[Help Please] (Oft-Freezing Nextbase 312G) => Corrupt .MOV Videos Playable In-Device but Unplayable On Computers

EJC

New Member
Joined
Jul 24, 2022
Messages
4
Reaction score
0
Location
England
Country
United Kingdom
Hello All,

(Sorry #1 if wrong sub-forum, couldn't see forum for older model 312G.)

(Sorry #2 for what will be a tl;dr first post.)

Background: elderly parents' dashcam, Nextbase 312G, bought mail-order, brand-new/boxed, approx 4 years ago on the basis of multiple recommendations. Only ever genuine Kingston Endurance 64GB SDHCs used in it. Almost continuously mounted in the car through its lifetime, but extremely low-mileage, and formatted every few weeks as per instructions (on the vinyl display-protector which has never even been peeled off).

Battery's healthy: from fully-charged, the standard test gives 20+ min run-time, though that needs qualifying. It's a case of if---big IF---it doesn't lock up.

It's been an ongoing (and growing) problem for more than a year: ignition on, then after the normal cheery-chime startup, they'd sooner or later notice the red LED would have stopped flashing, the device apparently frozen, all buttons unresponsive. So, they'd whip out a pen and Hard-Reset it. Sometimes needing several Hard-Resets before 'normal' operation resumed for good.

Same behaviour indoors on the workbench with just a USB charge cable! Unpredictably locks up.

Even tried it again this afternoon: charged, froze, hard reset, recorder-stop, menu-settings, screensaver off, exit menu, recorder-mode, then found the flashing red LED and the in-picture clock both froze after about five minutes, but the dashcam's little screen remained on more than an hour, 65 minutes, but more bizarre, actually showing the moving live scene in front of it, just with all buttons non-functional.

You can't be having to constantly monitor and reset it while driving, that's ridiculous - and dangerous!

But I see the forums littered with threads about these crappy devices freezing.

Far from impressive for a "recommended" item whose raison d'etre is supposedly Reliability of recording.

Still, in these four years there'd been no events actually requiring looking at the video footage.

At least, not until now. Couple of weeks back there was an incident, car damaged, recovery-trucked to garage, only just been repaired/returned.

So, dashcam footage? Yes and no!

The files are there. And the relevant video and audio are clearly there in some form, because with the SDHC card in the device, the recorded files play back on the dashcam's little screen and speaker.

But---big BUT, perplexing but---when transferred to a PC, first bad sign is they show without thumbnails or previews, and then the actual .MOV files will not play anywhere. And I say that as a Linux system nerd, who's tried all---ALL---the normal strategies and players, even ventured to the home of QuickTime and tried various utilities and players on a 10.6.8 Snow-Leopard Intel Mac. (For the sake of clarity, yes... mpv, VLC, kdenlive, handbrake, QuickTimePlayer, MPEGStreamClip, iMovie, FinalCutExpress, blah blah etc etc).

Even some basic commandline 'ffmpeg' - NOTHING (though admittedly I don't speak fluent ffmpeg)!

Worth repeating, despite nothing playing the corrupt files, the normal non-corrupt files DO all play fine everywhere.

Incidentally, the widely-mentioned media-repair utility 'untrunc' didn't work either, nor did mp4edit.

Which leaves me doubting there's A Solution out there publicly.

So I've started trying a little geek sleuthing myself.

The following script prints some core info from a selected file.

#!/bin/bash printf "\n\n\n" read -e -p " > Video File ? : " -i "${VIDEO_FILE}" VIDEO_FILE printf "\n\n\n" ffprobe -hide_banner -i "${VIDEO_FILE}" printf "\n\n\n" ffprobe -hide_banner -v error -show_format -show_streams "${VIDEO_FILE}" | grep -iE '(format_name|TAG:compatible_brands|codec_name=|codec_tag_string|codec_tag=|pix_fmt=)' printf "\n\n\n"

