| *** tpb has joined #timvideos | 00:00 | |
| tija | shenki: I was trying to install your driver for usb uart. sudo make install gives an error. | 00:08 |
|---|---|---|
| tija | make: *** No rule to make target `install'. Stop. | 00:08 |
| *** xfxf has quit IRC | 00:45 | |
| *** xfxf has joined #timvideos | 00:57 | |
| mithro | ayush3504: I'll be around in 30->45 minutes | 01:05 |
| mithro | shenki: are you on the fpgalink list? | 01:55 |
| mithro | ayush3504: opps, lunch time. Will be back in an hour. | 02:10 |
| shenki | tija: okay. we probably shouldn't make install it anyway, as it conflicts with the cdc-driver | 02:23 |
| shenki | tija: you need to follow the steps in the README.exar | 02:24 |
| shenki | tija: rmmod cdc_acm && rmmod vizzini && modprobe -r usbserial && modprobe usbserial && insmod ./vizzini.ko | 02:24 |
| shenki | im not sure how necessary all of those steps are | 02:25 |
| shenki | we really should get support for the device upstream | 02:25 |
| shenki | but that's not a concern for you at this stage | 02:25 |
| ayush3504 | mithro: no prob :) | 02:32 |
| tija | shenki: okay, will do that | 02:37 |
| shenki | tija: hows the progress? | 02:38 |
| tija | shenki: I was planning to use fpgalink library to bring debug data from fpga to pc | 02:38 |
| tija | turns out that since fx2 is already being used, it will be difficult. And I don't want to mux pins as I don't want to disturb the stream | 02:39 |
| tija | otherwise it would have been really elegant way to send data from fpga. | 02:40 |
| tija | Now I am planning to use the UART | 02:40 |
| shenki | mithro: no, im not on the fpgalink list | 02:41 |
| shenki | tija: we already have a serial pipe from the fx2 to the PC | 02:41 |
| shenki | tija: when you load our firmwrae on the fx2, you will notice it exposes a USB CDC-ACM device | 02:42 |
| tija | shenki: but isn't it using pins for video streaming | 02:42 |
| tija | shenki: I mean can I send serial data without affecting the video stream | 02:43 |
| shenki | tija: yes | 02:43 |
| shenki | you can | 02:43 |
| tija | shenki: oh okay! then i guess I need to figure that out. I don't have much knowledge about fx2 part of hdmi2usb firmware. | 02:45 |
| mithro | ayush3504: back now | 02:46 |
| mithro | ayush3504: just going through your responses to the github issues | 02:47 |
| mithro | ayush3504: then I wanted to chat about auto-detection | 02:47 |
| shenki | tija: yeah. whichever one works for you | 02:47 |
| shenki | tija: the advantage of that one is we already have some infrastructure set up | 02:48 |
| shenki | tija: but it's upto you; if you want to use the xr21v141x one that you've got going, then that's good too | 02:48 |
| shenki | whichever you find easier | 02:48 |
| tija | shenki: fine | 02:53 |
| mithro | ayush3504: can you give me edit permission to your repo so I can reopen the issues I need to? | 02:59 |
| shenki | tija: how are you going with the desgin document? | 03:18 |
| tija | shenki: I am planning to use switched to select which debug data should be sent serially. So first I am deciding what are the relevant debug data. | 03:29 |
| tija | shenki: But before that I need to decide which serial interface I am going to use. I am reading the code of cdc in fpga firmware. Actually that is being used to send commands to encoder and image selector. | 03:31 |
| mithro | tija: that is *part* of the design document | 03:31 |
| mithro | tija: IE it should include the things you looked at and why you chose to select the bits you did | 03:31 |
| mithro | ayush3504: poke? | 03:31 |
| shenki | tija: as mithro said, we need to get into a good habbit of specifying what we're going to do before we start hacking | 03:34 |
| shenki | tija: so before reading the code, write down what the problem is, how you plan to solve it given what you currently know | 03:35 |
