Canon 5d mark iii improved- olpf removed for higher resolution

I suppose people will get upset with me if I point out that it looks very pink, is only 9 seconds in length & shot in bright sunshine where the 5D3 resolution is going to look at its best but sorry but I'm still not a believer.

He only has a free Vimeo account so it's impossible to download the original to look at

Permission Denied
This video cannot be downloaded anymore today

If anyone has managed to download it then please contact me as I am happy to host the raw file on one of my servers.
 
Last edited:
Agreed it's not conclusive until we see some charts but I've found everything I've shot so far with the MKIII looks good with a touch of post sharpening and that footage didn't look like it needed any.
 
It's also not possible from that clip to see whether one can still get infinity focus after the surgery.
 
I think the procedure to remove a built in filter inside a camera like 5Dmk3 is bound to cause massive problems and will compromise this camera in how it performs after such a modification and all this just to maybe gain a smidgen more resolution in video. To even suggest users to do this is something I wouldn't do. By all means hack the firmware but leave the hardware as is. I think Canon made the camera to be at its best as it is leaving the factory.

If someone wants better video then get a better video camera.
 
I agree with all you guys. The difference I see in the camera after surgery is rather small, but the very downside is the IR contamination that occurs. I don't have an extra 5D3 at my convenient disposal so the very needed side by side test is not going to happen here.
The short Andrew clip that was posted does not look like the sharpness I am getting, same with James Miller's footage. Perhaps, something different is happening with the DIY process as opposed to the MaxMax. Their stuff is showing much larger gains in sharpness.
what I don't agree with is the idea that the 5D3 is somehow either crippled or an unworthy upgrade to the 5D2. It produces a beautiful image.
bill
 
"If someone wants better video then get a better video camera."

Or remove the filter as James and Andrew have done. The difference to these eyes is drastic. Some people just like pulling things apart, and I, for one, think this often leads to progress (and it's a lot of fun).
 
nigelbb said:
I suppose people will get upset with me if I point out that it looks very pink, is only 9 seconds in length & shot in bright sunshine where the 5D3 resolution is going to look at its best but sorry but I'm still not a believer.

He only has a free Vimeo account so it's impossible to download the original to look at

Permission Denied
This video cannot be downloaded anymore today

If anyone has managed to download it then please contact me as I am happy to host the raw file on one of my servers.
Click to expand...

No, Andrew's new vid is fully downloadable. Do so and take a look.
 
nigelbb said:
Emanuel said:
Barry,

I used the word disbeliever in plural form ;) different than 'believer' or the verb 'to believe'. I also used figurative speech, yes. But the fact is that there has been posted some samples with and without the filter removed which some people didn't take them as proof of anything. Some people prefer 'to believe' in charts or as we wish, to make assumption in 'scientific' charts proof. To me and many others in this industry, charts can be misleading because can give you deceptive conclusions once many real world scenarios can't be reproduced by them.

E :-)
Click to expand...

When I first got my 5D2 over 3 years ago I would have sworn that it had better resolution than my XHA1 or HV30 until I saw some proper comparisons done then did some of my own & realised that among other things sharpening & contrast can produce false detail that gives the illusion of increased resolution. I am not prepared to trust my eyes (or anyone else's) when it comes to deciding which camera resolves more detail just by looking at one or two compressed, edited, & graded videos. Only proper evaluation with charts or proper side by side comparison can be trusted. It's because I know not only my eyes fool my brain but everyone's do the same. It's just illogical to look at a clip & say it has more detail than another & categorically that this is because the OLPF has been removed. It's literally impossible to draw that conclusion.

Unlike most of the people involved in this discussion I actually own a 5D3 & would be absolutely ecstatic if some simple surgery on my camera would improve the image quality with no side effects even if it did void my warranty. The evidence that I have seen thus far is not just inconclusive it's non-existent.
Click to expand...
I guess I should have written instead that «the fact is that there have been posted some samples with and without the filter removed which some people didn't take them as proof of anything». But I didn't. Anyway, aside the OLPF controversy, it is true the perception is part of this business. Reason enough why charts don't tell all the truth, quite the opposite. They can be misleading. Some capture device that shines on charts can be a way less interesting than the other one with a softer outcome. Here is when many real world situations don't apply for. This has been one of the big mistakes of pixel peepers. The sooner you get it, the best tool you'll have from then on.
 
nigelbb said:
I suppose people will get upset with me if I point out that it looks very pink, is only 9 seconds in length & shot in bright sunshine where the 5D3 resolution is going to look at its best but sorry but I'm still not a believer.

