It has to reach KITT level functionality before I think people will spring for the whole package.
Tesla has had since 1982 to catch up..
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Upcoming integrated dashcam feature... How?
- Thread starter insaneoctane
- Start date
It has to reach KITT level functionality before I think people will spring for the whole package.
Tesla has had since 1982 to catch up..
Knightshade
Well-Known Member
Not sure why it would matter.... I don't think people are using dashcams for making commercial films or anything?
If it captures the license plate of the dude who bumped your car in the parking lot who cares if the blue is "blue enough" in the picture?
If it captures the license plate of the dude who bumped your car in the parking lot who cares if the blue is "blue enough" in the picture?
chronopc
Active Member
They can make it run when the car is parked if they wanted to. If they can get the HVAC system to run automatically based on cabin temperature, I don't see why they wouldn't be able to have the cameras start recording based on a censor going off.
They can make it run when the car is parked if they wanted to. If they can get the HVAC system to run automatically based on cabin temperature, I don't see why they wouldn't be able to have the cameras start recording based on a censor going off.
MCU never sleeps and HVAC works through MCU. I'm not sure which "sensor" need to go off to start recording. Accelerometer and gyrometer are on UBlox unit in APE, so it has to be up all the time, which probably create a problem with "vampire" drain. Another problem is to how to get recorded video out since APE and MCU has limited space (with M3 it's less of a problem ), but user accessible USB v2.
MCU never sleeps and HVAC works through MCU. I'm not sure which "sensor" need to go off to start recording. Accelerometer and gyrometer are on UBlox unit in APE, so it has to be up all the time, which probably create a problem with "vampire" drain. Another problem is to how to get recorded video out since APE and MCU has limited space (with M3 it's less of a problem ), but user accessible USB v2.
voip-ninja
Give me some sugar baby
I guess it comes down to what the footage is needed for. If you are trying to ID cars by license plate in a hit and run situation then I agree you want the highest quality video possible... even with 4K though you run into some limitations around lighting and such.
In many cases the dashcam footage is just needed in a he-said/she-said sort of situation, such as someone side-swiping you but claiming you entered their lane. In those situations (which I think are the more common ones) 720P, hell 480P, would be fine.
Now up to 5 pages so I may had over looked it. Will there be an option for always on? I think the biggest benefit of a dash cam is recording any movement. An example would be if someone happen to back into your car while parked ect ect. With out this always on feature, an aftermarket dash cam still appears to be worth the expense, IMHO.
chronopc
Active Member
Knightshade
Well-Known Member
Nobody (outside of Tesla) knows this yet. I would certainly think so though.
Nobody (outside of Tesla) knows this yet. The two obvious options would be local storage (via USB port for example since the car already has that) or via the cloud (which would likely be an added charge)
Lasairfion
Member
Have you seen the IRO line of built-in style cameras? Here's an example of one for the Model S AP2
VEHICLE CCTV - TESLA - IRO 1080P WIFI 140 Degree Dashcam for Tesla Model S AP2 I suppose they would bring out a Model 3 one at some point. Perhaps when the Model 3 is started being sold abroad?
Last edited:
Previous comments mentioned Computer Vision processing just like for AP. If it can drive itself, don’t think it’s too hard to detect vibrations, movements, or any other events for parking mode. Perhaps Superior to “sensors” like BlackVue. CV is used for Wipers Auto mode also. This is why I’ve been saying that Tesla can kill the dashcam market. I just dont know time frame.
Thanks for the comments about resolution. You’re right that 720p May just be enough for capturing faces and plates in proximity. But I do believe that they can always Stitch multiple cams together in those overlap areas for an even better resolution!
Good point. They’re all same cams.Well, we know what the rear camera looks like
They are not the same cam, especially for AP2 cars, 7 cameras are RCCC cameras (Red, Gray, Gray, Gray or 3xClear) so you would see only grayish color dominated by red. AP2.5 cameras would provide the colored image tho.