| shenki | tija: then we can help you by filling in the blanks | 03:35 |
| shenki | it's okay if you then start reading the code and realise you need to change the design | 03:35 |
| shenki | it's not a formal report or anything, just some notes in a google doc | 03:35 |
| shenki | tija: take this one for example: https://docs.google.com/document/d/1-E4SGuPz4Wjqs2BP9ck8GL6wrxWUUAnE6lsq7E3uBgE/edit | 03:36 |
| tpb | Title: I2C VHDL Cores & their comparison - Google Docs (at docs.google.com) | 03:36 |
| tija | shenki: okay I got it! My laptop battery is going to die (No power here). I will make a design doc asap. | 03:40 |
| shenki | tija: okay | 03:41 |
| shenki | tija: you made a commitment to get this debug stuff done by end of wednesday. do we need to revisit that deadline? | 03:41 |
| tija | shenki: No, I will put some extra hours and try to complete by then. Only thing is the exar driver should work which I will test once power is back. | 03:44 |
| shenki | tija: okay | 03:44 |
| shenki | tija: can you tell me more about the power situation? | 03:44 |
| tija | shenki: It goes off in the morning so I sleep then. Comes back in the afternoon so I wake up have my lunch and then start working. | 03:46 |
| shenki | okay | 03:47 |
| shenki | tija: it interests me, as i've not lived somewhere where i've had to deal with that | 03:48 |
| shenki | tija: how do you get net connectivity? | 03:48 |
| ayush3504 | mithro: in a while, no power... | 03:50 |
| tija | shenki: USB powered broadband modem. I am planning to go back to my parent's place and work from there. | 03:51 |
| *** tija has quit IRC | 03:54 | |
| ayush3504 | didn't notice that a discussion on power cuts is already going on :p | 04:06 |
| ayush3504 | mithro: back | 04:06 |
| mithro | ayush3504: great | 04:06 |
| mithro | still going through my email | 04:06 |
| ayush3504 | mithro: ok | 04:07 |
| mithro | ayush3504: so about auto-detection | 04:29 |
| ayush3504 | mithro: yes, did u understand what I did for auto-detection? | 04:29 |
| mithro | ayush3504: it looks like you are currently planning for every daughterboard to have EEPROM or microcontroller on it to identify them? | 04:29 |
| ayush3504 | mithro: yes, every daughterboard can have a cheap pin addressable eeprom | 04:30 |
| ayush3504 | mithro: costs some $0.2 | 04:30 |
| mithro | ayush3504: when saying prices you should always do it in $X.XX, ie $0.20 | 04:31 |
| mithro | ayush3504: I misread that as $0.02 at first | 04:31 |
| ayush3504 | mithro: $0.24 for 100 | 04:32 |
| ayush3504 | mithro: $0.31 for 1 | 04:32 |
| mithro | ayush3504: I was thinking that we could instead have a very small micro-controller on the motherboard which reads the RC identity values from daughterboards and presents the info via the I2C interface (effectively emulating the EEPROM) | 04:33 |
| ayush3504 | mithro: yes that's what i was initially doing | 04:33 |
| ayush3504 | mithro: the cost of micro is comparable to cost of 6 eeproms | 04:34 |
| ayush3504 | mithro: and it's much more complex | 04:34 |
| ayush3504 | mithro: the eeprom method is simple and can also store more data about the daughterboard | 04:35 |
| ayush3504 | mithro: that ID controller sheet you mentioned was using the microcontroller for RC detection, and emulating itself as an eeprom on I2C. | 04:35 |
| shenki | simple is good | 04:35 |
| ayush3504 | thanks shenki :) | 04:36 |
| mithro | ayush3504: what type of micro where you looking at btw? | 04:40 |
| ayush3504 | mithro: PIC16F1704 the 2nd cheapest micro i found on parametric search with i2c and sufficient pins. | 04:41 |
| ayush3504 | mithro: here's how the eeprom method would work - the atlys would poll on i2c on slave address 0 eeprom which contain ID and possibly some more data about the expansion board connected. If the ID is found to be of serial expansion board, it would read from slave address 1,2,3...7 to get ID strings from daughterboards. | 04:42 |
