[00:00] <paulk> oh wow and it's mipi csi-2
[00:00] <paulk> another thing is to have a driver for canon lenses
[00:01] <paulk> it's a SPI-like interface
[00:01] <paulk> and there are cheap EF to C-mount adapter
[00:01] <paulk> where you can drill holes to attach cables to the pogo pins to the lens
[00:01] <paulk> because tbh optics make a huge difference
[00:02] <kbingham> Ohh driving the lenses will be interesting
[00:02] <kbingham> Absolutely.
[00:02] <paulk> but for now I'd rather build a blutooth follow-focus for the blackmagic
[00:03] <paulk> rotary encoder kind of stuff, a few preset buttons
[00:03] <paulk> nrf something with zephyr
[00:22] *** xiaer1921 has quit IRC (Remote host closed the connection)
[00:22] *** xiaer1921 has joined #linux-media
[01:52] *** xiaer1922 has joined #linux-media
[01:52] *** xiaer1921 has quit IRC (Remote host closed the connection)
[01:52] *** xiaer1922 is now known as xiaer1921
[03:19] *** xiaer1922 has joined #linux-media
[03:19] *** xiaer1921 has quit IRC (Remote host closed the connection)
[03:19] *** xiaer1922 is now known as xiaer1921
[04:05] *** xiaer1921 has quit IRC (Remote host closed the connection)
[04:05] *** xiaer1921 has joined #linux-media
[05:21] *** eelstrebor has quit IRC (Quit: Ex-Chat)
[05:22] *** xiaer1921 has quit IRC (Remote host closed the connection)
[05:23] *** xiaer1921 has joined #linux-media
[06:12] *** bingbu has quit IRC (Ping timeout: 480 seconds)
[06:13] *** bingbu has joined #linux-media
[06:32] *** ayaka has quit IRC (Quit: byte)
[06:34] *** ayaka has joined #linux-media
[06:39] <pinchartl> mriesch: knowing what streams carries what data requires acquiring that knowledge in userspace "somehow". it can be considered an intrinsic property of that particular sensor driver that userspace would need to know about. better, it could be standardized in V4L2 for categories of devices (e.g. for sensor with multi-exposure, stream 0 is always short and stream 1 always long). or it could be 
[06:39] <pinchartl> exposed through an API (e.g. read-only controls, or adding a new ioctl to discover stream properties)
[06:40] <pinchartl> frame descriptors could be extended to provide more information within the kernel, possibly
[06:40] <pinchartl> and V4L2_SUBDEV_ROUTE_FL_SOURCE was dropped but will be replaced by something similar
[06:41] <mriesch> pinchartl: i could imagine you need a similar mechanism for the sensor driver you are working on?
[06:42] <mriesch> what kind of embedded data does it transmit?
[06:51] <pinchartl> the embedded data is a few lines at the start of the image that contains information such as the exposure time used to capture the frame
[06:52] <pinchartl> Sakari has posted a patch series to add embedded data support to V4L2
[06:58] *** jarthur has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[07:08] *** djrscally has joined #linux-media
[07:09] *** ao2 has joined #linux-media
[07:12] *** xiaer1922 has joined #linux-media
[07:12] *** xiaer1921 has quit IRC (Remote host closed the connection)
[07:12] *** xiaer1922 is now known as xiaer1921
[07:28] <mriesch> pinchartl: ah yes, found it.. 
[07:29] <mriesch> is there wip on the FL_SOURCE (or ITS successor)?
[07:36] *** GBenji has joined #linux-media
[07:38] <mriesch> ah, FL_INTERNAL seems to be the new thing... should have read the cover letter first
[07:42] *** frieder has joined #linux-media
[07:47] <pinchartl> yes, that's the one
[08:03] *** mripard has quit IRC (Quit: mripard)
[08:03] *** mripard has joined #linux-media
[08:05] *** darkapex has quit IRC (Remote host closed the connection)
[08:05] *** darkapex has joined #linux-media
[09:03] *** ao2 has quit IRC (Quit: Leaving)
[09:06] *** xiaer1922 has joined #linux-media
[09:10] *** xiaer1921 has quit IRC (Ping timeout: 480 seconds)
[09:10] *** xiaer1922 is now known as xiaer1921
[10:49] *** Moarc has quit IRC (Quit: i znowu NADMUCHAŁ BALONA)
[10:59] *** ao2 has joined #linux-media
[11:14] *** ao2 has quit IRC (Quit: Leaving)
[11:22] *** Moarc has joined #linux-media
[11:22] *** Moarc has quit IRC ()
[11:36] *** xiaer1922 has joined #linux-media
[11:37] *** xiaer1921 has quit IRC (Remote host closed the connection)
[11:37] *** xiaer1922 is now known as xiaer1921
[11:46] *** Moarc has joined #linux-media
[11:46] *** Moarc has quit IRC ()
[12:04] *** Moarc has joined #linux-media
[12:10] *** Moarc has quit IRC (Quit: i znowu NADMUCHAŁ BALONA)
[12:13] *** Moarc has joined #linux-media
[12:58] *** BrianG61UK has quit IRC (Read error: Connection reset by peer)
[13:33] *** mvaittin has joined #linux-media
[13:37] <jmassot> lplc: Hi, I'm looking at how to add support for pixel mode in the MAX9671{4,7} driver
[13:39] <jmassot> lplc: after looking at the MAX96717 FRONTTOP register, in case we don't select a datatype then all the traffic is forwared to the gmsl link
[13:40] <jmassot> lplc: so the current driver could work for you, are you selecting anything with the mem_dt* registers ?