The question will be if the implementation has all the devices needed to be useful. The only removable storage option is a USB stick or LTE/WiFi upload. There's definitely costs associated with LTE, currently paid by Tesla. They could make premium connectivity a requirement for any sort of LTE upload. I suspect any cloud retention would incur an additional subscription cost. I don't think the USB has the bandwidth required for any sort of continuous recording, so you would probably need the car to store some sort of 30 second buffer and let the driver choose to save the previous 30 seconds (or whatever) on demand if they're using USB.
I don't think there's any question that the car should be capable of adequate video recording while in motion, although the camera images may leave something to be desired. For parked recording, you need a sensitive shock sensor for impact or constant processing to detect motion in the camera's view. How many of the car's systems would need to be powered up and operating in order to process the images and what mechanism determines storage. I don't think there's enough LTE bandwidth upload 6 1080p streams constantly and it would definitely be cost prohibitive. What's the Wh draw going to be for the Tesla dashcam implementation vs. a dedicated unit?
I highly suspect any Tesla dashcam implementation is going to be subpar compared to a good dashcam, but maybe good enough for some people.
I don't think there's any question that the car should be capable of adequate video recording while in motion, although the camera images may leave something to be desired. For parked recording, you need a sensitive shock sensor for impact or constant processing to detect motion in the camera's view. How many of the car's systems would need to be powered up and operating in order to process the images and what mechanism determines storage. I don't think there's enough LTE bandwidth upload 6 1080p streams constantly and it would definitely be cost prohibitive. What's the Wh draw going to be for the Tesla dashcam implementation vs. a dedicated unit?
I highly suspect any Tesla dashcam implementation is going to be subpar compared to a good dashcam, but maybe good enough for some people.
Knightshade
Well-Known Member
Not sure why you think that?
Even USB 2 is 480 megabits per second (mbps), or 60 megabytes per second (MBps) on paper... and real world you're probably talking more like 70-80% of that.
that's massively more bandwidth than you need for 720p video streams- it's enough even for 8 of em
They're 720p cameras, not 1080p.... as to bandwidth... I don't think we know the speed of the LTE chipset in the Model 3, do we? LTE itself surely is capable of plenty fast enough speeds, but we've no idea if the actual service in the M3 is... (no browser to run speed tests on like the S/X)
That's the maximum signaling speed of the USB 2.0 bus, but the actual speed is significantly slower. I actually wasn't referring to the bus itself though, but typical USB flash drive speeds are very slow. I guess a high speed USB 2.0 drive would work okay for 720p streams. The class 10 card in my dash cam can handle 2 1080p streams, but the quality isn't perfect.
LTE should be capable of the bandwidth, but it seems like Tesla's deal with AT&T is definitely throttled. Browser download performance in the X is definitely throttled. Given that Tesla introduced the whole premium connectivity issue on the Model 3, I've got the impression (maybe I'm wrong) that data usage in the fleet was becoming expensive. Adding even 6 720p streams from the fleet is going to incur a large expense. Would people be willing to pay $20-50/month for upload and cloud storage? That number is obviously a guess, but Nest charges $7/month per camera for cloud storage (7 days) and that's over internet.
Personally, I find it hard to make out license plates and faces on 1080p unless they're real close.
We'll see, maybe Tesla will surprise me, but I'm really not expecting this dashcam feature to be that impressive.
Knightshade
Well-Known Member
Yeah- a quality USB2 drive will be plenty fast for 720p streams...
FWIW Blackvue, which everyone seems to love for dashcams, has cloud storage offerings.
They state their 720p uploads use 600kbps of bandwidth... so about 5mbps very roughly for 8 streams if you wanted/needed to upload all the cameras at once, assuming comparable compression of the videos.
You only get 5GB total storage from them for free...for $10/mo it goes up to 10 GB. For $20/mo it's 15GB.
So yes, it seems like folks are happy to pay that for cloud dashcam storage... (and I expect Tesla would be able to offer even better pricing)
I certainly don't expect this will be perfect on day 1 (or maybe ever)- but I expect it'll be "good enough" for most users in some reasonable amount of time.
I suppose it depends though on if this is being worked on by folks like the EAP team that do good work, or the folks who do infotainment whose product is absolutely horrible.
Similar threads
