| 09:39 | BogdanXOR | left the channel |
| 10:04 | BogdanXOR | joined the channel |
| 16:39 | BogdanXOR | left the channel |
| 16:50 | BogdanXOR | joined the channel |
| 16:55 | BogdanXOR | left the channel |
| 17:00 | se6astian | MEETING TIME, who is here?
|
| 17:00 | Bertl | is here
| | 17:01 | se6astian | any news to report Bertl?
|
| 17:02 | Bertl | not much, spent most of my time investigating the HDMI MIPI bridge setup further
|
| 17:02 | Bertl | preparing for a suitable plugin board but still missing manufacturer data
|
| 17:03 | se6astian | great
|
| 17:03 | Bertl | some test showed that we can probably get away with a very reduced component count
|
| 17:04 | Bertl | without sacrificing quality that is
|
| 17:06 | se6astian | anuejn and me had a work session last thursday and made lots of progress with firmware upgrades and additions, still some things left to complete though
|
| 17:06 | se6astian | anuejn also expressed interest in evaluating the rockpi5 instead of the raspberry pi5
|
| 17:06 | BogdanXOR | joined the channel |
| 17:06 | se6astian | as it has hdmi in built in already IIRC
|
| 17:06 | se6astian | for the axiom recorder
|
| 17:07 | se6astian | you have one Bertl right?
|
| 17:07 | Bertl | yeah, I have one
|
| 17:07 | se6astian | do I remember correctly that the hdmi in was already tested but showed no signal from the beta at the time?
|
| 17:07 | Bertl | I didn't do the testing back then IIRC
|
| 17:07 | se6astian | maybe anuejn did it
|
| 17:08 | Bertl | but I certainly can test it with the new firmware
|
| 17:08 | se6astian | would be interesting to try again now with the hdmi fixes
|
| 17:08 | se6astian | if hdmi works
|
| 17:08 | Bertl | yes, but I do not see that it could be a real competition for the rpi5
|
| 17:08 | se6astian | yes please hook it up if you have time
|
| 17:08 | Bertl | costs way more, only has one HDMI input, etc
|
| 17:09 | se6astian | what are the pros/cons? not needing additional PCBs for mipi/hdmi input is a plus
|
| 17:09 | intrac | left the channel |
| 17:09 | se6astian | and the nmve ssd is also built in already IIRC
|
| 17:09 | se6astian | the slot I mean
|
| 17:09 | intrac | joined the channel |
| 17:09 | Bertl | we won't need additional PCBs, we will have direct MIPI out on the upcoming plugin
|
| 17:10 | Bertl | two MIPI channels and one HDMI
|
| 17:10 | Bertl | but MIPI might work with the rockpi as well
|
| 17:11 | BogdanXOR | left the channel |
| 17:11 | se6astian | well, we dont even know if we will get chips from toshiba no?
|
| 17:11 | Bertl | but the rock-pi 5 with cooler comes about 2-3x the price of the rpi5
|
| 17:13 | Bertl | we definitely should do a feature comparison
|
| 17:13 | Bertl | (recorder specific that is)
|
| 17:14 | se6astian | its also slightly bigger (100x72mm) than the rp5 (85x56mm)
|
| 17:15 | se6astian | yes a good overview table would be great
|
| 17:15 | se6astian | anuejn: size you brought that choice up would you dive into this?
|
| 17:16 | se6astian | *since
|
| 17:19 | se6astian | ok, anyone else who wants to share?
|
| 17:20 | se6astian | BAndiT1983 made some progress with the asciidocs auto generated documentation IIRC?
|
| 17:20 | BAndiT1983 | a bit, will adjust the pipeline
|
| 17:20 | BAndiT1983 | so we can get proper theming and additional diagrams via kroki etc.
|
| 17:23 | se6astian | great, anything else?
|
| 17:25 | BAndiT1983 | nope, not yet, but will try to resume other things soon
|
| 17:25 | se6astian | good good
|
| 17:25 | se6astian | ok then thanks everyone! MEETING CONCLUDED
|
| 17:26 | Bertl | thanks for the moderation!