| shenki | you don't need an i2c periperhal; you can bitbang i2c | 04:46 |
| ayush3504 | shenki: true, I wanted to use the same micro for other things. Like this micro has a UART which makes it useful for me to use in daughterboards. | 04:53 |
| shenki | ayush3504: sounds good | 04:54 |
| ayush3504 | mithro / shenki: RC detection seems to have another issue that it relies on logic level thresholds of the I/O pins. A slight error in that will change the measurements greatly. How much will it change is not known because the errors in logic level thresholds are not given in datasheet. It can slightly vary depending on temperature, VCC, noise and production | 05:01 |
| ayush3504 | batch. A comparator with precision reference could be used, but then it makes things even more complex. | 05:01 |
| mithro | sorry, got called into quick meeting | 05:05 |
| mithro | back now | 05:06 |
| mithro | ayush3504: this is another case where I would recommend having a spreadsheet which shows the information comparing the cost of the two solutions | 05:06 |
| shenki | i don't like the idea of RC detectio | 05:09 |
| shenki | i don't like the idea of RC detection | 05:09 |
| shenki | for the reasons you mention ayush3504 | 05:09 |
| shenki | i mean, we should easily be able to do a few different levels | 05:10 |
| shenki | but are we really that short on pins? couldn't we use resistors to encode a number on a few pins? | 05:10 |
| mithro | shenki: you should point ajit to Rohit's blog and this post in particular -> http://dreamsxtrinsic.blogspot.com.au/2014/06/gsoc-daily-log-trails-errors-part-2.html | 05:12 |
| tpb | Title: [GSoC Daily Log]: Trails & errors - Part 2 | Dreams eXtrinsic (at dreamsxtrinsic.blogspot.com.au) | 05:12 |
| shenki | mithro: ok | 05:13 |
| mithro | RC detection is only a good solution if it's much cheaper than an EEPROM solution - It should be 5c compared to 35c to make it worth doing. | 05:14 |
| mithro | does I2C allow "chaining / passthrough" like SPI? Its normally configured as multi-tap bus style.... | 05:23 |
| mithro | ayush3504: ping? | 05:23 |
| ayush3504 | mithro: yes | 05:24 |
| ayush3504 | mithro: simply connecting them in parallel works | 05:24 |
| mithro | ayush3504: but that requires them to all have unique addresses, right? | 05:24 |
| ayush3504 | mithro: yes, those are given by address pins A0, A1, A2 | 05:24 |
| mithro | ayush3504: if we where using SPI you'd just attach them all serially and then just keep reading information until you got an empty byte, thus auto-detecting the number of devices attached | 05:25 |
| ayush3504 | mithro: yeah, but daisy chaining has a problem, for example if you remove one of the SPI device in the chain, it breaks it. | 05:28 |
| mithro | ayush3504: sure, it means you need to populate the slots in an increasing order or have a jumper which shorts the SDI/SDO pins around the missing chip | 05:29 |
| ayush3504 | mithro: yeah, i2c is better in that regard | 05:29 |
| mithro | ayush3504: I still think we don't have the right solution for the serial board / auto-detect yet | 05:34 |
| ayush3504 | mithro: RC is not suitable for the technical reasons i gave | 05:34 |
| mithro | ayush3504: yeah, I'm leaning towards an I2C based solution - the question is how to make that work | 05:35 |
| ayush3504 | mithro: did you not clearly understand the i2c eeprom method? | 05:36 |
| mithro | ayush3504: no, I understand it - each daughterboard has an i2c eeprom who's address is set via the slot it is plugged into | 05:37 |
| ayush3504 | mithro: exactly | 05:37 |
| mithro | ayush3504: the motherboard has a i2c eeprom too | 05:38 |
| ayush3504 | mithro: yes | 05:38 |
| mithro | ayush3504: having the FPGA scan the serial motherboard and then the daughterboards seems not the best | 05:39 |
