Elite 9 - initial thoughts & comparison to ARC 900

[Onetenner]

Firstly, this isn't a 'review' as such, just my opinions & observations.
This 'story' started when I had need to check some recordings from my Thinkware Q850, which I'd been happy with since I fitted it just over a year ago. The problem was the registration plates were unreadable in parking mode, at night - in a well-lit street....
That led me to purchase a Thinkware U3000 Pro - for me, too invasive and the wrong form factor which led to about 30% of the image being obscured. Next, Thinkware ARC 900, currently unavailable in UK so imported from Canada, it's a 'better' fit but the images are not as good as I think they could, or maybe should be for the money.
Having looked at pre-release reviews & images, I've now installed an Elite 9, initial images are good but so far, conditions have been optimal - I'm looking for 'edge case' improvements. Time will tell.

Installation was straightforward-ish, the power connector is the same type as used by Thinkware, so happy days, just plug it in eh? No. It's wired differently.... It'd probably be electrically protected from polarity reversal (I'd hope it would!) but I'm glad my 'never assume, always check' mindset kicked in.
Unlike the ARC 900, the Elite 9 has live view, which makes positioning and alignment so much easier, the ARC 900 has a small screen but it's no use when you're trying to position a rear camera in a big car.

Connection to the camera through the phone application was straightforward - although I did have a ??? moment, I'd scanned the QR code in the box to set up the application prior to installation, then when I opened the app once the camera was connected to power, on trying to add the camera (which seems the obvious thing to do) I got an 'already registered' error message.
Overall, I have to say that the software and pairing process is more 'Apple' than 'Android' - it just works, no faffing around turning off auto-join for nearby saved networks, and no having to scan a QR code every time you want to connect to the camera, unlike the ARC 900.

One feature I do like is the ability to deselect areas of the image for park recording, I'd seen no mention of it - maybe it's a standard Blackvue feature?

It's early days yet - and I'll be hot-swapping the two cameras around over the next couple of weeks (I don't want to have both on the screen at the same time, I'd find it to visually distracting!)

Talking of hot, I have noticed there is a noticeable amount of heat generated by the Elite 9, something I'd not noticed from any of the Thinkware models - they just about generated enough to keep the screen clear around the lens, this feels like it'll keep the ice off the pan. roof too! Of course, it takes electrical energy to generate heat so I'll be keeping an eye on the state of battery charge, I noted this once the camera had been in park mode for just over an hour.

A couple of screenshots from a drive out earlier.

I've got a couple of CPL filters en-route, so hopefully that'll sort out the reflections, I can already see some improvement in daytime registration plate readability, as mentioned, for me, it's night-time static and dynamic readability that I want to gain. I'm not expecting ANPR class results and I don't want to go down the Chinese brand route although admittedly they do seem to have the edge...

 

[Onetenner]

It's worth pointing out that those screen grabs were from Onedrive's native player, I've since downloaded them and watched using VLC - the registration plates are nowhere near as blurred as above.
A short drive 'there and back' to take the dogs for their daily run in the woods produced some good results:-
White van man doing his usual....

And some perplexing ones:-

Oncoming traffic (30mph + 30mph so closing speed of 60mph) renders the registration readable but the car in front, effectively stationary as we were both moving at the same speed is overexposed.
'Contrast' that with this shot - both moving at slow speed ~10-15mph, but perfectly readable

It did occur to me that maybe this car had an anti-camera coating on the plate, but it didn't seem 'the type'...

Anyway, lots of food for thought, including the images from my ARC 900 in similar circumstances

More here if you're at all mildly interested. https://dashcamtalk.com/forum/threads/arc-900-initial-thoughts-images.59387/

 

[Dashcamuser567]

How is parking mode? I currently have a U1000 and the parking mode is fantastic. The G sensor is super sensitive which does create a lot of noise in recordings but has always caught things even a door ding. I am interested in how sensitive the sensor is and how reliable it seems.

Also is downloading clips via the phone app fast over 5ghz?

