Required 30 min or 1 hour mp4 files, 24+ hours capacity

[dash riposki]

I'd order the super capacitor kit for the mobius, also, since you're going to be recording for extended periods.

I think the V3 lens has the best night video capability.

 

[Questions]

Thank you again.

Related to the Mobius, a 1 or 2 second gap in the video is a deal breaker. Recording in MOV is an issue as well unless there is a very fast conversion tool. The time and hassle of converting the files to mp4 is why I don't use a regular DVR. If it takes 5 minutes to create a 30 minute block then that's 4 hours spent on just file conversion for a 24 hour period. That's 4 hours 'wasted' and not reviewing the video.

Thanks for the offer on uploading the mp4 files for testing, but I've got plenty I can use.

Do we know if the MateGo has the same issue with loosing seconds between files? I see it buffers video for crash recording use, so hopefully the saved files will be seamless when merged. But it would be good to know before purchase.

Or are there any other units recommended.

Thanks again everyone for all the help on this.

Andrew

 

[2000rpm]

OK, a bit beyond my level of experience but I'll suggest it:

1. Mobius (capacitor converted)
2. 720p 30fps (narrow)
3. 128GB card
4. exFAT file system
5. File length setting: maximum

Shouldn't that theoretically allow the creation of one enormous file covering somewhere between 25-30 hours?

 

[Questions]

...Shouldn't that theoretically allow the creation of one enormous file covering somewhere between 25-30 hours?

One enormous file is acceptable, as long as there is a tool that will break it down quickly into 30 or 60 minute segments. (I haven't been able to look at Registrator Viewer yet to see if that is one of its functions.) I can't be spending a lot of time converting the video into usable file types or blocks of time.

 

[Questions]

If you want @Questions i can upload 3-4-5 of my 3 minute files from a driving session you can then DL them and toy with them in registrator viewer, but really "RV" is pretty strait forward to use, just too bad Vadim the Russian that made it passed avay.

It looks like I'll need to take you up on the file uploads. I was thinking RV would merge any mp4 file, but I find that's not the case. I'm guessing the mp4 file needs to have some sort of information that defines it as a dash cam (?). So if you don't mind, I would appreciate sample files I can try.
Thanks again,
Andrew

 

[Module 79L]

Do we know if the MateGo has the same issue with loosing seconds between files?

I think you understood it wrongly: no dashcam/action cam loses seconds between clips. What the large majority of cameras do is repeat the last second of each clip, not jump 1 second between clips. That's why the files have 1:01 mins, 3:01 mins, 5:01 mins, etc. What happens when you join clips is that at the end of each one you notice a jump back in the video.
Then there are some really seamless cameras that don't repeat that 1 second at the end of the clips. Whatever the case, you never lose any footage.

Regarding the MateGo, the files are MP4, so you'd have the conversion problem solved, and if you want above average night video quality at an affordable price, it's a good option. If you also need the audio from the recordings then you'll have to choose something else because the MateGo's audio is too low and muffled and at the distance you're recording from the sound source you'll probably only record silence.

 

[kamkar]

4 files from my SG9665GC is uploading as a rar file now, they are not mp4 format but mov.

If you are dead set on mp4 files to play with in registrator viewer i have to get a hold of my mobius and change its settings to mp4 format, thats not really a big problem aside for the camera beeing down in the back yeard.

Its slow going as MEGA.NZ only allow me to UL with 500 KB/s at the moment, kinda sad cuz thats less than 10% of what my upload speed can do

The SG9665GC camera record with a 1000 ms overlap on files.

 

[kamkar]

Here is the link to the 1.2 Gb zip file with 4 consecutive video clips.

https://mega.nz/#!ChIg3Kqa!-1KlGe4Kh-w1YxzWCAMItAWV4jEFXeRlnSeCYyqC_Sw

 

[Questions]

Here is the link to the 1.2 Gb zip file with 4 consecutive video clips.

• Downloading now. However, it may be tomorrow before I can look at them.