| ayush3504 | mithro: the fpga logic will read expansion board ID on slave address 0. It wouldn't read higher slave addresses unless it finds a serial expansion board. | 05:39 |
| mithro | ayush3504: I'm wondering if the motherboard could end up being totally passive and we could just scan for the daughter boards directly | 05:39 |
| ayush3504 | mithro: or any other board that's modular in nature | 05:40 |
| mithro | ayush3504: yeah | 05:40 |
| mithro | ayush3504: for example, if we ended up with two expansion boards which didn't use all the pins then we could have a VHDCI splitter or something | 05:41 |
| ayush3504 | mithro: hmm, yes. technically it's not necessary to have eeprom on motherboard if the FPGA is made to handle exceptions when it doesn't find eeprom on a particular slave address. | 05:42 |
| mithro | ayush3504: I'm trying to figure out how we could change the auto-detection method to allow multiple daughterboards | 05:43 |
| mithro | ayush3504: s/daughterboards/VHDCI expansion boards/ | 05:43 |
| mithro | ayush3504: then your motherboard could just end up as VHDCI -> PMOD converter | 05:44 |
| ayush3504 | mithro: i think it's straight forward. It would depend on what ID it reads on slave 0. Let's reserve slave 0 to contain data only about the expansion board. If it doesn't read anything on slave 0, it would scan higher addresses to find possible daughterboards. This can also work if a another expansion board and two daughterboards are connected using the | 05:46 |
| ayush3504 | splitter idea. | 05:46 |
| mithro | ayush3504: that requires the splitter to be more than just wires (it needs to shift one of the daughterboards to another address) | 05:47 |
| ayush3504 | mithro: well first let's not make it a condition to read higher addresses depending on ID on slave 0, second, we won't populate slave 0 eeprom on motherboard. So it doesn't require changing any addresses | 05:48 |
| ayush3504 | mithro: for this top work the daughterboard addresses should always start from 1 | 05:49 |
| mithro | ayush3504: I was thinking that each expansion board used a different address, and we just scan the i2c bus for devices | 05:50 |
| ayush3504 | mithro: yes, the address could however start from 1. 0 can be reserved from expansion board ID | 05:52 |
| mithro | sadly it doesn't look like i2c ever thought about multiple devices at the same address | 05:52 |
| ayush3504 | mithro: also think about the case, if you were to include daugherboard circuits on the PCB, i.e. a customized pre-loaded. in such case you would want to use only one eeprom at address 0. | 05:53 |
| ayush3504 | mithro: technically we can have multiple slave devices at same address only if they listened to the bus, but that's not very useful in our case. | 05:55 |
| mithro | ayush3504: yeah | 05:58 |
| mithro | ayush3504: what EEPROM are you proposing to use? | 05:58 |
| mithro | ayush3504: still waiting on your weekly progress report too | 05:58 |
| ayush3504 | mithro: 24AA014 | 05:58 |
| ayush3504 | mithro: http://in.mouser.com/ProductDetail/Microchip-Technology/24AA014-I-SN/?qs=gySgac5WNsHB6VrGKNil7A== | 05:58 |
| tpb | Title: 24AA014-I/SN Microchip Technology | Mouser (at in.mouser.com) | 05:58 |
| mithro | Oh, is there a standard for EEPROM I2C interface? | 05:59 |
| ayush3504 | mithro: I think so, because there's a standard library to talk to 24xx eeproms in arduino. | 06:00 |
| mithro | 24xx EEPROMs are only one type of EEPROM - 24xx is probably a family of EEPROMs right? | 06:01 |
| mithro | 24xx might be "standard" enough though.... | 06:04 |
| ayush3504 | mithro: yes a family | 06:04 |
| mithro | ayush3504: it looks like that EEPROM is 1kbit? IE 128 bytes? | 06:05 |
| ayush3504 | mithro: yes. I found this to be the cheapest one in parametric search with cascading support (pin adressing) | 06:05 |
