DSLR vs dashcam

[Outbacknomad]

A comparison between 2008 Canon 5D2 35mm sensor with 14mm lens 5616x3744 pixels vs Garmin 56 2560x1440 pixels.

14mm lens FOV


Garmin 56 FOV


14mm crop & scaled down x2.2


Garmin crop.


This was a simple comparison simply to see what basic differences these 2 cameras have. One camera is over a decade old with large sensor & the other is a modern camera with tiny sensor for a car.

 

[Nigel]

Am I right in thinking that you cheated on the Canon and used photo mode instead of video mode? Which of course would be useless for security recordings!

The Garmin has more fisheye distortion, but that doesn't affect its use as a dashcam, so no need to reduce its score for that.

The Canon has less noise, but it appears to be through more noise reduction, and it has removed the texture details in the bricks, so maybe a win for the Garmin on noise if we are looking to record plenty of detail?

The Canon has better exposure for the highlights and sky, but not a big difference.

So the only big difference is in the colour "noise", Garmin has failed to correct the chromatic distortion, maybe a lack of attention to detail rather than a hardware issue, and there is far more colour noise from the compression, but that is from using too low a bitrate for the video compression while the Canon is I think cheating by using photo mode!

Ohh, and even though the Canon is only available in well used 10 years old form, it is still twice the price, and comes without a lens! The lens may also cost more than the Garmin?

 

[Outbacknomad]

The 14mm F2.8 lens is about au$2600 so a touch more than a dashcam, then you need a DSLR! The two photo were simply photos. The primary use of the DSLR is photos, certainly in 2008.

The chromatic aberration & distortion can be reasonably removed and remapped from the dashcam image, I used to do this with pre digital medium and large format cameras, even with basic LAB lightness & AB channel editing a lot can be improved compared with RGB editing.

The dashcam is easily more than adequate for its purpose & stands up amazingly well against a DSLR.

At the end of the day one does not need pretty pictures from a dashcam.

 

[Nigel]

The chromatic aberration & distortion can be reasonably removed and remapped from the dashcam image,

Can be much improved on that by the dashcam processor itself if they had bothered during development, unless their lenses are inconsistent.

However removing the distortion is generally not desirable since it will cause some blurring of detail towards the edges of the image, and for a dashcam, reading plates is more important than having straight walls to buildings, which logically shouldn't actually be straight anyway on a very wide angle image, even if photo cameras do all try to make them straight.

You really don't need a big sensor for a wide angle image, and there are some disadvantages such as depth of field becoming too small for dashcam use, unless you then use a very small aperture, which then results in motion blur when you start moving.

 

[Dashmellow]

Can be much improved on that by the dashcam processor itself if they had bothered during development, unless their lenses are inconsistent.

However removing the distortion is generally not desirable since it will cause some blurring of detail towards the edges of the image, and for a dashcam, reading plates is more important than having straight walls to buildings, which logically shouldn't actually be straight anyway on a very wide angle image, even if photo cameras do all try to make them straight.

You really don't need a big sensor for a wide angle image, and there are some disadvantages such as depth of field becoming too small for dashcam use, unless you then use a very small aperture, which then results in motion blur when you start moving.

That could really be dealt with using better corrected lenses with less or even zero barrel distortion rather than digital correction but that would likely push a dash camera out of its intended price point.

 

[Outbacknomad]

When I was using large format, 5 x 4 inch Velvia transparency (film loaded individually in complete darkness into film backs) I would often use f64 to get enough depth of field for architectural work. Still today I look back at the quality of those images & they are impressive & have a "look" one does not get today.

But like everyone else I went digital early simply from the convenience. In the early days using raw files and stacking images to squeeze every last drop out of an image, or using a shift lens to increase resolution. Often doing a modestly zoomed in section in the middle as the edges on an image people would not really notice. Remapping RGB channels to squeeze everything out of a lens & using LAB An awful lot of processing, but it was worth it.

Quite a few tricks done in early digital images to give a far higher quality image than what might first appear.

If you saved a raw file in the early days even though you may have used a jog. If you have a few critical family images. Go back & process those original photos. You will be surprised of what can be fixed with the original raw data file.

It is impressive what you get out of cameras with tiny plastic lenses these days.

