Well, as the great (I think) v8.1 autopilot footage was released it was a bit marred by a quickly followed v9 update where Tesla improved their offering quite a bit (took away some of the new thingies since then).
More importantly, they started using all the cameras and that created quite some problems in how to capture those esp. without Tesla actually cooperating.
The solution I chose was to just limit framerate on all the cameras but main one (the storage driver on ape can handle only at most 80MB/sec). 9fps is used for the side cameras and 6 fps for fisheye (the car actually gets 36fps feeds from all cameras bu backup where it gets 30fps) and even that tends to overwhelm the drive from time to time and this is visible in still frames here and there.
One of the problem is the backup camera is actually quite a bit unreliable and in many runs there's no output captured from it. As such I decided not to collect it at all for now. (you know it's not working on your cam when on a trip the cars come behind your car real close at a traffic light and nothing shows on the IC, surprisingly the CID backup cam display still works, so Tesla decided to just paper over the old "freeze frame" issue but not hte autopilot problem of the same).
It's also notable that different cameras get different detection rates it appears and since I cannot predict which ones are which sometimes detections seem to be a little bit off - know it's most likely a sampling artifact.
@DamianXVI was hard at work meanwhile improving visualizations - we now get some basic car state, 3D bounding boxes and other smaller things.
The video combines 6 cams and the narrow while captured is not shown simply because there's no space for it (as is we are already almost at 4K with native resolutions of all cams used and one more row would bring us over). If you have any cool ideas about how to visualize this differently and fit more cams in - let me know and I can provide individual (time-synced) streams of the 7 cameras.
Keep in mind all detections are independently run on every camera with little regards to other cams so you cannot just stitch them all together into 360-like video unfortunately.
Anyway here's the video, it includes different traffic modes in a North-American megapolis, though luckily no traffic jams, try not to go too cross-eyed when trying to watch 6 video-feeds at the same time:
Also another important question that I am sure is on everybody's mind is "how does the pothole/animals/othe road hazards" are handled by autopilot. Well the bad news is it's still nto handled at all as is shown in the video below. The quality is somewhat worse than the premier video above because I did not get my capture scripts to a good enough shape by then, but I don't run into tires in the middle of highways every day, so cannot really choose from multiple options.
I have a couple more of processed clips, but they are similar in nature to the first video and I am not sure if there's any interest in publishing that?
Post whatever interesting timelines you see.
More importantly, they started using all the cameras and that created quite some problems in how to capture those esp. without Tesla actually cooperating.
The solution I chose was to just limit framerate on all the cameras but main one (the storage driver on ape can handle only at most 80MB/sec). 9fps is used for the side cameras and 6 fps for fisheye (the car actually gets 36fps feeds from all cameras bu backup where it gets 30fps) and even that tends to overwhelm the drive from time to time and this is visible in still frames here and there.
One of the problem is the backup camera is actually quite a bit unreliable and in many runs there's no output captured from it. As such I decided not to collect it at all for now. (you know it's not working on your cam when on a trip the cars come behind your car real close at a traffic light and nothing shows on the IC, surprisingly the CID backup cam display still works, so Tesla decided to just paper over the old "freeze frame" issue but not hte autopilot problem of the same).
It's also notable that different cameras get different detection rates it appears and since I cannot predict which ones are which sometimes detections seem to be a little bit off - know it's most likely a sampling artifact.
@DamianXVI was hard at work meanwhile improving visualizations - we now get some basic car state, 3D bounding boxes and other smaller things.
The video combines 6 cams and the narrow while captured is not shown simply because there's no space for it (as is we are already almost at 4K with native resolutions of all cams used and one more row would bring us over). If you have any cool ideas about how to visualize this differently and fit more cams in - let me know and I can provide individual (time-synced) streams of the 7 cameras.
Keep in mind all detections are independently run on every camera with little regards to other cams so you cannot just stitch them all together into 360-like video unfortunately.
Anyway here's the video, it includes different traffic modes in a North-American megapolis, though luckily no traffic jams, try not to go too cross-eyed when trying to watch 6 video-feeds at the same time:
Also another important question that I am sure is on everybody's mind is "how does the pothole/animals/othe road hazards" are handled by autopilot. Well the bad news is it's still nto handled at all as is shown in the video below. The quality is somewhat worse than the premier video above because I did not get my capture scripts to a good enough shape by then, but I don't run into tires in the middle of highways every day, so cannot really choose from multiple options.
I have a couple more of processed clips, but they are similar in nature to the first video and I am not sure if there's any interest in publishing that?
Post whatever interesting timelines you see.