| mithro | We decided on I2C because rohit doesn't have any spare pins right? | 06:07 |
| *** CarlFK has joined #timvideos | 06:11 | |
| *** ChanServ sets mode: +v CarlFK | 06:11 | |
| mithro | wb CarlFK, I'm guessing your probably asleep? | 06:20 |
| mithro | ayush3504: where did you do the parametric search? | 06:20 |
| ayush3504 | microchip | 06:21 |
| CarlFK | mithro: I will be soon - spent the day trying to setup a site for presenters to edit their talk video metadata, and nothing went right, and rushed out the door before I could get to calling up Fedx | 06:22 |
| ayush3504 | mithro: http://www.microchip.com/ParamChartSearch/chart.aspx?branchID=70002&mid=&lang=en | 06:23 |
| tpb | Title: 24xx01 Products - Microchip Technology Inc (at www.microchip.com) | 06:23 |
| mithro | ayush3504: Using Microchip's parametric search is only going to give you microchip's solutions | 06:23 |
| ayush3504 | mithro: $0.19 here | 06:23 |
| mithro | ayush3504: can you try at something like Mouser or Digikey? | 06:24 |
| ayush3504 | mithro: tried that first. found the parametric search crappy. i'll try again | 06:25 |
| mithro | ayush3504: yeah - maybe just use it to find out alternative's to microchip and search on their sites directly | 06:26 |
| mithro | ayush3504: how are you going updating your requirements document? | 06:27 |
| ayush3504 | mithro: Yes there's a lot of updating there to do. I only have to add things that we have discussed and agreed upon.But I should do it a bit later, because there are other the pending tasks - like getting the PCB done | 06:29 |
| *** slomo has joined #timvideos | 06:30 | |
| ayush3504 | mithro: in mouser parametric search, the cheapest 8 pin eeproms are from microchip. Anything less than 8 pins (like SOT-5) will not have address pins. | 06:36 |
| mithro | shenki / Joelw / anyone else with an opinion. If the HDMI2USB device had an embedded processor inside it - what architecture should it be? http://opencores.org/usercontent,doc,1393289758 seems to be a good summary of the options. | 06:54 |
| *** rohitksingh has joined #timvideos | 06:54 | |
| shenki | mithro: inside the fpga? | 06:55 |
| mithro | shenki: yeah | 06:55 |
| mithro | heyo rohitksingh | 06:55 |
| rohitksingh | mithro: hi :) | 06:56 |
| shenki | probably some variant of the leon | 06:56 |
| shenki | depends what we wanted it to do | 06:57 |
| shenki | microblaze would work fine too | 06:57 |
| techman83 | What is the Leon doing? | 06:57 |
| techman83 | me = Leon :P | 06:57 |
| shenki | mithro: ^ | 06:57 |
| shenki | hah | 06:57 |
| mithro | shenki: PicoBlaze might be a solution too | 06:57 |
| shenki | i got the impression that needed to be coded in asm | 06:58 |
| shenki | (the code that runs on it needs to be) | 06:58 |
| mithro | shenki: for doing things like managing the "control" of the system (such as reading from the control serial port, responding with information, probing and configuring the expansion board and other peripherals, etc) | 06:58 |
| shenki | ok | 06:59 |
| shenki | i'd go with the microblaze | 06:59 |
| mithro | shenki: PicoBlaze is not well suited to C style code | 06:59 |
| rohitksingh | mithro: referring the logs, yes, the I2C was mainly decided due to lack of any extra pins | 06:59 |
| shenki | the downside is it's xilinx specific | 06:59 |
| mithro | shenki: there are open implimentations of it from what I can see | 06:59 |
| mithro | don't know how good / compatible they are | 06:59 |
| mithro | Using something like PIC / AVR / 8052 would mean we have better compiler support but more of an impedance mismatch with FPGA (hence possible more LUT usage?) | 07:00 |
| shenki | upside is that it's easy to use, and i have experience with it | 07:00 |
| shenki | an external PIC/AVR/8051? | 07:01 |
| shenki | https://www.youtube.com/watch?v=mnwTCfGyeqk | 07:01 |