For a typical good file from the dashcam, the script's utility 'ffprobe' reveals:
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/path/to/video/files/220721_160907_084.MOV': Metadata: major_brand : qt minor_version : 0 compatible_brands: qt creation_time : 2022-07-21T16:11:08.000000Z original_format : Nextbase original_format-eng: Nextbase comment : NBDVR312G-B-R00.2 comment-eng : NBDVR312G-B-R00.2 Duration: 00:02:00.00, start: 0.000000, bitrate: 15330 kb/s Stream #0:0(eng): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1920x1080, 14739 kb/s, 30 fps, 30 tbr, 60k tbn, 60 tbc (default) Metadata: creation_time : 2022-07-21T16:11:08.000000Z handler_name : VideoHandler encoder : h264 Stream #0:1(eng): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, mono, fltp, 112 kb/s (default) Metadata: creation_time : 2022-07-21T16:11:08.000000Z handler_name : SoundHandler codec_name=h264 codec_tag_string=avc1 codec_tag=0x31637661 pix_fmt=yuv420p codec_name=aac codec_tag_string=mp4a codec_tag=0x6134706d format_name=mov,mp4,m4a,3gp,3g2,mj2 TAG:compatible_brands=qt

All pretty much standard info-overload output in the world of AV files. And that's a good file.

But the needed file, the unplayable file, the bad file, sadly that gives zilch. Well, just an error:
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x5617a4887e40] Format mov,mp4,m4a,3gp,3g2,mj2 detected only with low score of 1, misdetection possible! [mov,mp4,m4a,3gp,3g2,mj2 @ 0x5617a4887e40] moov atom not found /path/to/video/files/CORRUPTED_FILE_20220707_160148_075.MOV: Invalid data found when processing input

I've even started looking inside the binary files with a hex-editor (Okteta), and it's clear the content is very different between good and bad files: repeated text-strings and structures are either present or absent. But I'm rapidly getting out of my depth.

I've no familiarity with MP4 containers and tracks, nor this crucial "moov atom" or whatever, beyond a rapidly-gleaned understanding that these MOV/MP4 files are fragile. They need access-enabling index-data in the file, but that index is ONLY WRITTEN AFTER they've finished recording the actual audio/video tracks. Hence why MP4 files are corrupted if a recorder fails before it's written the index stuff. It's not robust and failsafe - not like, for example, the '.TS' transport-stream formats which deliver our DVB-T FreeView Digital Terrestrial Television over the airwaves and into our PVRs.

Anyway, off-topic, that's gonna the basis for another post, the question of which dashcams record .TS files.

Back to this post, the fact the dashcam itself still plays the needed video/audio shows IT IS STILL IN THERE SOMEWHERE.

It makes me suspect the software/firmware in the dashcam is by-passing normal PC media-player higher protocols and accessing the RAW AV data, probably an optimisation for speed.

Ideally, some Nextbase engineers would help here. <:big pleading eyes, like Puss-In-Boots to Shrek and Donkey:>

How can I repair/reclaim/extract the existing video from the corrupt .MOV file???

As I said, 'untrunc' couldn't manage it. So, I've even been pondering using the commandline utility 'dd' to try to do experimental tranplants/grafts of the structures/blocks I found using the hex-editor. Pondering, but not tried yet.

So, this oversize post is all to ask:

??? Can anyone chip in and nudge me closer to extracting what's needed ???

Please, pretty please!

If you've read this far you probably deserve a medal.

Hopefully someone can help.

Many, MANY thanks!
 
Hello,
Thanks for your message. I've spoken to our senior engineer and due to the inconsistencies of other brands SD cards, it sounds like you're having an SD card data writing failure. As you're using a third party SD card I would recommend trying a replacement SD card, which should resolve the issue.

For a 312G I'd recommend a Dash Cam compatible 8GB -128GB U1 Micro SD card, specifically designed for Dash Cam or CCTV usage. Using a dash cam compatible SD card instead should then allow your Dash Cam to function as designed.

For guaranteed functionality and warranty of the SD card itself, I'd suggest a Nextbase Micro SD card.

Kind regards,
Millie
Nextbase Technical Support
 
Hello,
Thanks for your message
<snips>
... SD cards ... SD card ... SD card ... replacement SD card ... Dash Cam compatible 8GB -128GB U1 Micro SD card ...
... dash cam compatible SD card instead ... SD card ... Nextbase Micro SD card ...
<snips>
Kind regards,
Millie
Nextbase Technical Support


Hello Millie @NextBaseSupport

Thank you for responding. But I suspect we're speaking different languages and levels.