 

[TonyM]

Oncoming traffic (30mph + 30mph so closing speed of 60mph) renders the registration readable but the car in front, effectively stationary as we were both moving at the same speed is overexposed.
'Contrast' that with this shot - both moving at slow speed ~10-15mph, but perfectly readable

The Honda Jazz plate is just above the headlight cutoff, whereas the Mazda plate is lower down and within the main beam.

 

[Onetenner]

Thankfully, I've not had any recordings triggered by the impact sensor so far, all have been motion detection - which is on a par with the Q850 & ARC900, the only difference being the Elite 9 allows you do pick areas to exclude from detection. If I get any impact detections, I'll post them up - I might be able to 'sneak up' on the car whilst at work...
🙂
I suppose it's one of those things that will vary between vehicles, those with more insulation, thicker windscreens, larger bodies and longer spent on NVH tuning will inevitably suffer from a lower number of triggers from minor body impacts - think of the number of transitions through panels & differing materials the shockwave has to travel through, including the worst for transmitting 'noise', the high density glass adhesive, plus the VHB pad itself, it's a wonder anything is picked up at all.

The download speed over 5GHz is, as you'd expect, minimal - no doubt helped by the files being only ~1 minute in length. I was pleasantly surprised at the time taken to copy the files to my phone, probably comparable to copying via USB-C on the ARC 900

 

[Onetenner]

The Honda Jazz plate is just above the headlight cutoff, whereas the Mazda plate is lower down and within the main beam.

I understand your reasoning, but I'm not so sure - the Mazda is white, so lots of 'bounce' from my lights, the Honda has it's brake lights on and is a darker colour, so less saturation. I do think there is an issue with lighter coloured cars and UK registration plates. Here's a couple of screengrabs, my headlights are lower (i.e. I was further back as it became a 40mph zone), the image of the registration plate became almost readable due to the headlights of the oncoming vehicle causing the image sensor to react to the increased light.

I'm well aware of the limitations of sensors - whatever they are measuring and I don't expect any manufacturer to be able to capture registration plate images in all circumstances, catering for the 99th percentile is realistically the best anyone can strive to attain, whether they regard the UK market, with light coloured cars, refractive registration plates and LED headlights part of that 99th percentile is the real question.

Edit:- I'd also add that the car on the left in this image from the ARC 900 has it's registration plate illuminated by my headlights but is easily readable.

https://dashcamtalk.com/forum/attachments/screenshot-364-webp.88865/

The two on the right are a different matter however.

 

[Dashcamuser567]

Thankfully, I've not had any recordings triggered by the impact sensor so far, all have been motion detection - which is on a par with the Q850 & ARC900, the only difference being the Elite 9 allows you do pick areas to exclude from detection. If I get any impact detections, I'll post them up - I might be able to 'sneak up' on the car whilst at work...
🙂
I suppose it's one of those things that will vary between vehicles, those with more insulation, thicker windscreens, larger bodies and longer spent on NVH tuning will inevitably suffer from a lower number of triggers from minor body impacts - think of the number of transitions through panels & differing materials the shockwave has to travel through, including the worst for transmitting 'noise', the high density glass adhesive, plus the VHB pad itself, it's a wonder anything is picked up at all.

The download speed over 5GHz is, as you'd expect, minimal - no doubt helped by the files being only ~1 minute in length. I was pleasantly surprised at the time taken to copy the files to my phone, probably comparable to copying via USB-C on the ARC 900

Thats good to hear. A good way of testing the g sensor for me was to open and close the door while im sitting inside to hear the beep. I could very lightly close the door with my finger and the g sensor would go off on the High setting.

Hopefully your gps connection delay isnt a hardware issue. I think i saw your comment in the other thread.