| tpb | Title: Andreas playing Mario on a Gameboy emulator running on 4 Microblaze CPUs - YouTube (at www.youtube.com) | 07:01 |
| shenki | or you mean a soft one? | 07:02 |
| shenki | if we were going to put a core on there, you'd want it to be something that's easy to program | 07:02 |
| shenki | otherwise why are you using a core instead of coding HDL | 07:03 |
| mithro | shenki: soft one | 07:03 |
| shenki | ok | 07:03 |
| shenki | so that would mean something where you can go apt-get install foo-bar-gcc && make && make run | 07:03 |
| mithro | shenki: AVR/8051/PIC all have support via tools like sdcc | 07:03 |
| mithro | ayush3504 / rohitksingh: meeting? | 07:03 |
| shenki | yeah, that's what i mean | 07:03 |
| shenki | you want to be able to just use gcc | 07:03 |
| shenki | avr is supported by gcc | 07:03 |
| ayush3504 | mithro: i'm ready | 07:03 |
| rohitksingh | mithro: ready | 07:04 |
| mithro | I'm not seeing anyone else in the call? | 07:04 |
| ayush3504 | mithro: says "This party is over..." on mine | 07:04 |
| aps-sids | slomo: hi, what could be the reason for this error? gstbasesrc.c(2865): gst_base_src_loop (): /GstPipeline:pipeline-overlay-video/GstBin:bin0/GstAppSrc:source: streaming task paused, reason not-negotiated (-4). Whole debug log is located here - http://paste.ubuntu.com/7693402/ error occurs around line 1200 | 07:32 |
| tpb | Title: Ubuntu Pastebin (at paste.ubuntu.com) | 07:32 |
| slomo | aps-sids: segment event before caps, segment having the wrong format... and it looks like the caps are also rejected by your set_caps() function or something | 07:39 |
| slomo | aps-sids: what's the pipeline? are there custom elements in there? | 07:39 |
| aps-sids | slomo: I don't think there are any custom elements. Pipeline should be there in the log. | 07:41 |
| slomo | aps-sids: can you extract it from the log for me? :) | 07:42 |
| aps-sids | slomo: also, segment before caps warning is there for other components but they still work fine. I was more concerned about the critical errors. | 07:43 |
| aps-sids | slomo: just a min. I'm on mobile | 07:43 |
| slomo | aps-sids: the caps in there are not valid 1.x caps in any case | 07:44 |
| slomo | there is no bpp, depth, *_mask anymore | 07:44 |
| slomo | it's just video/x-raw,format=somestring | 07:45 |
| aps-sids | Oh | 07:45 |
| slomo | that might be the reason why the caps are rejected | 07:45 |
| slomo | if you get me a log with GST_DEBUG=6 (not 5 as that one), i can probably tell you more | 07:45 |
| slomo | but also the event misordering that happens there is worrying | 07:45 |
| *** rohitksingh has quit IRC | 08:14 | |
| aps-sids | slomo: I removed bpp, depth, *_mask, still the error. Here's the DEBUG=6 log -- http://s000.tinyupload.com/download.php?file_id=00971454773040291985&t=0097145477304029198526811 | 08:22 |
| tpb | Title: TinyUpload.com - best file hosting solution, with no limits, totaly free (at s000.tinyupload.com) | 08:22 |
| aps-sids | thanks for helping :) | 08:22 |
| *** rohitksingh has joined #timvideos | 08:28 | |
| slomo | aps-sids: try removing endianess too ;) | 08:33 |
| slomo | aps-sids: and then a new log | 08:33 |
| slomo | aps-sids: see line 37220 and the few lines above | 08:33 |
| aps-sids | slomo: still error, uploading new file | 08:40 |
| aps-sids | slomo: http://s000.tinyupload.com/?file_id=54525262765684129696 | 08:42 |
| tpb | Title: TinyUpload.com - best file hosting solution, with no limits, totaly free (at s000.tinyupload.com) | 08:42 |
| slomo | aps-sids: video/x-raw, width=(int)320, height=(int)240, framerate=(fraction)0/1 in anything we support | 08:44 |
| slomo | allowed caps video/x-raw, width=(int)320, height=(int)240, framerate=(fraction)0/1, format=(string)AYUV, pixel-aspect-ratio=(fraction)1/1, interlace-mode=(string)progressive, colorimetry=(string)bt601; video/x-raw, width=(int)320, height=(int)240, framerate=(fraction)0/1, pixel-aspect-ratio=(fraction)1/1, interlace-mode=(string)progressive, format=(string){ RGBA, BGRA, ARGB, ABGR, AYUV } | 08:44 |