He only has a free Vimeo account so it's impossible to download the original to look at

Permission Denied
This video cannot be downloaded anymore today

If anyone has managed to download it then please contact me as I am happy to host the raw file on one of my servers.
Click to expand...
Why upset? : ) It is true, it is pinky. Fact. In this case, it is not opinion :-)

3hats said:
I agree with all you guys. The difference I see in the camera after surgery is rather small, but the very downside is the IR contamination that occurs. I don't have an extra 5D3 at my convenient disposal so the very needed side by side test is not going to happen here.
The short Andrew clip that was posted does not look like the sharpness I am getting, same with James Miller's footage. Perhaps, something different is happening with the DIY process as opposed to the MaxMax. Their stuff is showing much larger gains in sharpness.
Click to expand...
It is very possible something they're doing differently. They're not replacing both filters but are getting much sharpness, you say. What is rather funny to say the least.

3hats said:
what I don't agree with is the idea that the 5D3 is somehow either crippled or an unworthy upgrade to the 5D2. It produces a beautiful image.
bill
Click to expand...
Nobody here says the contrary. The facts are: you need either post processing or to partially/totally remove the OLPF in order to get extra sharpness. As well, you'll have clean HDMI output if you'll pay 4x the price ;-) Is it necessary to add anything else on the subject matter? :P


I will be buying 1-2 1D/C units as soon as the camera will be out, so I rule out as crybaby with no money for the purchase. My 2 cents of euro in behalf of the facts, in any case.
 
Last edited:
Must have been a glitch. I just downloaded the file. It's not as I expected the raw .MOV file off the camera as I see from running exiftool that it's been processed by Premiere Pro. The only part of the frame actually in focus is a small portion over on the left that looks OK. The car looks nice & shiny. The pink/red cast is more pronounced in the Vimeo version but is still present here. Apart from the colour cast it looks no different than I would expect from a 5D3 or 5D2 in bright sunshine. If it's meant to demonstrate improved resolution it's a very poor example. Just stopping it down so more of the frame was in focus would have made for a more convincing demo.

Here is an example of what video from the 5D2 can look like in bright sunshine. Download the 250MB 1920x1080 .MOV file to evaluate the full quality.

 
Last edited:
3hats said:
I agree with all you guys. The difference I see in the camera after surgery is rather small, but the very downside is the IR contamination that occurs. I don't have an extra 5D3 at my convenient disposal so the very needed side by side test is not going to happen here.
The short Andrew clip that was posted does not look like the sharpness I am getting, same with James Miller's footage. Perhaps, something different is happening with the DIY process as opposed to the MaxMax. Their stuff is showing much larger gains in sharpness.
what I don't agree with is the idea that the 5D3 is somehow either crippled or an unworthy upgrade to the 5D2. It produces a beautiful image.
bill
Click to expand...

You are the only person in the world that had BOTH OLP filters removed, James and Andrew only removed the first one. The second one is epoxied on the sensor and can only be removed professionally in a clean room.
 
marvinhello said:
You are the only person in the world that had BOTH OLP filters removed, James and Andrew only removed the first one. The second one is epoxied on the sensor and can only be removed professionally in a clean room.
Click to expand...
Interesting. You are suggesting the second OLPF is the magic element. I like it! Or, their footage mysteriously made it's way into
some post sharpening to just give it that little something extra. Which probably proves the majority viewpoint that post sharpening works just fine. I'll post a resolution test today but not having another 5D3 to A-B with it will be somewhat meaningless. I kind of like my slightly sharpened, distinctly pink camera. It's either half broke or half fixed. I definitely more of a half broke kind of guy.
Bill
 
