hardware. On simulation it is working good.
17:07
Bertl
okay, so has been any testing done on real hardware yet?
17:08
dev__
left the channel
17:09
apurvanandan[m]
Configuring modelsim what damn irritating
17:09
apurvanandan[m]
was*
17:09
apurvanandan[m]
But now everything is set up and working peacefully
17:10
apurvanandan[m]
Yes, but only was receiving random data. First need to verify complete design on simulation.
17:10
apurvanandan[m]
Thats why I installed MIcrosemi IDE for Modelsim
17:12
apurvanandan[m]
At this moment I am trying to make the 8:1 to 10:1 work, rest all is working fine
17:12
apurvanandan[m]
The ALIGNWD pin makes the link training very easy.
17:19
Bertl
okay, please create some VCD files for the simulations
17:19
sraj
left the channel
17:19
apurvanandan[m]
Link training is working for x8 gearing
17:19
BAndiT1983
left the channel
17:20
apurvanandan[m]
Do I to send right now?
17:21
Bertl
no, just create some during testing so that I get an idea what you're doing and how the progress is
17:22
Bertl
originally the bandwidth tests should have been finished for the first evaluation
17:22
apurvanandan[m]
I will send you tonight!
17:29
RexOrCine|away
joined the channel
17:29
Nira|away
joined the channel
17:29
BAndiT1983|away
joined the channel
17:29
se6astian|away
joined the channel
17:29
se6astian|away
changed nick to: se6astian
17:29
philippej|away
joined the channel
17:29
BAndiT1983|away
changed nick to: BAndiT1983
17:29
philippej|away
changed nick to: philippej
18:58
illwieckz
joined the channel
19:14
se6astian
changed nick to: se6astian|away
19:15
se6astian|away
changed nick to: se6astian
20:17
BAndiT1983
changed nick to: BAndiT1983|away
20:17
BAndiT1983|away
changed nick to: BAndiT1983
20:20
Bertl
off for now ... bbl
20:20
Bertl
changed nick to: Bertl_oO
20:39
BAndiT1983
changed nick to: BAndiT1983|away
21:07
se6astian
off to bed
21:07
se6astian
good night
21:07
se6astian
changed nick to: se6astian|away
21:45
apurvanandan[m]
Hey Bertl, I am using K28.5 for link training. Can I directly compare equality with the encoded RD+ and RD- words for link training ie without decoding them as decoder takes two clock cycles.
21:50
illwieckz
left the channel
22:05
illwieckz
joined the channel
22:52
illwieckz_
joined the channel
22:55
illwieckz
left the channel