Sadly you've missed some key points.

1). Kingston High-Endurance cards are Class-10 UHS-I U1 premium cards specifically for dashcams. (Parents may have naively bought the NextBase camera after recommendations, but, as a geek, I researched and RTFM before I set it up for them, which included proper card-choice from the outset.)

2). When this unreliable dashcam freezes up (non-flashing red-LED/unresponsive buttons) it seems to continue recording something to the card, but these recordings are corrupted files, unplayable on PCs.

3). Maybe you don't know what a "hex-editor" is, Millie, what it's for, what it reveals. But your "senior engineer" should appreciate exactly how we use these tools to investigate and troubleshoot.

4). The binary-data in the (recorded-whilst-frozen) corrupted file is missing the normal MP4/MOV metadata-structures and layout, but...

5). The corrupt recording plays fine in the dashcam. We've watched it on the little screen, heard it from the little speaker. The file obviously contains both Video and Audio which are clearly readable and playable by the dashcam's firmware/software.

None of this is hard for a techy to grasp.

(Unless you're cynically or incompetently engineering your device to a single brand, which even I don't think likely, then...)

Let me make it clear: the card is not the issue here.

The issue is a flaky little dashcam writing improperly formatted media files when aforesaid dashcam is throwing a wobbler.

The fact there's a "Hard-Reset" button on the side of the device is tacit admission that these dashcams are just another little computer-system which is expected to fail. Just another hardware-software device designed by fallible humans and constructed from fallible components.

"Hard-Reset" exists because you--we--know devices cannot be trusted, and that's okay. Complex chaotic real-life systems are never 100% predictable and trustworthy.

But a device marketed and hyped as a "Reliable Evidence Recorder For When You Need It" should have considered the 99.9% typical scenarios---including most of the "edge-cases" and "corner-cases"---and then planned for those. And what wasn't caught pre-release needs support with ongoing Firmware once the paying punters have done further real-world (beta-)testing for you.

The bitter pill is realising there've been deliberate design-decisions to have them "not Fail Safe" - a term which all flavours of engineer are supposed to have big on their radar.

You want to go count---not just here at Dash Cam Talk, but wider, Amazon reviews, Bazaarvoice, etc etc---just how many comments there are online saying these things freeze and lock up, utterly unreliable.

Yet time and again I read the same cop-out fob-em-off not-me-guv cut'n'paste responses.

What is it about companies, eh?

They never just hold their hand up and say "Sorry, we got it wrong."

Anyone else remember a string of scandals with corporations denying there was any problem with their products, like, say, bursting into flames---cars, phones, laundry-machines, whatever?

It's wrong-headed to ask "what's wrong with Businesses?" - the proper question is "what's wrong with those People in the businesses, running them, staffing them?" Real people who go home and sleep at night.

Anyway, back to immediate topic.

If you really want to help, Millie, here's some useful sensible tangible things you can do:

A). Get your "senior engineer" to tell us HOW your little dashcam manages to play the corrupt file, without the file even having an "ftyp" header, a "moov-atom" or an "mdat" block and whatever, eh?

B). If you can't(/won't) help me rescue/repair the Digital form (retaining HD quality), at least tell me what's the "pinout" and specs of the AV-Out socket on the 312G - online research suggests it's a 2.5mm diameter 4-pole T.R.R.S. plug/socket, is that correct? But what's the exact Length Dimension and Video/Audio connection specifics to those Tip Ring Ring Sleeve terminals? At least then I might be able to source (or solder) a cable to get the recording out of the camera, albeit in (I'm assuming) low-res Analogue form.

C). Update your Firmwares to account for these myriad Bug Reports, and introduce a TS (transport-stream) option for robust, failsafe recording.

D). Bring the Community in and go "Open-Source" - let NextBase be the Raspberry-Pi of dashcams. Look at the XDA-Developer forums, look at the Maker groups, look at myriad Linux-ecosystem projects, the amazing progress achieved by the bright minds of enthusiastic volunteers.