3hats said:
Interesting. You are suggesting the second OLPF is the magic element. I like it! Or, their footage mysteriously made it's way into
some post sharpening to just give it that little something extra. Which probably proves the majority viewpoint that post sharpening works just fine. I'll post a resolution test today but not having another 5D3 to A-B with it will be somewhat meaningless. I kind of like my slightly sharpened, distinctly pink camera. It's either half broke or half fixed. I definitely more of a half broke kind of guy.
Bill
Click to expand...

haha, you removed both filters so there is no low pass going on at all. Pixel binning has an identical effect to low pass filtering so you probably won't see moire.
Have you shot any fabrics? like carpet, stripped shirt or any clothes that have complex patterns. It could be used to judge the impact of no OLPF on extreme detail handling.

For the pink cast, have you tried turning the WB shift a bit?
 
3hats said:
Interesting. You are suggesting the second OLPF is the magic element. I like it! Or, their footage mysteriously made it's way into
some post sharpening to just give it that little something extra. Which probably proves the majority viewpoint that post sharpening works just fine. I'll post a resolution test today but not having another 5D3 to A-B with it will be somewhat meaningless. I kind of like my slightly sharpened, distinctly pink camera. It's either half broke or half fixed. I definitely more of a half broke kind of guy.
Bill
Click to expand...

The first firmware update for 5D3 is out, have you tried updating it? is everything alright after update? any surprises?
 
I have posted a "half" resolution test, that is, a 5D3 with no OLPF but it's missing an unmodified 5D3 to compare with, sorry.
https://vimeo.com/40950236
Marvin(?)
thanks, for the update, I will check to see if the camera explodes.
So, you are saying the second filter is a low pass filter but without it (in my case) there is less moire? So, keeping it in the path would be a boost to resolution?
Just trying to understand.
thanks
 
3hats said:
I have posted a "half" resolution test, that is, a 5D3 with no OLPF but it's missing an unmodified 5D3 to compare with, sorry.
https://vimeo.com/40950236
Marvin(?)
thanks, for the update, I will check to see if the camera explodes.
So, you are saying the second filter is a low pass filter but without it (in my case) there is less moire? So, keeping it in the path would be a boost to resolution?
Just trying to understand.
thanks
Click to expand...

Thanks for the test video. What I meant was, the OLP filters are there to minimise moire as much as possible (primarily in stills mode), you removed them both so there may be risks of moire in stills mode (try take some pictures of super fine details like fabric, and view them 100%). In video mode, since Canon is using pixel bining, this downscaling method has identical effect of lowpass filtering hence there might not be moire, but there could be some in extreme conditions, that's why I suggested you to shoot some video of fabric to see if moire exists, and how serious it is.
 
marvinhello said:
haha, you removed both filters so there is no low pass going on at all. Pixel binning has an identical effect to low pass filtering so you probably won't see moire.
Have you shot any fabrics? like carpet, stripped shirt or any clothes that have complex patterns. It could be used to judge the impact of no OLPF on extreme detail handling.

For the pink cast, have you tried turning the WB shift a bit?
Click to expand...

This is not true.

If I had no OLPF and shot a scene where a red pattern was exactly not getting projected over the red pixels, you'd NEVER see it. Therefore, after the image is recorded, there is no way to know the red was even there because it never fell on the red pixels. Binning will not show you what was there.

If I had an OLPF, the red pattern would be duplicated in 1 or 2 directions depending on how many layers. The OLPF would be sufficient that the red pattern would actually make it to the Red pixels and would be captured by the sensor.

Their results will be different. The OLPF is designed to spread information out to they bayer pattern filter which doesn't color sample over 100% of the area. If you can't get that information, you can't recreate it.
 
Cosimo Bullo said:
"If someone wants better video then get a better video camera."

Or remove the filter as James and Andrew have done. The difference to these eyes is drastic. Some people just like pulling things apart, and I, for one, think this often leads to progress (and it's a lot of fun).
Click to expand...

At the cost of compromising the cameras performance in other areas with a questionable result when it comes to the resolution gain.
 
You must log in or register to reply here.
Back
Top