[13:44] <lplc> jmassot: hi, yes, I'm filtering only the pixel data for now. The sensor also outputs a bunch of statistics and test patterns with different data types that I don't need
[13:45] <jmassot> lplc: ok same here
[13:46] <lplc> Later, I'll work on implementing the embedded data as a different stream, in which case I'll need to alow that data type to pass as well. For now, I have it embedded in the image
[13:47] <jmassot> lplc: that's exactly what I'm doing with the vgxy61 driver on top of the sailus metadata branch
[13:48] <jmassot> lplc: https://pastebin.com/raw/WNKJP422
[13:49] <lplc> jmassot: our corporate firewalls do not allow access to pastebin, unfortunately... :/
[13:50] <jmassot> lplc: hehe I added multistream support for the vgx61 so that it correctly reports a second metadata stream on an internal pad
[13:51] <jmassot> lplc:really similar to https://git.linuxtv.org/sailus/media_tree.git/commit/?h=metadata
[13:51] <jmassot> lplc: so now that I have two streams available on the sensor I can start looking at dt filtering in the serializer
[13:52] <lplc> for just 2 streams you only need to set FRONTTOP_16 and FRONTTOP_17 with the right DT
[13:53] <lplc> to allow the data to pass
[13:53] <jmassot> lplc: yes but the annoying thing will be to check if all the selected datatype have the same bpp
[13:54] <jmassot> lplc: and we will have to maintain a table with all the mipi datatype and their bpp
[13:55] <jmassot> lplc: the other solution is not filter anything so all DT goes to the GMSL link
[14:00] <lplc> jmassot: not sure I get it why you need to check or why they need to have same bpp. By default, max96717 uses DT to know the bpp of the stream
[14:01] <jmassot> lplc: Oh I was thinking that we should not mixed different streams of different bpp in the same video pipe
[14:03] <lplc> in my case the pixel data is RAW16 and the embedded data is 8-bit
[14:03] <lplc> when I was experimenting with having embedded data separate, I could see the data just fine on a MIPI analyzer
[14:07] <pinchartl> lplc: you have a CSI-2 analyzer ? lucky you
[14:08] <lplc> pinchartl: I know :) 
[14:09] *** eelstrebor has joined #linux-media
[14:10] <jmassot> lplc: pinchartl: I want one too :)
[14:11] <pinchartl> I'm sure kbingham has a camera module or two he would like you to test ;-)
[14:12] *** mvchtz has quit IRC (Ping timeout: 480 seconds)
[14:12] <jmassot> lplc: so I can prepare something for the pixel mode which filter the dt using mem_dtx_sel 
[14:13] <jmassot> lplc: one other question in pixel mode is how to retrieve the stream id properly on the deserializer 
[14:16] <lplc> don't remember which deserializer you were using, but the one we're using (max96724) has 16 filter registers per video pipe and you can actually reassign an incoming dt on a vc to another vc and/or dt
[14:20] <jmassot> lplc: I don't plan to reassign dt/vc on the max96714 but there is a stream id to select at the deserializer side. In fact it just works because the max96717 and max96714 are using '2' (Pipe Z) by default, but I may be lucky
[14:21] <jmassot> lplc: looks like that by default on max96724 each pipe can select all stream id  VIDEO_PIPE_EN (0xF4) bit 4
[14:34] <kbingham> Oh - just ship the CSI2 analyser to me. I'll look after it indefinitely... it's no bother.
[15:09] *** xiaer1922 has joined #linux-media
[15:15] *** xiaer1921 has quit IRC (Ping timeout: 480 seconds)
[15:15] *** xiaer1922 is now known as xiaer1921
[16:27] *** greenjustin has joined #linux-media
[16:43] *** jarthur has joined #linux-media
[17:02] *** xiaer1921 has quit IRC (Remote host closed the connection)
[17:02] *** xiaer1921 has joined #linux-media
[17:05] *** mripard has quit IRC (Remote host closed the connection)
[17:13] *** mvchtz has joined #linux-media
[17:17] *** GBenji has left 
[17:42] *** gouchi has joined #linux-media
[17:46] *** xiaer1921 has quit IRC (Remote host closed the connection)
[17:47] *** xiaer1921 has joined #linux-media
[18:31] *** ao2 has joined #linux-media
[18:48] *** xiaer1921 has quit IRC (Remote host closed the connection)
[18:49] *** xiaer1921 has joined #linux-media
[19:04] *** xiaer1921 has quit IRC (Remote host closed the connection)
[19:04] *** xiaer1921 has joined #linux-media
[20:08] *** xiaer1922 has joined #linux-media
[20:13] *** xiaer1921 has quit IRC (Ping timeout: 480 seconds)
[20:13] *** xiaer1922 is now known as xiaer1921
[20:16] *** BrianG61UK has joined #linux-media
[20:21] *** gouchi has quit IRC (Remote host closed the connection)
[20:42] *** gouchi has joined #linux-media
[21:03] *** frieder has quit IRC (Remote host closed the connection)
[21:11] *** ao2 has quit IRC (Quit: Leaving)
[21:29] *** gouchi has quit IRC (Quit: Quitte)
[22:12] <ndufresne> hverkuil: hello, had a note in my todo that you ask to add visl in https://git.linuxtv.org/v4l-utils.git/tree/contrib/test/test-media . Considering this is a wrapper on top of v4l2-ctl and we have no plan to implement stateless codecs in there, I'd like to drop this task. SebastianCollabora and I will pick it up anyway later in some form when we
[22:12] <ndufresne> have the FDO infra ready. We already have that running in GStreamer CI https://gitlab.freedesktop.org/seungha.yang/gstreamer/-/jobs/53770557
[22:17] *** xiaer1922 has joined #linux-media
[22:22] *** xiaer1921 has quit IRC (Ping timeout: 480 seconds)
[22:22] *** xiaer1922 is now known as xiaer1921