VIOFO A329S - Testing/Review - RCG

[Mtz]

Maybe you had no SSD plugged in, and an empty microSD?

If somebody can use some digital stopwatch editing video to analyze my video we can see if the wake-up time is less than 2.75s tested by rcg530 on a 3CH system.
You know that LPID can be used only with SD card so I don't expect the dashcam to initialize any connection to the USB-C port, so the SSD is not affecting the wake-up time.

If my wake-up time is shorter than rcg530, maybe it is because I used the Multiplexing feature.
If this will shorten the wake-up time maybe it should be needed an option to Multiplex or not the LPID mode. Why not standard or forced Multiplexing? Because can be some situations when some user wants to keep some area around his car as private and if using Multiplexing then all areas will be recorded in one video so later he needs to crop and save the needed area from the Multiplexed video. Or to blur some part of the video but this will not look trustworthy if some official will want to analyze what was happening. If showing only two videos you can say that is everything you have.

If you will ask me, I consider the Multiplexed videos should be used in all Parking modes not only LPID because they are covering almost all sides around the car in just one video and will be more easy for every one to analyze such a video.
Another thing is that a parking situation is on a stationary mode which means the objects around the car will not have a high speed to produce motion blur so the details are not so important like the ones in driving mode when I suggest the telephoto camera to not be multiplexed for A329T and the front camera to not be multiplexed for A329S.

I am sure rcg530 will test the differences when the parking files are saved as multiplexed or individual. We already see that when only two channels are used in parking mode the wake-up time is shorter by a little.

 

[Nigel]

If somebody can use some digital stopwatch editing video to analyze my video we can see if the wake-up time is less than 2.75s tested by rcg530 on a 3CH system.

I can't see when it actually records the first frame after waking up.

I can see that the screen lights up after 27 frames, so under 1 second at 30fps.

I can see that the record light comes on after 72 frames, so 2.4 seconds, and the G-sensor indicator comes on after that, so I wouldn't be too surprised if it actually takes 2.75s until it is actually recording. However I was expecting some video buffering between the impact detection and the point when the first frame is written to memory card, so for all I can tell from your video, it might be 0.1 seconds to recording start.

I have heard a time from Viofo, which is between your time and rcg530's time.

I sort of expect that rcg530 would record a stopwatch of some type on the camera, and have another camera that isn't asleep, recording the impact plus stopwatch, so that he can actually see the time of the impact and the time recorded in the first frame, so has got the actual time, but I don't know, maybe he just looked at the record LED coming on, but that could be a second late, or a second early.

You know that LPID can be used only with SD card so I don't expect the dashcam to initialize any connection to the USB-C port, so the SSD is not affecting the wake-up time.

Correct, there is no need to initialise the USB-C or SSD.
But I don't know if it does or does not with the current firmware. There is probably potential for improvement yet.

 

[rcg530]

I have an Android tablet visible within the field of view of the dash camera's various cameras that displays the current time down to the 100th of a second. All impact tests and mode transition events [parking mode to normal recording] are triggered at the top of the minute/second/100th of a second. For example, if the current time is 09:03:50:15, I wait until the time is 09:04:00:00 to trigger the impact event or mode transition event.

I monitor the "current time" value and when it reaches the top of the minute, I hit/tap the test rig I have the dash camera mounted to for the test. Currently, I do not have a second camera "watching" the test. As the top of the next minute approaches when I need to hit/tap the test rig, I perform a mental countdown as I watch the current time on the Android tablet, so that I can hit/tap the test rig as close as humanly possible to the exact top of the minute/second/100th of a second. After performing multiple impact event tests for the parking mode type being tested, I power down the dash camera and copy the video files generated during the tests. I examine the video files to see what the "current time down to the 100th of a second" the first visible frame contains for each camera recording. The time difference between the time observed in the video file's first frame and the top of the minute/second/100th of second value is what I calculate as the time for the dash camera to wake up and start recording. Is this a perfect and 100 error free process, no. I perform the test several times to try and minimize any human error caused by my not exactly hitting the test rig at the exact top of the minute/second/100th of a second. My stated times may be slightly higher than the actual time by as much as 0.20 seconds [roughly]. I'll see if I have another camera to setup an external monitoring camera with a high FPS recording ability. I'm all for making my test results as accurate as possible! It's just how much time and effort is necessary for the additional monitoring and how much does it increase the accuracy of my existing testing process. I'll see what I can come up today.