I think you understood it wrongly: no dashcam/action cam loses seconds between clips. What the large majority of cameras do is repeat the last second of each clip, not jump 1 second between clips. That's why the files have 1:01 mins, 3:01 mins, 5:01 mins, etc. What happens when you join clips is that at the end of each one you notice a jump back in the video.
Then there are some really seamless cameras that don't repeat that 1 second at the end of the clips. Whatever the case, you never lose any footage.

Regarding the MateGo, the files are MP4, so you'd have the conversion problem solved, and if you want above average night video quality at an affordable price, it's a good option. If you also need the audio from the recordings then you'll have to choose something else because the MateGo's audio is too low and muffled and at the distance you're recording from the sound source you'll probably only record silence.

Ah, okay. I definitely misunderstood. Thank you.

• Do I understand correctly that one of the functions of Registrator Viewer is to remove the overlapping frames in order to produce a smooth video when files are merged. If so, and the merge process is fast then I'm okay. If not then I have an issue.
  • In my video playback tool I enter the time stamp shown on the video. As vehicles pass, I press a key. The tool logs the time stamp and the key press. The time stamp the tool logs is based upon the current playback position of the video. I have to match this up with reports. If the video playback position shifts a second or so every few minutes then it's going to make matching the log to reports difficult. I need a smooth/seamless video recording.
• Audio isn't required at all, so I'm good there.

Thought / Question:

• A straight out 30 minute video file is optimal, but it could be 'close'. If I can set the max file size so that I get approximately 30 minutes of video per file then that is an option I can consider. If the videos vary between roughly 25 and 35 minutes each then I can live with that.

 

[kamkar]

Thats allright buddy, if you have anything else i might be able to help you with just hit me up, i have far too much spare time on my hands.
I will go for a drive tomorrow to my sisters house, will retrive my mobius after that so i will be able to generate some mp4 files for you if you wish, just let it record a little in my window to generate some files.

 

[Questions]

Thats allright buddy, if you have anything else i might be able to help you with just hit me up, i have far too much spare time on my hands.
I will go for a drive tomorrow to my sisters house, will retrive my mobius after that so i will be able to generate some mp4 files for you if you wish, just let it record a little in my window to generate some files.

All of you at this forum are amazing. I can't tell you how much your help is appreciated. And yes, I'd love some sample files that I can experiment with.

Thank you,
Andrew

 

[Feitelijk]

I use mencoder and cut off the first second of the sections. takes about 10 seconds for appending 5 sections.

Here, they put a rubber tube across the road

Pah. Over here we have induction loops in the highway, about every 500m on all lanes, and more importantly, they are publically available.
So if you know what you're doing you could make a live traffic map, do all sorts of statistical analysis and have a lot of fun.

Oh wait look a this: (more fun at dutch peak traffic hours)

 

[Module 79L]

Do I understand correctly that one of the functions of Registrator Viewer is to remove the overlapping frames in order to produce a smooth video when files are merged. If so, and the merge process is fast then I'm okay.

Yes, RV removes the "extra" second so there's no "jumps" in playback, and as it was already said in this thread, it's a fast process.

If not then I have an issue.

• In my video playback tool I enter the time stamp shown on the video. As vehicles pass, I press a key. The tool logs the time stamp and the key press. The time stamp the tool logs is based upon the current playback position of the video. I have to match this up with reports. If the video playback position shifts a second or so every few minutes then it's going to make matching the log to reports difficult. I need a smooth/seamless video recording.

Then that extra second won't be a problem if you keep it there (it's just a repeat of the 1st second of the next clip), it won't change the video's time stamp. However, if your tool relies on it and it has to match the reports, I think your main concert would have to be keeping the camera's clock perfectly synchronized with whatever system produces the reports.

 

[jokiin]

http://www.mediafire.com/view/4nsvpj9on2y426c/Configure-Registrator-Viewer-for-SG9665XX-series.pdf how to remove the overlap

http://www.mediafire.com/view/dp4p9gpribkz0rt/Export-files-from-Registrator-Viewer.pdf how to export

 

[Questions]

...Over here we have induction loops in the highway, about every 500m on all lanes

Loops every 500 meters, wow! I sure hope someone is doing something with that data, otherwise it's a phenomenal waste of resources. Even if they are just single loops used for counting and not loop combinations that can be used for vehicle classification, speed, volume, density, etc., that's still a lot of information.