| slomo | try adding interlace-mode=progressive there | 08:49 |
| slomo | but quite curious that it's required in all cases | 08:49 |
| slomo | it should only for the conversion cases | 08:49 |
| aps-sids | slomo: this is actually a conversion case, an overlay component | 08:50 |
| slomo | why are you using alphacolor btw? | 08:51 |
| aps-sids | slomo: i am not really sure since i am porting from gst-0.10 where it was used. | 08:53 |
| aps-sids | slomo: format option is not there in pipeline. Should it be there? | 08:53 |
| slomo | aps-sids: try replacing alphacolor with videoconvert :) | 08:53 |
| aps-sids | slomo: oh,i has replaced ffmpegcolorspace with video convert as well. | 08:56 |
| aps-sids | s/has/had | 08:56 |
| slomo | yeah, ffmpegcolorspace is gone and videoconvert does the same | 08:57 |
| slomo | but alphacolor is mostly useless, was already in 0.10 | 08:57 |
| aps-sids | slomo: still no luck | 09:01 |
| slomo | good | 09:02 |
| slomo | how does it fail with videoconvert? | 09:03 |
| *** rohitksingh has quit IRC | 09:03 | |
| aps-sids | slomo: same error | 09:03 |
| aps-sids | slomo: should i send a new log? | 09:04 |
| slomo | yes | 09:05 |
| aps-sids | slomo: and should format option be there in pipeline? (I've tried both with and without) | 09:06 |
| slomo | depends on what you want to do there | 09:06 |
| slomo | i still don't know the pipeline and what problem it's trying to solve at that part ;) | 09:07 |
| aps-sids | slomo: http://s000.tinyupload.com/index.php?file_id=83297190493166896037 | 09:11 |
| tpb | Title: TinyUpload.com - best file hosting solution, with no limits, totaly free (at s000.tinyupload.com) | 09:11 |
| slomo | aps-sids: ok | 09:15 |
| slomo | aps-sids: where do these caps even come from? | 09:15 |
| slomo | video/x-raw, width=(int)320, height=(int)240, framerate=(fraction)0/1, interlace-mode=(string)progressive | 09:16 |
| slomo | i mean | 09:16 |
| aps-sids | slomo: This is a CAPS_TEMPLATE in the code. just wait, I'm pushing the code changes to repo. | 09:16 |
| slomo | aps-sids: and that's the caps on a source element? | 09:17 |
| aps-sids | slomo: I've ported Pango and PangoCairo and there might be some other problems | 09:17 |
| aps-sids | yes, source element | 09:18 |
| aps-sids | slomo: here's the commit https://github.com/aps-sids/flumotion-orig/commit/9b7c88c682cc62138a8b2dd465ef6014a89fa33e | 09:24 |
| tpb | Title: Ported Overlay component · 9b7c88c · aps-sids/flumotion-orig · GitHub (at github.com) | 09:24 |
| slomo | aps-sids: ok, so you have a custom source element | 09:29 |
| slomo | and you don't provide complete caps in it | 09:29 |
| slomo | fields are missing | 09:30 |
| slomo | especially format=BGRA (test if that works, otherwise it's ARGB ;) ) | 09:31 |
| slomo | and pixel-aspect-ratio=(fraction)1/1 | 09:31 |
| slomo | if you use the API from libgstvideo you wouldn't have to worry about that | 09:31 |
| aps-sids | okay. I should add these to CAPS_TEMPLATE right? | 09:32 |
| slomo | yes | 09:34 |
| slomo | or you just use the GstVideoInfo API to create the caps for you | 09:34 |
| slomo | instead of doing stuff with strings | 09:34 |
| aps-sids | slomo: nstreaming task paused, reason not-negotiated (-4) is gone | 09:35 |
| aps-sids | GStreamer-CRITICAL **: gst_segment_to_running_time: assertion `segment->format == format' failed is still there | 09:35 |
| slomo | yes and also the event misordering is still there | 09:36 |
| aps-sids | yeah | 09:36 |
| aps-sids | how to go about that? | 09:36 |
| slomo | when do you set the caps on the srcpad? | 09:36 |
| slomo | and are you configuring a time segment? | 09:36 |
| slomo | basically you just have to implement GstBaseSrc::get_caps() and return your caps there | 09:45 |
