ALL Dash Cams Have Distortion- "Bending Images"

[jokiin]

dewarp is done by the processor, it is not edited, you can do dewarp in post processing and that would be considered edited footage

 

[synoptic12]

not at all



Haven't seen the C3 myself, I have a C1 and it works well, I feel Innovv have enough resources and experience to get good results from the C3 so expect it should be much the same performance wise, just a different form factor

I've seen a couple of reports of heat issues with the Mobius, keep in mind they've sold 20,000 or so, given the numbers I would say heat related issues are the exception rather than the rule

Can't the manufacturers employ some type of heat sink? With the size of the dash cams, one would think that size would be a factor by including this component. Cannot anyone find a way to dissipate the heat generated from these dash cams. Again, the most important factor remains heat.

 

[synoptic12]

dewarp is done by the processor, it is not edited, you can do dewarp in post processing and that would be considered edited footage

Don't you guys ever sleep? Thank you for the information.

 

[jokiin]

Can't the manufacturers employ some type of heat sink? With the size of the dash cams, one would think that size would be a factor by including this component. Cannot anyone find a way to dissipate the heat generated from these dash cams. Again, the most important factor remains heat.

It has a heat sink already

 

[synoptic12]

It's more a result of the sensor chips being flat with a rectangular grid of pixels instead of being matched to the image produced by the lens. Having collected the data from that rectangular grid of pixels the only way to get it straight is to reorganise the pixels in software which has to result in a fair amount of real data getting lost. Throwing away data from a dashcam image in order that the image looks straight instead of warped doesn't make sense as that lost data may have been useful, better to leave it warped, you can always dewarp it later on your computer. Having said that, the next generation of sensors may well have far more pixels than get recorded in the data stream thus allowing data to be lost during dewarp while still filling the full bitrate.

That very well may be, but most heat sinks are not very efficient. You mentioned earlier that a narrow lens wouldn't suffer the drawbacks of a wider lens, "but would not be viable as a dash cam". I find that to be basically correct in the field of scientific properties, but differ on the end result. Employing a 35 mm camera has no diverse effects in using the camera for a dash cam. Getting back to the Russian forums, one Russian suggested 'taping' the camera to the dashboard. He presented this aspect on his own, without any initial argument from any.

Seeing that you are awake, and you have provided an opinion on some possibilities, what camera( dash cam) would have the least 'barrel distortion'?

 

[synoptic12]

dewarp is done by the processor, it is not edited, you can do dewarp in post processing and that would be considered edited footage

Whatever "dewarp" is, certainly not a word; I was not aware that the processor calculated this fact into memory. More so, I have no knowledge of this process in post editing. If this feature is present in the initial stages by the 'processor', the remedial returns are 'poor' in my personal opinion. Every dash cam suffers from this fault, and plainly stated, the corporations, or entities making these products, just wish to turn a huge profit. Again, the circuit boards are very cheap. The many just seek to have a working unit.

 

[Nigel]

I'm unaware of the information you presented, but understand the synopsis fully. It stands to reason that pixelation is the main ingredient on rendering video. To correct video footage using the process to 'dewarp' is unfamiliar to me. This would be more work, using the time in correcting faults. Not that I wish to delve upon litigation; but editing,or re-editing video would not suffice in presenting evidence in a court of law. Now, you may have a clearer picture on the end result we are trying to achieve.

What would you suggest, or advise in a dash camera? In your writing, I can notice that you may be from England by certain wording you use. I have acquired some very good information from the 'Brits', more so than in the U.S.

I don't understand your requirement. In my view the fact that the images are warped isn't an issue as far as being used as evidence in an accident, it is an issue if you are making a movie that you want to look right. The warped images are not actually wrong, they are optically correct for the lens used and so can provide decent evidence and even be used to accurately measure the scene if done by an expert, they just don't give the image that our brains expect. If they were dewarped in the camera then that would not be counted as editing when used for evidence any more than adjusting the exposure, although it would make it more difficult to take measurements of the scene and you would loose some clarity in the corners of the image, even if done later on the computer then I doubt it would be an issue as long as there is proof of what processing has been done.

The question you need to answer is - what is wrong with the warped images? Only then can someone help...

 

[synoptic12]

Whatever "dewarp" is, certainly not a word; I was not aware that the processor calculated this fact into memory. More so, I have no knowledge of this process in post editing. If this feature is present in the initial stages by the 'processor', the remedial returns are 'poor' in my personal opinion. Every dash cam suffers from this fault, and plainly stated, the corporations, or entities making these products, just wish to turn a huge profit. Again, the circuit boards are very cheap. The many just seek to have a working unit.

P.S. As you mentioned, the 'form factor' is the only difference. Most every(good) dash cam has the ambarella, and novatek. In this respect, there is no difference in quality of components. This would border upon the, 'foundary' where the components are assembled, "quality control". The Chinese are laughing, as they are highly intelligent in the electronics industry.

 

[jokiin]

This function was done in editing program's long before it became a feature of the processor, saying it's poor is odd since you've not seen it implemented as yet, if you think all circuit boards are cheap then you need to look further afield than whatever you've looked at so far, in $88 public model product though you are going to see product that's built to a price, not a standard

 

[russ331]

...Mobius (heat issue for sure)?

I thought we'd buried this urban myth ages ago. Mobius has been on sale for over a year, & has proven it's performance, reliability, toughness & exceptional value for money in a tiny discreet package, to tens of thousands of owners.

Mobius does not have a heat issue. The video you refer to is a freak incident that is irrelevant & should be ignored. People persist in referring to Mobius cameras being hot to the touch. The camera body doesn't get hot, the exposed heat sink does, which it's supposed to while dissipating heat from the DSP. People should stop touching the heat sink!