 

[Onetenner]

That's unlikely to be a good test for me, the doors have double seals and 'soft close' - part of the reason I park on the periphery when shopping.
BTW, why do all the fatties get to park closest to the shop doors? they're the ones that could do with the exercise....
I've updated the firmware - I doubt it's solar flares or other weather related issues, as a check I switched on my Garmin GPSMAP64x that I keep in the car, despite not being used for a few months - and last used over 150 miles away, it got a GPS lock inside the car within 90 seconds. I have a telematics module in the car that provides tracking data (that I can access) so providing time/date/location proof, should it ever be needed (e.g. in the case of cloned registration plates) can be done, it'd be nice if the GPS data was embedded in the Video as advertised. If it turns out that the screen thickness or heating elements are the cause - then I'd counter that with 'everything else can get GPS data through it - it's a popular production vehicle, why have you not taken account of these parameters?' On my list of features, it's currently a con. for the Elite 9 and a pro. for the ARC 900.

 

[Onetenner]

Some photos from earlier tonight, both rear camera - odd that the white car is blurred out of all recognition but the darker car behind me isn't?

Street lighting was present but minimal, rear screen was dirty, but clean enough to read the plate of the following car, it was on a retail park approaching a pedestrian crossing and then a set of traffic lights so speed was probably ~20mph

 

[Onetenner]

As mentioned elsewhere, GPS lock is still taking longer than expected, even after firmware update 1.003 - something that was never a problem with Thinkware Q850 / U3000 Pro / ARC 900.
I could understand it if the unit had been powered down and moved several thousand miles, but it hasn't. The timezone and the sky according to the 'last known' set of visible satellites is the same. This is basic functionality that the cheapest Chinese chipset is capable of achieving, I have other GPS devices that will achieve a lock in seconds, or, in the case of those that have not been used in weeks or months, in a single number of minutes. Yes, I have a heated windscreen, that is thicker than average - but - the GPS receiver is (presumably) positioned optimally in the housing, the one in my phone is in my pocket, on it's side, and still manages to give an accurate position - better still, I switched on my Garmin handheld that hasn't been used in months and was last used a good distance away - it achieved a lock in less than a minute.

Maybe the default location is somewhere in the US - as the map would suggest using the Blackvue viewer, rather than starting out with the last known position, which would result in a faster lock?

 

[Dashcamuser567]

As mentioned elsewhere, GPS lock is still taking longer than expected, even after firmware update 1.003 - something that was never a problem with Thinkware Q850 / U3000 Pro / ARC 900.
I could understand it if the unit had been powered down and moved several thousand miles, but it hasn't. The timezone and the sky according to the 'last known' set of visible satellites is the same. This is basic functionality that the cheapest Chinese chipset is capable of achieving, I have other GPS devices that will achieve a lock in seconds, or, in the case of those that have not been used in weeks or months, in a single number of minutes. Yes, I have a heated windscreen, that is thicker than average - but - the GPS receiver is (presumably) positioned optimally in the housing, the one in my phone is in my pocket, on it's side, and still manages to give an accurate position - better still, I switched on my Garmin handheld that hasn't been used in months and was last used a good distance away - it achieved a lock in less than a minute.

Maybe the default location is somewhere in the US - as the map would suggest using the Blackvue viewer, rather than starting out with the last known position, which would result in a faster lock?

Have you tried turning on the dashcam outside of the car via like a power bank or something to see if the GPS locks on faster?

 

[Onetenner]

These screenshots would seem to support the US-centric GPS initial location
The first one, a few seconds after setting off - only parked for <15 minutes.

Second one, a few seconds later, presumably the short parking time meant the GPS location wasn't 'stale' and didn't require the full refresh it seems to require after parking for a few hours

 

[Onetenner]

Have you tried turning on the dashcam outside of the car via like a power bank or something to see if the GPS locks on faster?

No, but if that worked, it would only prove that the GPS implementation / receiver was sub-optimal. Many vehicles have thicker windscreens and heating elements. Some, such as lorries, have vertical windscreens so will likely have a lower view of the sky, thereby reducing the number of satellites viewed. I don't think you can blame the vehicle for not being fully compatible with the dashcam, surely the dashcam should be sufficiently designed to work with the 99th percentile of it's intended market? It's not like I've got one of the solar control screens popular in the US and JP.