Here's a frame grab from one of the impact tests. The "impact" occurred as close as humanly possible to 11:40:00:00. I log this as a 2.83 second delay before starting to record.

 

[Nigel]

You are probably within a few frames on your tap, given that you can see the stopwatch approaching, so don't need to react to anything. A second camera would get to the exact frame +/-1.

 

[rcg530]

I used a GoPro 10 recording at 240 fps as the "watcher" camera for the LPID impact tests I ran this morning. I found that I was tapping the test rig ever so slightly before the top of the minute [between 0.14 to 0.08 seconds before the top of the minute]. The frame used to determine the "start time" from the GoPro 240 fps video is when I see the start of physical movement in the test rig. I'm tapping the test rig just to the right of the front camera. I have the A329S setup as a 3-channel setup, LPID parking mode, but the interior camera is disabled in parking mode. In my previous post, I reported this configuration would start recording within 2.55 to 2.65 seconds. Today's tests show values within that stated time frame.

 

[Mtz]

I think you should look also at the red dot when it is blinking first time on the screen and see if it is coordinated with the front or the rear camera.

I hope your next news will be about the wake-up time in Multiplexed mode by using the same rig.

 

[rcg530]

I think you should look also at the red dot when it is blinking first time on the screen and see if it is coordinated with the front or the rear camera.

While knowing if the first red blinking dot corresponds to any of the cameras beginning their recordings, I'm not sure how helpful it is. I've always determined the recording start time by the first video frame captured by each camera. The user interface may or may not be in sync with the camera recordings. But it's definitely interesting to check out. 🙂

Here are some frame grabs from the GoPro 240 fps video during the 09:53 impact event test. The first red dot appears 2.75 seconds after the triggering impact event. The red dot appears on the screen approximately 0.13 seconds after the first video frames are captured from the cameras.

09:53:00:77 - 09:52:59:86 = 0.91 seconds / Screen turns on [red dot still off]

09:53:02:08 - 09:52:59:86 = 2.22 seconds / Mic LED turns on [red dot still off]

09:53:02:52 - 09:52:59:86 = 2.66 seconds / "Alert" symbol first appears on the screen [red dot still off]

09:53:02:61 - 09:52:59:86 = 2.75 seconds / Red recording dot first appears on screen

I hope your next news will be about the wake-up time in Multiplexed mode by using the same rig.

I haven't tested the multiplexed wake-up times. I've been working on the non-multiplexed AED parking mode behavior and power consumption when writing parking mode video files to the microSD card. The combinations that now need testing have increased since the interior camera can be turned off in parking mode, the IR LEDs can be "off" [daytime], the IR LEDs can be "on" [nighttime], and the destination for AED parking mode files can be the microSD card or the SSD.

I did find that the announcement that an impact occurred while the vehicle was parked will occur when exiting AED parking mode, but only if an impact event was detected. AED motion events do not result in any announcement when exiting AED parking mode back to normal recording mode.

 

[Mtz]

09:53:00:77 - 09:52:59:86 = 0.91 seconds / Screen turns on [red dot still off]

So in my first test on this dashcam I was looking only when the screen turned on and not when the red dot was blinking on the screen.

09:53:02:61 - 09:52:59:86 = 2.75 seconds / Red recording dot first appears on screen

So we can assume that the appearance of the red dot is marking the starting of the event recording of the LPID.

 

[viofo]

When there is a parking event, does the message and sound signal sound when exiting all parking modes or only when exiting LPID parking mode?

This feature is only for testing and will not be available in the mass production version.

 

[Dex_]

If this feature not will be released, somebody need to save this beta FW file with parking events message feature.