The only thing Mobius requires for it's reliable operation is an external power supply of at least 1A.

The standard Mobius A lens looks suitable for your needs. It's 85° HFOV (Horizontal Field of View) is user-adjustable to 75° HFOV in 1080p/30 mode, 46° In 720p/60 mode.

 

[jokiin]

Still have no idea what a foundry is

 

[synoptic12]

I don't understand your requirement. In my view the fact that the images are warped isn't an issue as far as being used as evidence in an accident, it is an issue if you are making a movie that you want to look right. The warped images are not actually wrong, they are optically correct for the lens used and so can provide decent evidence and even be used to accurately measure the scene if done by an expert, they just don't give the image that our brains expect. If they were dewarped in the camera then that would not be counted as editing when used for evidence any more than adjusting the exposure, although it would make it more difficult to take measurements of the scene and you would loose some clarity in the corners of the image, even if done later on the computer then I doubt it would be an issue as long as there is proof of what processing has been done.

The question you need to answer is - what is wrong with the warped images? Only then can someone help...

You present a logical explanation of defining the so-called principles in 'dewarping'. Actually, there is nothing wrong with the 'warped' images, other than those developing these dash cams are highly aware of the setbacks. Money remains a prevalent factor in production; 'keeping cost low'. As a consumer, I do not appreciate this type of production, whereas improvements can be upgraded. Nothing is more important than 'money' for the producers of this product. That is the bottom line.

Recommend a dash cam, could you?

 

[synoptic12]

This function was done in editing program's long before it became a feature of the processor, saying it's poor is odd since you've not seen it implemented as yet, if you think all circuit boards are cheap then you need to look further afield than whatever you've looked at so far, in $88 public model product though you are going to see product that's built to a price, not a standard

Military specs would suffice.

 

[maccardiolog]

Whatever "dewarp" is, certainly not a word; I was not aware that the processor calculated this fact into memory. More so, I have no knowledge of this process in post editing. If this feature is present in the initial stages by the 'processor', the remedial returns are 'poor' in my personal opinion. Every dash cam suffers from this fault, and plainly stated, the corporations, or entities making these products, just wish to turn a huge profit. Again, the circuit boards are very cheap. The many just seek to have a working unit.

Because I take an interest in this issue (a dashcam with minimal distorsion AND day and night video quality ) , I could enumerate some of them : G1W , Mobius A lens , Vico TF EZ , Lukas - 7900 ACE , Transcend DrivePro 200 , Innovv . Generally speaking , some dashcams with less than 140 degree lens .

 

[maccardiolog]

Because I take an interest in this issue (a dashcam with minimal distorsion AND day and night video quality ) , I could enumerate some of them : G1W , Mobius A lens , Vico TF EZ , Lukas - 7900 ACE , Transcend DrivePro 200 , Innovv . Generally speaking , some dashcams with less than 140 degree lens .

And the new one G90

 

[Nigel]

You present a logical explanation of defining the so-called principles in 'dewarping'. Actually, there is nothing wrong with the 'warped' images, other than those developing these dash cams are highly aware of the setbacks. Money remains a prevalent factor in production; 'keeping cost low'. As a consumer, I do not appreciate this type of production, whereas improvements can be upgraded. Nothing is more important than 'money' for the producers of this product. That is the bottom line.

Recommend a dash cam, could you?

What I am saying is that there is nothing wrong with the warped images and since they contain more information than de-warped ones they are preferable for use as evidence. It is not an issue of cost or quality, they are left warped because that is better for a dashcam than what is currently possible with de-warping.

 

[synoptic12]

Because I take an interest in this issue (a dashcam with minimal distorsion AND day and night video quality ) , I could enumerate some of them : G1W , Mobius A lens , Vico TF EZ , Lukas - 7900 ACE , Transcend DrivePro 200 , Innovv . Generally speaking , some dashcams with less than 140 degree lens .

You did not mention the FineVu CR500 HD. I am well aware of the cams you have listed; researching them thoroughly. Tell me about the G90.

 

[synoptic12]

What I am saying is that there is nothing wrong with the warped images and since they contain more information than de-warped ones they are preferable for use as evidence. It is not an issue of cost or quality, they are left warped because that is better for a dashcam than what is currently possible with de-warping.

There may be 'nothing' wrong with warped planes to the many, but we see this as a very distracting feature. Most eyes are focused upon the 'action' (main focal point), not really the adverse effects to say. This is where the developer banks on selling their units. In all reality, if it still works why change, right?

 

[maccardiolog]

You did not mention the FineVu CR500 HD. I am well aware of the cams you have listed; researching them thoroughly. Tell me about the G90.

I did not mention because of lower bit rate and higher cost . I am not a specialist , I only inform myself from this excelent forum . I don't have a G90 , I only have one Mobius A lens with battery .

 

[Dashmellow]

This function was done in editing program's long before it became a feature of the processor, saying it's poor is odd since you've not seen it implemented as yet, if you think all circuit boards are cheap then you need to look further afield than whatever you've looked at so far, in $88 public model product though you are going to see product that's built to a price, not a standard

Reading through this thread I was going to mention the same thing. "Dewarping" is essentially a camera firmware equivalent of a built in Photoshop filter. There are Photoshop filters that can deliberately create a pronounced fish-eye effect or do the inverse and correct for barrel distortion. The inverse is called "pin-cushioning". Of course, in Photoshop you can control the amount of effect you want but in a dash camera it has a singular purpose.