 

[Dashmellow]

I too used to do large format architectural work using Fuji Velvia. Wonderful film! I miss those days! In many ways, large format transparencies offered Ultra High Definition, tonal range and depth that even today can't really be matched by digital. It has a different quality altogether. I used to joke that the difference between transparency film and digital was like the difference between gourmet coffee and instant coffee but things have come a long ways. I held off going digital for quite some time, relying on 6x7 medium format with occasional digital DSLR work. That was partially because I had a lab doing my own color printing and digital printing wasn't quite as capable, affordable or practical as it is nowadays.

 

[kamkar]

My niece sure do appreciate the pictures i have taken of her dogs, and then get printed in 60 X 90 CM format to be put in a frame.
She never said ( as a phone generation kid ) but clearly the output of my rather low end Nikon Dslr impress.

The prints are actually pretty cheap, when i get a place to live where i would actually bother putting something on the walls, i will have to troll thru my picture archive, and maybe snap a few new ones too.

 

[Dashmellow]

I have 30 x 40 inch prints from 4x5 inch internegatives made from Velvia transparencies. Those would impress your niece!

 

[Outbacknomad]

I too used to do large format architectural work using Fuji Velvia. Wonderful film! I miss those days!

The ultimate drug!

Also still got my door stop Mamiya 645 & 67RZ. Maybe one day I will get a digital back. There is something completely different in using these larger cameras.

Also still miss the fact that you could move & angle the film & lens plains in relation to each other with a massive range.

People today have no idea of the flexibility & scope of these cameras.

Even I am getting dumbed down as well!

 

[Nigel]

That could really be dealt with using better corrected lenses with less or even zero barrel distortion rather than digital correction but that would likely push a dash camera out of its intended price point.

With lenses of reasonable cost, you will get a sharper image doing it in software, as most photo cameras do!

Our real view of the world is spherical, straight lines should not appear straight!

 

[Dashmellow]

With lenses of reasonable cost, you will get a sharper image doing it in software, as most photo cameras do!

Our real view of the world is spherical, straight lines should not appear straight!

These are cameras, not eyeballs.

Curious that you now claim you'll get a sharper image via software when earlier you said:

However removing the distortion is generally not desirable since it will cause some blurring of detail towards the edges of the image

Nothing like a high quality, well corrected optic in my book!

 

[Dashmellow]

The ultimate drug!

Also still got my door stop Mamiya 645 & 67RZ. Maybe one day I will get a digital back. There is something completely different in using these larger cameras.

Also still miss the fact that you could move & angle the film & lens plains in relation to each other with a massive range.

People today have no idea of the flexibility & scope of these cameras.

Even I am getting dumbed down as well!

View attachment 57182

I have an RZ67 too! Fantastic camera but indeed rather heavy. Funny thing, mine came with a neck strap! It attaches to those silver lugs on the upper right corner of the camera body seen in your photo. I have the older version of the prism but also no digital back. Pricey, pricey.

Also own a Manfrotto head (and a couple of their other products).

 

[Outbacknomad]

Some more door stop material. 50mm, 75 shift lens, 90, 180, 360, 1.4 extender, Polaroid back & 3 film backs. Not to forget one really needs to carry 16kg of trypod. If only I could justify a digital back.

I had some Polaroid left over & made a cardboard pin hole camera for my niece. She was blown away with the results of our home made camera.

 

[Nigel]

These are cameras, not eyeballs.

The real problem is not in the camera, but in displaying the image onto a flat screen instead of a spherical one, no better than trying to display a map of the earth on flat paper instead of a globe!

 

[EGS]

The real problem is not in the camera, but in displaying the image onto a flat screen instead of a spherical one, no better than trying to display a map of the earth on flat paper instead of a globe!

Hey!,, soon we'll start seeing convex photo sensors. Quite practical.

 

[Nigel]

Hey!,, soon we'll start seeing convex photo sensors. Quite practical.

 

[Dashmellow]

The real problem is not in the camera, but in displaying the image onto a flat screen instead of a spherical one, no better than trying to display a map of the earth on flat paper instead of a globe!

Not sure why you are posting this sort of goofy stuff but the problem is not, "displaying the image onto a flat screen instead of a spherical one", it is correcting the image projected by the lens onto a sensor or sheet of film. That task is accomplished by using highly corrected rectilinear lenses that are optimized for throwing a flat field. We primarily view our images and videos on flat screens, after all.

 

[EGS]

It was so predictable you were going to come back with that stuff...

 

[Outbacknomad]

If you have 3D software you can project an image onto any shape and then render an image of that shape from any angle. Even through a shift lens. Before people had heaps of computing power it was the only way to divide up an image and render it on multiple computers to speed a job up.