Current Server Time: 08:31 (Central Europe)

#apertus IRC Channel Logs

2019/04/27

Timezone: UTC


00:10
Spirit532
left the channel
00:10
Spirit532
joined the channel
02:57
futarisIRCcloud
left the channel
03:27
Bertl
off to bed now ... have a good one everyone!
03:28
Bertl
changed nick to: Bertl_zZ
04:43
futarisIRCcloud
joined the channel
06:33
BAndiT1983|away
changed nick to: BAndiT1983
06:53
futarisIRCcloud
left the channel
07:09
BAndiT1983
changed nick to: BAndiT1983|away
07:12
BAndiT1983|away
changed nick to: BAndiT1983
09:09
futarisIRCcloud
joined the channel
09:33
se6astian|away
changed nick to: se6astian
10:41
BAndiT1983
changed nick to: BAndiT1983|away
11:19
se6astian
changed nick to: se6astian|away
11:21
se6astian|away
changed nick to: se6astian
11:35
Bertl_zZ
changed nick to: Bertl
11:35
Bertl
morning folks!
11:44
apurvanandan[m]
Good morning!
11:59
Bertl
hey, how's it going?
12:17
futarisIRCcloud
left the channel
12:23
apurvanandan[m]
Not much, Just going through the codes and reading about new things.
12:24
Bertl
like for example?
12:24
apurvanandan[m]
Bertl I think major part of island data won't be required. Just info about the resolution needs to be sent which can be sent once in the starting. Am I right?
12:25
Bertl
are you talking about HDMI now?
12:27
apurvanandan[m]
I mean that island data is sort irrelevant to the usb module.
12:27
Bertl
ah, yes
12:31
Bertl
the information is still relevant, but no need to transfer the island data over USB
12:31
Bertl
i.e. there are different resolutions and color modes on USB as well
12:32
Bertl
but there is no point in using the convoluted island data for it
12:32
Bertl
in the UVC case, all this needs to be set for the FTDI
12:32
Bertl
in the raw data case, there needs to be some kind of header information
12:33
BAndiT1983|away
changed nick to: BAndiT1983
13:11
apurvanandan[m]
Bertl why is that the hdmi_ch0 to 3 go to color correction module before being encoded while hdmi_ch4 is directly sent to rgb encoder module?
13:19
Bertl
channel 0-3 contain the image data (R,G1,G2,B) while ch4 contains metadata
13:21
Bertl
(i.e. the overlay)
13:23
apurvanandan[m]
Oh I see, there is a overlay process for this.
13:28
Bertl
aSobhy: what's up on your side?
13:47
Y_G
joined the channel
14:11
aSobhy
I'm taking a tour in lattice diamond
14:14
aSobhy
didn't download vivado yet but I'm trying
14:17
aSobhy
beside Study networks (just refreshment) :)
14:17
Bertl
ah, Lattice Diamond :)
14:17
Bertl
I remember that it was a little tricky first to find all the parts to run it from commandline
14:18
Bertl
I do not remember much else about it though
14:21
aSobhy
yeah collecting my commands to save them
14:22
Bertl
there is a make file for the routing fabric bitstream
14:22
Bertl
http://vserver.13thfloor.at/Stuff/AXIOM/BETA/LATTICE/
14:22
Bertl
(might come handy)
14:26
aSobhy
nice, seeing them now ^_^
14:30
apurvanandan[m]
Bertl, could you explain me what is this overlay here?
14:33
Bertl
sure, currently the sensor image data is burst into memory in a square four sensel arrangement
14:33
Bertl
i.e. a square of four sensel (R, G1, G2, B) with 12 bit each are stored together in a 64bit memory location
14:34
Bertl
4x12 bit = 48bit which leaves 16bit 'unused'
14:35
Bertl
we decided to use this space for an overlay which affects the full HD pixels generated from the 2x2 sensel words
14:36
Bertl
if you look at overlay.vhd, there the 16bits are interpreted to form said overlay image
14:37
Bertl
two bits are used to control transparency, the other control color and there are special symbols for a blink effect
14:38
Bertl
(this happens outside of the overlay entity)
14:42
BAndiT1983
changed nick to: BAndiT1983|away
14:49
Bertl
the alpha mixing is done in overlay_proc in top.vhd
14:50
Bertl
well, looking at the code, everything is done in overlay_proc in the current version :)
15:12
Bertl
any further questions or all clear? :)
15:13
apurvanandan[m]
This part is clear , but more questions coming soon. ^^'
15:15
apurvanandan[m]
Earlier I was planning to use the lut after the color correction for doing the 8b/10b encoding but now I think hd_code is the signal to be sent over.
15:17
Bertl
for raw output, we definitely want unmodified raw data from the sensor
15:18
Bertl
i.e. the 12bit sensor data (R,G1,G2,B) without any LUT or color matrix
15:18
Bertl
for the UVC case, we have several options
15:19
Bertl
we can do color space conversion on the MachXO2 or we can do that in the ZYNQ fabric
15:21
apurvanandan[m]
So that means I will use the data coming from the fifo which is connected to axi hp reader.
15:22
Bertl
for raw output, yes
15:31
apurvanandan[m]
Great. Now it is more clear i.e. basically this fifo data needs to reach MachXO2. :)
15:42
akriti
joined the channel
15:43
Spirit532
left the channel
15:43
akriti
left the channel
15:56
Spirit532
joined the channel
16:15
BAndiT1983|away
changed nick to: BAndiT1983
17:17
Y_G
left the channel
17:41
Dev_
joined the channel
17:45
Dev_
left the channel
17:57
BAndiT1983
changed nick to: BAndiT1983|away
18:51
Bertl
off for now ... bbl
18:51
Bertl
changed nick to: Bertl_oO
19:27
Bertl_oO
changed nick to: Bertl
19:27
Bertl
back again ...
19:32
Nira|away
changed nick to: Nira
19:46
Bertl
evening Nira!
19:47
Nira
Hello!
19:47
Bertl
How are you?
19:54
Nira
Great! Weekends always look so fast
19:55
se6astian
welcome back, how are you getting along with the code so far?
19:57
Nira
I'm reading PIC32 documentation now, I've been checking the code and I'm trying to understand what I don't know yet!
20:35
se6astian
changed nick to: se6astian|away
20:51
Nira
changed nick to: Nira|away
20:51
Nira|away
changed nick to: Nira
22:05
aSobhy
can I have a look at any VHDL code on both sides machxo2 and zinq ?
22:07
Bertl
there is some VHDL code for the Lattice which I linked not too long ago (check IRC logs for today)
22:08
aSobhy
also the past year their was a student working on the project as I submitted where do I get it
22:08
Bertl
https://github.com/apertus-open-source-cinema/axiom-beta-firmware/tree/master/peripherals
22:08
Bertl
here you can find the Zynq (soc_main) and MachXO2 (routing_east) examples
22:10
aSobhy
is the same code for both MachXO2 FPGAs
22:21
BAndiT1983|away
changed nick to: BAndiT1983
22:33
Bertl
at the moment we only have code in one routing fabric
22:33
Bertl
i.e. the other one is empty and jtag instructions are used to toggle pins :)
22:37
Nira
changed nick to: Nira|away
22:40
se6astian|away
changed nick to: se6astian
22:40
se6astian
changed nick to: se6astian|away
22:44
BAndiT1983
changed nick to: BAndiT1983|away
22:55
Bertl
off for now ... bbl
22:55
Bertl
changed nick to: Bertl_oO
22:57
futarisIRCcloud
joined the channel
23:06
aSobhy
mmm okaaay