...Then that extra second won't be a problem if you keep it there (it's just a repeat of the 1st second of the next clip)...

Nope, it would be a problem. I explained myself poorly. The time stamp recorded when I press a key should be in sync with the time stamp on the video, but it won't be if seconds are periodically added to the video.

• Say the video starts at 15:00:00 and runs for 30 minutes.
• I have a field where I enter 15:00:00
• I look at the playback position of the video file. It shows zero.
• I go to the end of the video and it shows a playback position of 1800 (basically 1800 seconds, which is 30 minutes).
• As the video plays I press keys as vehicles pass. The application has no idea what is showing on the display, but it knows the playback position of the video is 60 of 1800, so it logs 15:01:00.
• The next vehicle passes at playback position 120 of 1800, so it logs 15:02:00.
• The next vehicle passes at playback position 180 of 1800, so it logs 15:03:00.
• The video was joined at this point and an extra second was added.
• The next vehicle passes at playback position 240 of 1800, so it logs 15:04:00, but it should really be 15:03:59.
• Near the end of the 30 minutes I'll be making log entries about 10 seconds off. That's a problem.

Now actually the end playback position would be something like 1810 and not 1800 (I simplified it for the example), but I'm still interjecting extra seconds that will gradually shift off the synchronization of the time on the display and the playback position of the video that I'm actually recording. Some recorders will show an end playback position very different from 1800 for 30 minutes. I have a field for a calibration multiplier that calculates what will be the time stamp on the display for the playback position and log the time stamp accordingly. Syncing reports to the video is not a big deal. I do the comparison in Excel. It only takes a moment to find the time offset between the video's time stamp and the report's time stamp for the same vehicle. Once that offset is found then it's simply applied to all the log values and I'm golden.

Of course, with Registrator Viewer stripping out the duplicated frames then none of this is an issue. Since you were curious, I just thought I'd explain.

 

[Feitelijk]

Yeah loads of data, I m downloading 5MB every minute!

They really are for the detection of jams, then the electronic signalling is telling drivers to slow. And that's why the dutch traffic is one of the safest in the world.

 

[Feitelijk]

Even if they are just single loops used for counting and not loop combinations that can be used for vehicle classification, speed, volume, density, etc., that's still a lot of information.

They log speed and volume (intensity?), and some also vehicle length.


I like to look at the speeds in these kinds of plots:

color is speed, x-axis is time and y axis is position.
Media like to call them phantom traffic jams, but as you can see clearly, it's just a travelling wave. I Had some nice results capturing these waves by using single pixel correlation (from my fluiddynamics days, actually in the group started by Burgers!)
They travel backwards at around 20km/h.

x-axis timeshift, y-axis postion and the color represents the correlation coefficient around 51,0km.


So, what is your research?

 

[Feitelijk]

Just to show what can be done:

Yesterday there was a speed trap on 51.4 A16 right. It's the black line ther. I made speed, number of lanes, intensity and density plots to show that they were causing a jam!

 

[Questions]

Just to show what can be done...

That's sweet, really sweet. Even our best dashboards here in the states can't give that level of detail, as the gantries and loops are typically a 1600m or more apart and only on tolled roads. Some are good enough to alert the traffic center of an incident before it's called in, but that's about it. I wonder how and how quickly V2V (V2X) technology will replace loops in intelligent transportation systems.

 

[Questions]

Oh my! I just tried the MOV files sent earlier that were recorded by the Street Guardian unit. My review tool can play that video, both the individual MOV files and the merged file created by Registrator Viewer! Other MOV (and OGG) files I've tried would not play in the tool. This is awesome!

So, does the ability to play MOV files change camera recommendations from the Mobius and MateGo? Are there any others that only output MOV that I should consider? I guess most importantly are there any that output MOV and can record in 30 minute blocks? If so, and my tool will play the files then that would be terrific (although I'd want to get some sample files to assure they will play back properly).

I'm still somewhat stunned over this discovery. I don't yet know if it's a game changer or not.