| 02:40 | aombk | joined the channel |
| 03:22 | rton | left the channel |
| 03:30 | seaman | left the channel |
| 06:58 | davidak[m] | left the channel |
| 06:58 | anuejn | left the channel |
| 06:58 | flesk_ | left the channel |
| 09:29 | anuejn | joined the channel |
| 09:29 | flesk_ | joined the channel |
| 09:29 | davidak[m] | joined the channel |
| 09:36 | anuejn | left the channel |
| 09:36 | flesk_ | left the channel |
| 09:36 | parasew[m] | left the channel |
| 09:36 | hof[m]1 | left the channel |
| 09:36 | elkos | left the channel |
| 09:36 | derWalter | left the channel |
| 09:36 | MilkManzJourDadd | left the channel |
| 09:36 | davidak[m] | left the channel |
| 09:42 | se6astian|away | changed nick to: se6astian
|
| 09:50 | se6astian | changed nick to: se6astian|away
|
| 10:09 | Bertl_zZ | changed nick to: Bertl
|
| 10:09 | Bertl | morning folks!
|
| 10:30 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 10:30 | derWalter | joined the channel |
| 10:31 | parasew[m] | joined the channel |
| 10:34 | rton | joined the channel |
| 10:35 | supragya | joined the channel |
| 10:35 | supragya | hi Bertl!, available?
|
| 10:35 | supragya | hi BAndiT1983 ;)
|
| 10:35 | BAndiT1983 | hi supragya
|
| 10:36 | supragya | you here at this time (morning) on a week day?
|
| 10:37 | BAndiT1983 | it's holiday today, so i'm at home and tomorrow in home office
|
| 10:37 | BAndiT1983 | will work on apertus tasks, to get more progress
|
| 10:38 | supragya | working on emulation myself!
|
| 10:38 | Bertl | supragya: what's up?
|
| 10:39 | supragya | do you plan on sending meta using Gbe for video streams Bertl ?
|
| 10:39 | Bertl | would definitely be an option
|
| 10:39 | supragya | and how will interframe markers look like on high speed channel?
|
| 10:39 | Bertl | depends on the format
|
| 10:40 | supragya | suppose it is just RAW12 that comes out
|
| 10:41 | Bertl | well, 'just' raw12 won't work, as all codes would be used by the image data
|
| 10:42 | Bertl | but in case we decide to go with 'just' 12bit words, then we have to reserve one or two for marking
|
| 10:43 | supragya | any Idea how would we do it? offset calculations would work?
|
| 10:43 | supragya | what if some bits get lost (is that even a possibility) ?
|
| 10:43 | Bertl | no idea, all this depends on the actual channel
|
| 10:44 | BAndiT1983 | and dont't forget about possible header
|
| 10:44 | Bertl | for example if we use a 'simple' USB3 stream (like it is planned for the USB3 module)
|
| 10:44 | BAndiT1983 | we need some secondary control channel, so the recording app gets a signal that new data will start to arrive
|
| 10:45 | Bertl | we will certainly use some packet structure to contain the actual image data
|
| 10:45 | supragya | would it be hard to induce a header such as (marker bits| sequenceID bits) before frame data
|
| 10:45 | supragya | and the same sequenceID to be used when pushing meta on Gbe?
|
| 10:45 | supragya | so that we can correlate which meta is for which frame?
|
| 10:46 | supragya | in short: seqID|Frame and seqID|meta on different channels
|
| 10:46 | Bertl | something like this has to happen, probably time codes are a better 'id' though
|
| 10:47 | supragya | in a packet structure, would you go for meta/RAW12 piggyback (or meta in same packet as frame) so that we don't need a Gbe?
|
| 10:49 | supragya | Also, for the recorder end, should I expect camera to provide me with structured meta? or should I expect it to provide me with a simple register dump for sensor/lens controller
|
| 10:57 | supragya | *will check logs later
|
| 10:58 | supragya | left the channel |
| 10:59 | BAndiT1983 | some meta data cannot be dumped, as lens controller is not there yet and also other stuff is residing in other areas, which probably cannot be just dumped
|
| 11:05 | flesk_ | joined the channel |
| 11:05 | anuejn | joined the channel |
| 11:09 | Bertl | In general I think there are a number of options and it would be way better to 'design' data/metadata coherence independant of file formats and data streams
|
| 11:09 | Bertl | *independent
|
| 11:10 | Bertl | IMHO some kind of camera identifier (users choice) and a time code would be appropriate
|
| 11:10 | Bertl | but that is just my idea ...
|
| 11:12 | Bertl | regarding the USB3 interface, it would make sense to send metadata over the USB3 link instead of a GbE side channel, but in the beginning we won't be able to do that easily
|
| 11:12 | Bertl | so I'm optimistic that this will happen sooner or later but probably not at first
|
| 11:17 | davidak[m] | joined the channel |
| 11:17 | hof[m] | joined the channel |
| 11:17 | elkos | joined the channel |
| 11:17 | MilkManzJourDadd | joined the channel |
| 11:30 | BAndiT1983 | Bertl, that also my idea, to send the separate, later to concatenate by using some frame id
|
| 11:34 | Bertl | off for now ... bbl
|
| 11:34 | Bertl | changed nick to: Bertl_oO
|
| 12:38 | seaman | joined the channel |
| 14:21 | RexOrCine|away | changed nick to: RexOrCine
|
| 14:51 | danieel | if you seek for framing, then look up UVC (for USB) or GigEVision (for GbE), why I am again seeing reinventing the wheel?
|
| 15:21 | BAndiT1983 | danieel, and what are the licensing options?
|
| 15:21 | BAndiT1983 | one can pay a lot of money for that stuff or have to invent something himself
|
| 15:22 | BAndiT1983 | "GigE Vision is not an open protocol, and as such a special license is required to develop GigE camera drivers."
|
| 15:50 | danieeel | joined the channel |
| 15:53 | danieeeel | joined the channel |
| 15:54 | danieel | left the channel |
| 15:57 | danieeel | left the channel |
| 16:15 | Bertl_oO | thanks to proprietary solutions and licensed standards the wheel has to be reinvented over and over again ...
|
| 16:31 | danieeel | joined the channel |
| 16:32 | danieeeel | left the channel |
| 16:33 | danieeel | left the channel |
| 16:37 | Kjetil | https://xkcd.com/927/
|
| 16:44 | BAndiT1983 | Kjetil, and if 14 other standards are proprietary, what should be done then?
|
| 16:58 | BAndiT1983 | changed nick to: BAndiT1983|away
|
| 17:12 | Kjetil | make a 15th and make it proprietary
|
| 17:13 | Bertl_oO | sounds like a plan ...
|
| 17:24 | RexOrCine | Objection.
|
| 18:21 | TofuLynx | joined the channel |
| 18:23 | TofuLynx | Good Afternoon Everyone!
|
| 18:54 | felix_ | just use some weird legacy protocol that includes checksums in the stream and then encapsulate that into rtp packets? ;P
|
| 19:08 | TofuLynx | left the channel |
| 19:16 | slikdigit | joined the channel |
| 19:31 | danieel | joined the channel |
| 20:02 | tjstyle | joined the channel |
| 20:35 | slikdigit | left the channel |
| 20:36 | slikdigit | joined the channel |
| 20:47 | slikdigit | left the channel |
| 21:12 | Kjetil | RTP? Surely you mean HTTP
|
| 21:14 | TofuLynx | joined the channel |
| 21:49 | RexOrCine | changed nick to: RexOrCine|away
|
| 22:16 | TofuLynx | left the channel |