 

[joe384]

Nice to see they implemented this request. One idea was to enable or disable such announcement in camera settings, another was to tell how many impacts were during the parking period by a voice or by the number of the beeps. Another idea was to be a time-out for some impact to be not announced like when you are leaving the car and closing the doors, because such an action is creating a LPID file and it will be announced. Some timer like 45-60 seconds to allow people going out of the car can avoid creating an unnecessary impact file.

Some viofo dashcams already have an option to delay switching to parking mode "Enter Parking Mode Timer" 90s

Not sure if this works with the LPID mode though!

I'd be interested to find out about any changes in g-sensor sensitivity for parking mode/s - The current high sensitivity is not enough

 

[Karagandinez]

This feature is only for testing and will not be available in the mass production version.

Too bad. Many users dream of this feature.
If it were implemented in the production version of the model, the number of your users would increase.

 

[haiou_f]

When monitoring 329 parking, it is said that the night vision function is very powerful and can achieve brightness as bright as daylight, right？

 

[Karagandinez]

When monitoring 329 parking, it is said that the night vision function is very powerful and can achieve brightness as bright as daylight, right？

Post in thread 'Vortex's Viofo A329 Review'

Sep 30, 2024

When recording 1fps timelapse in SNV mode, does the rear camera have the same exposure increase as the front camera?

Starting with firmware from September 29, also the rear camera have Night Vision (SNV).
Looking at these screenshots it is like the night is day.


If you will look at the filename of the original video you will see that I already recorded more than 41,000 videos on my SSD during A329 tests.

• Mtz

• 

 

[haiou_f]

Post in thread 'Vortex's Viofo A329 Review'

Sep 30, 2024

When recording 1fps timelapse in SNV mode, does the rear camera have the same exposure increase as the front camera?

Starting with firmware from September 29, also the rear camera have Night Vision (SNV).
Looking at these screenshots it is like the night is day.


If you will look at the filename of the original video you will see that I already recorded more than 41,000 videos on my SSD during A329 tests.

• Mtz

• 

Can it still be so bright in a completely dark environment without streetlights

 

[Augustus]

When monitoring 329 parking, it is said that the night vision function is very powerful and can achieve brightness as bright as daylight, right？

Does the T800 not do this?, the X800 did it

 

[haiou_f]

Does the T800 not do this?, the X800 did it

I haven't tried it before, but the T800 parking monitor shouldn't light up like the A329, right?

 

[Mtz]

There are about two type of IR LEDs: visible LED light and almost invisible LED light. I expect the visible LED to be more powerful.
But having more LEDs or more powerful LEDs will affect the power consumption in parking mode. If they are too powerful the parking recording time will be reduced and maybe the more light provided by them is not necessary. So it must be found some balance.
Do you really need very powerful LEDs for the interior camera? If yes, what do you expect to achieve by using more powerful IR LEDs?

I'd be interested to find out about any changes in g-sensor sensitivity for parking mode/s - The current high sensitivity is not enough

Right now the G-Sensor sensitivity is at maximum I ever met on a Viofo dashcam, at least on LPID mode and I consider the setting to be very OK. You can see my above video when I am touching the dashcam with my finger.

 

[GPak]

This feature is only for testing and will not be available in the mass production version.

What is so difficult to implement this notification?

Is it because high sensitive G-setting can create unacceptable amount of false event detections, or is it because events are detected due to doors being closed after entering parking mode and doors being opened before exiting parking mode?
If the latter, then all that's needed is to ignore any parking event detected for, say, 90 seconds after entering parking mode and for, say, 90 seconds before switching to normal recording.
This could complement or replace the existing 90-second delay option for entering parking mode.

 

[Mtz]

I already wrote somewhere about this feature and its possible problems. You can add a delay after turning off the car engine and enter the parking mode and in this way to avoid closing the door when leaving the car. But you will not have any solution when you will enter the car and the camera is in LPID mode.
This announcing feature was a request, it was not originated by Viofo. Maybe Viofo added that feature to see how it is working but after analysing the overall situation they can consider that it will add more claims than benefits. Also I agree to not implement such a feature which will make some people yelling on forums or internet because they have false alarms.