|
| 17:41 | BogdanXOR | joined the channel |
| 17:48 | BogdanXOR | left the channel |
| 17:50 | BogdanXOR | joined the channel |
| 17:55 | BogdanXOR | left the channel |
| 18:07 | BogdanXOR | joined the channel |
| 18:13 | BogdanXOR | left the channel |
| 18:27 | BogdanXOR | joined the channel |
| 18:33 | BogdanXOR | left the channel |
| 18:36 | BogdanXOR | joined the channel |
| 18:42 | BogdanXOR | left the channel |
| 18:50 | BogdanXOR | joined the channel |
| 18:55 | BogdanXOR | left the channel |
| 19:00 | BogdanXOR | joined the channel |
| 19:06 | BogdanXOR | left the channel |
| 19:35 | BogdanXOR | joined the channel |
| 19:41 | BogdanXOR | left the channel |
| 19:49 | BogdanXOR | joined the channel |
| 19:55 | BogdanXOR | left the channel |
| 20:24 | BogdanXOR | joined the channel |
| 20:29 | BogdanXOR | left the channel |
| 20:37 | BogdanXOR | joined the channel |
| 20:39 | anuejn | se6astian: I can try
|
| 20:39 | anuejn | but i dont really think we should do a pro con session but rather discuss the points
|
| 20:39 | anuejn | since bertl has strong feelings for the rpi and I have feelings for the rockchip, we should probably just have a chat
|
| 20:42 | anuejn | so here is my argumentation for the rockpi:
|
| 20:42 | anuejn | I believe, that it is hardware-wise a bit easier since we do not need to develop anything new (can use the existing plugin modules)
|
| 20:43 | anuejn | In addition to that, the rockpi offers a solution for fast storage that has loads of headroom (pcie 3x4)
|
| 20:44 | BogdanXOR | left the channel |
| 20:44 | anuejn | these are my main points
|
| 20:44 | anuejn | the raspi is cheaper
|
| 20:45 | anuejn | and with some work we can probably get some more input bandwith (via 2x mipi with 4 lanes)
|
| 20:45 | anuejn | but it has a lot less output bandwidth
|
| 20:46 | anuejn | on the upside it has a working vulkan driver (which the rockpi lacks)
|
| 20:49 | anuejn | the RK3588 is roughly twice as fast as the BCM2712
|
| 20:49 | anuejn | (cpu performance)
|
| 20:50 | anuejn | (multicore; single core seems roughly the same)
|
| 20:55 | anuejn | gpu performance we cannot really compare since the axiom recorder cannot currently utilize the rockchip gpu due to no existing driver
|
| 20:56 | anuejn | I have the feeling that the rockpi _should_ be more or less just a matter of plugging it in, now that the HDMI is working
|
| 20:57 | anuejn | but it would be nice to test that
|
| 20:57 | anuejn | (and some tweaks on the axiom recorder to fix the so far untested cpu-only mode)
|
| 20:57 | anuejn | the pi is also fine, if we get write performance sorted out
|
| 20:58 | anuejn | any more thoughts, Bertl?
|
| 20:59 | anuejn | ah: the RK3588 has probably a bit more headroom to do image compression in software
|
| 20:59 | anuejn | but unclear if that benefits us much because we have a hard time getting the data into the soc
|
| 21:03 | paulk | joined the channel |
| 21:03 | paulk | left the channel |
| 21:03 | paulk | joined the channel |
| 21:11 | BogdanXOR | joined the channel |
| 21:40 | se6astian | do we benefit from the increased write performance to SSD via pcie if the input is 1080p60 max anyways?
|
| 23:03 | anuejn | no.
|
| 23:03 | anuejn | the maximum datarate we can shove over the hdmi is below 3.96 gigabit / second (495 megabyte / second)
|
| 23:04 | anuejn | so more than that we will also not write
|
| 23:27 | Bertl | anuejn: write performance on the rpi5 is fine for the PCIe NVMe
|
| 23:27 | Bertl | what we investigated recently is mostly alternatives over USB3
|
| 23:27 | anuejn | ah, i see
|
| 23:27 | anuejn | then it does not really matter
|
| 23:29 | anuejn | that is nice
|
| 23:29 | Bertl | we should also look at size/volume as se6astian already hinted and we should probably check power consumptuon
|
| 23:29 | Bertl | *consumption
|
| 23:30 | anuejn | the rockchip is probably a bit more power efficient (newer process node)
|
| 23:30 | Bertl | all in all, a comparison table (of all the devices) would be interesting
|
| 23:30 | anuejn | but I expect that both do not burn _crazy_ amounts of power
|
| 00:31 | BogdanXOR | left the channel |