Always sad to see the ostrich-headed 'Business Types' trapped in their throwback proprietary mindsets, convinced they have the One True Fountainhead and monopoly on Good Ideas and troubleshooting Solutions ... utterly failing to see just how resourceful and talented the Community can be.

Make NextBase a modder's platform and you'd see things grow in ways you'd never imagined! And actually boost your little Welsh outfit's presence and reputation.

As if.

Still, here's hoping for a helpful response next time, Millie.

At least please address A). and B).

Thanks. (?)

Fingers crossed.

EJ.
 
With respect to your knowledge

While your waiting for Nextbase, can you run the card through 'Speedout 0.5' and report it's write speed.

I've not used that particular card but for modern higher resolution dashcams U3 is always recommended.

As an example, here is my U3 card write speed :

O24xllu.jpg


I think Nextbase recommend a minimum of 48MB/s
 
Last edited:
The speed of the SD card is not the most important part; its the ability to maintain a consent speed that's critically important. each of our cards are made with the same components each and every time, while other card manufacturers use slightly different components which causes a difference between cards. While one may work for 6 months others may only work for 1 weeks, we cannot grantee the function of the Dashcam with other branded SD cards.
As you are using a third party card, I would recommend that changing your SD card to a compatible SD card will resolve the issue.

If you still have any questions once using a Nextbase SD card, please let us know. However I expect the issue you're having to be fully resolved.

Kind regards,
Millie
Nextbase Technical Support
 
With respect to your knowledge

While your waiting for Nextbase, can you run the card through 'Speedout 0.5' and report it's write speed.

I've not used that particular card but for modern higher resolution dashcams U3 is always recommended.

As an example, here is my U3 card write speed :

O24xllu.jpg


I think Nextbase recommend a minimum of 48MB/s


Hello @Kremmen

Thanks for looking in. Speedout. Erm, as a Windows-only closed-source executable blob, it'll never be running here. I still say the card's irrelevant in this case, but to answer you with write-speed, from the terminal Linux-users can do:

$ dd bs=1M count=256 status=progress conv=fdatasync,fsync oflag=direct if=/dev/zero of="/path/to/NEXTBASE/DCIM/256-MiB-SPEEDTEST.ddzeros"

I've run it a few times (with card in adapter in PC slot), giving the typical result:

268435456 bytes (268 MB, 256 MiB) copied, 27.3037 s, 9.8 MB/s

Sometimes 9.7, sometimes 9.9, sometimes even 10.2.

Under 30 secs to write quarter-of-a-gigabyte of data. That's consistent with Kingston's rating as a U1 card.

What Millie posted as Required Spec, was pretty-much just quoting what's on their 312G support page: "... U1 ... specifically designed for Dash Cam ..."

More numbers to consider:

Typical file-size of the dashcam's 2-minute-duration clips (both healthy and corrupt) is only 219 MiB.

In other words, only 25s of the real-time 2 minutes is needed for the actual recorder-writing.

Three-quarters of the real-time there's no need for it to be writing anything!

The Kingston U1 card meets the stated requirements for this device.

The 312G is only an old 1080p HD dashcam, not modern 4k kit.

I believe the 312G even PRE-DATES U3, because U3 dates from 2017.

Thanks anyway.

-
 
.. SD card .. our cards .. other card manufacturers .. cards. .. other branded SD cards.
.. third party card, .. changing your SD card .. compatible SD card .. a Nextbase SD card ..

Kind regards,
Millie
Nextbase Technical Support


Sigh...

It's simple logic.

I'll say again, the card is not the issue here.

WITH CARD FULLY REMOVED this flaky device is STILL PRONE TO FREEZE.

IF NOT RECORDING---OR NO CARD AT ALL---
IT DOES NOT MATTER WHOSE BRAND!!!!!!!

DUH!!!


However I expect the issue you're having to be fully resolved.

1) - 5) ignored. A) - D) ignored.

I could have added a point number 6 in my reply to Millie, although I fear reading and comprehension is not a strong point round here.

6). Clear in the first post: when doing the battery-test, the camera was NOT RECORDING at all, but proceeded to lock up.

It was idle, merely in Video-Mode after leaving the Menu-Mode. The camera froze after about 5 minutes.