| slomo | and call gst_base_src_set_format() with GST_FORMAT_TIME during initialization | 09:45 |
| slomo | and after that not worry about caps or segments or any other events | 09:45 |
| *** rohitksingh has joined #timvideos | 11:35 | |
| *** FeltonChris has quit IRC | 11:47 | |
| *** tija has joined #timvideos | 16:53 | |
| *** rohitksingh has quit IRC | 17:34 | |
| *** rohitksingh has joined #timvideos | 17:34 | |
| *** CarlFK has quit IRC | 18:23 | |
| tija | rohitksingh: any issuse with the UART? | 18:25 |
| rohitksingh | tija: not yet...working wonderful in windows...for past two days I haven't ran my board...will do today...but no issues as of now! :) and thanks for the code! | 18:27 |
| tija | rohitksingh: you mentioned in your blog you had to regenerate the core? | 18:27 |
| rohitksingh | tija: no i removed the core...because i already had 50MHz clock present :) | 18:28 |
| rohitksingh | tija: maybe i should look what i may have written in the blog inadvertently | 18:29 |
| tija | rohitksingh: okay! Did you try it in linux? Btw shenki has updated the drivers for newer kernel. | 18:30 |
| tija | the code is available in his git repo | 18:31 |
| rohitksingh | tija: great! will try it tonight! | 18:32 |
| *** CarlFK has joined #timvideos | 18:41 | |
| *** ChanServ sets mode: +v CarlFK | 18:41 | |
| *** rohitksingh has quit IRC | 18:41 | |
| *** slomo_ has joined #timvideos | 18:41 | |
| *** tija has quit IRC | 18:43 | |
| *** slomo has quit IRC | 18:43 | |
| *** rohitksingh has joined #timvideos | 18:49 | |
| *** slomo has joined #timvideos | 20:03 | |
| *** slomo has joined #timvideos | 20:03 | |
| *** slomo_ has quit IRC | 20:04 | |
| *** tija has joined #timvideos | 21:22 | |
| tija | rohitksingh: Did you test the linux driver? | 21:31 |
| *** CarlFK has quit IRC | 21:31 | |
| rohitksingh | tija: will do in just few more minutes... | 21:33 |
| *** CarlFK has joined #timvideos | 21:34 | |
| *** ChanServ sets mode: +v CarlFK | 21:34 | |
| tija | rohitksingh: I am having some problems. If it works let me know. | 21:34 |
| rohitksingh | tija: sure! | 21:35 |
| *** puck has joined #timvideos | 21:41 | |
| rohitksingh | tija: can't get any output...devices comes up as ttyUSB0 and dmesg also reports fine | 21:54 |
| tija | rohitksingh:same here | 21:55 |
| rohitksingh | tija: inserting vizzini module blocks my beaglebone black's usb-serial from working too | 22:26 |
| rohitksingh | tija: actually removing the vizzini module...the Exar one enumerates as ttyACMx and is working fine | 22:28 |
| tija | rohitksingh: what is working to fpga uart or the beagle bone | 22:29 |
| rohitksingh | tija: both...beaglebone shows up at ttyACM0 and FPGA's uab-serial as ttyACM1 and i can see my project's debug output. BTW which client do you use for hex output from serial? | 22:31 |
| rohitksingh | tija: ^ that is after removing the vizzini module | 22:32 |
| tija | rohitksingh: okay will do that. btw I am use gtkterm | 22:32 |
| tija | *using | 22:33 |
| rohitksingh | tija: ah! got it! gtkterm has hexadecimal view options from "View" menu..thanks...now i can see proper values! :) previously i was using cutecom, but it wasn't good | 22:36 |
| tija | rohitksingh: good going! let me get my thing working. | 22:38 |
| rohitksingh | tija: great! | 22:38 |
| tija | rohitksingh: btw why do you need a beagle bone? | 22:42 |
| rohitksingh | tija: for rapid prototyping of AD9984A I2C initialization, and to view its status registers for debugging using I2C | 22:43 |
| tija | okay | 22:43 |
| aps-sids | mithro: please ping when you have 2 minutes | 22:53 |
| *** sarwarc has joined #timvideos | 23:32 | |
| *** rohitksingh has quit IRC | 23:35 | |
| *** sarwarc has quit IRC | 23:41 | |
| *** rohitksingh has joined #timvideos | 23:55 | |
Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!