Edit:- Also found this from an earlier journey, vehicle had been parked for ~1.5 hours and I'd been travelling for over 7 minutes (it's a regular journey so I know the timings) - GPS data not available.

 

[TonyM]

Some photos from earlier tonight, both rear camera - odd that the white car is blurred out of all recognition but the darker car behind me isn't?View attachment 88941View attachment 88940

Street lighting was present but minimal, rear screen was dirty, but clean enough to read the plate of the following car, it was on a retail park approaching a pedestrian crossing and then a set of traffic lights so speed was probably ~20mph

There's perhaps 40mph relative speed differential between you and the white car, compared to almost none for the car following you. That would account for the blur on the white car.

 

[Onetenner]

The white car had stopped at the junction. I do think there is a 'problem' caused by the combination of UK plates and white/light coloured cars.

 

[Dan-UK1415]

GPS has been very poor on BlackVue models since models after the DR900 range, when they stopped allowing GPS in parking mode to reduce power consumption.
They tried to improve it by adding the ability to use the external GPS antenna in the CM100GLTE module when connected in both DR and Elite ranges.
But overall they're still quite poor.

BlackVue dashcams can now take 'upto' 20 minutes of driving before they pick up a signal, and when flagging this with Blackvue they say that timescale is within their acceptable range.
They suffer from very sensitive interface, so if you have the dashcam near anything that's electrical or has it's own signal then it reduces the GPS quality on the dashcam.
For example, having close to a rear view mirror mount that also houses power cables or windscreen rain sensor etc.
You can help by positioning it a bit further away.

 

[Onetenner]

That's great info, thanks, it's a pity they insist on 'cold booting' the GPS, storing last known good position data (like every car manufacturer does in their telemetry & navigation modules) would at least get that down to double, if not single digit seconds as they could use the time and position to look for 'expected' satellites rather than start with 01,02,03....

The ARC 900 (which seems to be 'persona non grata' on this site) does at least get a fast GPS lock, like every other Thinkware dashcam I've owned.
I was planning on refitting it this weekend in place of the Elite 9, reading your comments, it might be a permanent refit - something to think about, anyway 👍

 

[Dashcamuser567]

Welp this is weird timing. My battery pack for my U1000 seems to not work anymore only 2.5 years in. Camera works from the cig port but not from the battery. I may use this as an excuse to upgrade my dashcam if leaving the car running for a bit to try to charge it doesnt work. So either U3000 Pro or Elite 9 🤔

 

[Onetenner]

Not that I have anything like the experience of some of the reviewers on here, in my opinion both the U3000 Pro & Elite 9 are underwhelming.
Having ruminated on the subject whilst out with the dogs this evening (a time I do my finest thinking!), I think I'm going to refit the ARC 900 and wait out until Vueroid come up with a more discrete version of the S1 - 4k

U3000 Pro - I like the idea of the RADAR for parking detection, the form factor of the front camera is way too big and the camera is too far away from the screen, the imahge processing doesn't meet expectations.
Elite 9 The form factor is ideal for me, parking mode doesn't detect minor impacts as well as the ARC 900 (this may be due to the mount and/or the 'solidness' of my car, either way, it's not as expected), the image processing also needs improvement compared to the competition but the big downside is the poor GPS behaviour.

Time will tell, I have Christmas & New Year to ponder a bit more - and of course, read more reviews, opinions and watch more videos...

 

[Dashcamuser567]

I was hesitant with Blackvue in the past over their weird decisions such as sticking a potato 1080p cam in the rear with their last camera when the competition was putting in 2k and even 4k. And I think they used a different SoC than the others which made videos seem soft IIRC. Thinkware isnt immune either, the U3000 non pro was a waste of money IMO and I kept waiting for a worthwhile upgrade.

Looking at the ARC 900, this seems pretty interesting. I need to do more research on what the differences are between that and the U3000 Pro other than losing radar and gaining a screen. This dashcam wasn't even on my radar tbh (pun intended).