Symptoms: the clock on the LCD display stops updating and the buttons became unresponsive. Bizarrely the displayed video was still 'live' in that it showed the scene moving in front of it, for the rest of the 65 minutes till the battery ran flat.

Now, unless the camera is using the card as some sort of operational Swap- or Scratch-Memory (which would be a HORRENDOUS engineering/design decision, and for which I see no sign among the files on the cards), then---when it's not actually in Recording-Writing-Mode---the card-speed and "component-quality" become irrelevant, a red-herring, a distraction, a cynically convenient fig-leaf.

Hell, it'll lock up while sat idly charging from a USB wall-wart. Sometimes not even 20 seconds in, before the auto-screen-off, which is how I first noticed it, the fact that it hadn't auto-offed, screen still on, but clock stopped and buttons non-functional. I've seen it before.

Here, today, right now. Again. On the desk. Menu => Screensaver:OFF. Power down. CARD REMOVED FROM CAMERA. Inserted USB charging cable. Screen came on, message "PLEASE INSERT SD CARD" then it returned to the live-view. Less than 3 minutes later the clock stopped, buttons unresponsive. Had to press the Hard-Reset.

NO CARD INVOLVED, but the camera still randomly locked up.

Let me repeat: NO CARD IN CAMERA, but still it froze. THAT IS A FAULTY DEVICE, not a faulty card.

Ironically and unhelpfully, since the Hard-Reset it's been sat ticking away apparently fine, showing the live view across the room for the last half hour.

Nope. Update. Turned it off (had to go do other stuff elsewhere). Came back. Re-inserted USB cable. Same thing, "PLEASE INSERT SD CARD" then it returned to the live-view. Exactly 16:45:00. Three minutes and twelve seconds later it's frozen, stuck on 16:48:12, and no button has any effect.

NO CARD IN IT, STILL FREEZES!!!!!!!

What is it you find so difficult to grasp, Millie???????


Redeem yourself, address A) and B) in earlier previous post.

Anyone else wonder if Millie's on a bonus-scheme to shift own-brand cards?

The inconvenient truth is the camera unpredictably locks-up even when it's not actually recording. It freezes regardless of whether there's a card in use!

My gut feeling is it's a fundamental flaw in the programming of the device, the firmware, or a part of the electronic circuitry is flaky, or both.

Whatever. This NextBase is prone to failure, and it does not Fail Safe!

Oh, and in case anyone wondered, I did install the latest Firmware (as indicated in ffprobe's output above, NBDVR312G-B-R00.2) when it first appeared in 2020, and even re-installed it some months back in the vain hope it might have fixed the freezing issue for them.

So no, not cards, not batteries, the central issue is that these devices randomly fail and freeze, and while failing they write malformed media files.

My uphill struggle is to repair/extract the footage from one such malformed file, so my parents can use it for evidence.

I'd just hoped there might've been others who'd been through it and could help.

Or that the manufacturer might step up.

Shame,

EJ.
 
Hello,

Thank you for your polite messages.
You did not previously state in your messages that the dashcam had issues without the SD card inserted. As you have now updated me with this information, and the dashcam is now outside of warranty, the best way forward would be to send the dashcam in for investigation and repair at our repair centre. If your parents would like to do this, please ask them to email support@nextbase.com and our team will book the camera in for them.

Regards,
Millie
Nextbase Technical Support
 
Hello,

Nextbase has a Zero Abuse Tolerance Policy which I'm afraid I'm having to enforce in this case. Having tried to be patient and assist you with your previous messages, your aggressive, undermining and abusive behaviour contravenes our abuse policy.

We provide Technical Assistance on this forum as an extra level of Technical care to our users and to provide Nextbase users with another platform for technical help, whether inside or outside of warranty. This is in addition to our direct technical support and is a free service. We respect every member on this platform and prior to your thread, we have never had to enforce an abusive policy ban.

Our apologies that you did not want to accept our assistance, however abusive behaviour towards our Team is completely unacceptable. Due to your behaviour, we will not be able to assist you further.

The Nextbase Technical Support Team
 
Thread starter Similar threads Forum Replies Date
G 312GW 5
C 312GW 4
J 312GW 7
Back
Top