UPDATE: James Miller removes optical low-pass filter from 5D Mark III for resolution increase / new footage

[Andrew Reid]

[html]

Update: James has had a full day’s shooting now with the modified 5D Mark III. Check out the footage above.

Warning: please wait for this to all shake out. Don’t hastily modify your 5D Mark III without the necessary technical knowledge and research. Opening the camera voids the warranty and risks irreparable damage.

[url="http://twitter.com/#!/millerandmiller/status/186085781683703809/photo/1"]James Miller[/url] has decided to open up his brand new Canon 5D Mark III, tear it down and remove the blurring anti-aliasing filter from in front of the sensor. It does seem like a particularly strong optical low pass filter on this camera, which produces very soft results in video mode.

[url="http://www.eoshd.com/content/7727/james-miller-removes-optical-low-pass-filter-from-5d-mark-iii-for-resolution-increase/"]Read full article[/url]

[/html]

[Blastoff]

Interesting that the point of focus is different for both these shots, look at the glass of the lamp post in the top left hand corner, I'm sure that removing the low pass filter will improve detail greatly, but its harder to judge with these 2 images as the plane of focus of the picture with the filter removed does not meet with the row of buildings as in the 400% crop.

[cameraboy]

canon cameras have two aa filters ...
one blur y axis and second for x axis...
u can remove only one because second is glued with sensor ...
thats only half job but it does give more resolution ....
only problem is that u loose dust filter and u need additional uv/ir filter ...
gh2 is much better for that it has only one olpf and dust filter stay in the camera ....
second gh2 sensor is much more dense because smaller pixel and there is less possibility for moire ...
i saw some videos from modified gh2 and it beats c300 in resolution ...
but maxmax doesnt send cameras in my country .... :(
let see what d800e can do but  nikon is already got problems with aliasing  ....

[TJB]

Or for $200 less just grab yourself a Nikon 800E. I'm passing on the 5DMK3 and sticking with the MK2 and a newly acquired GH2. The 5DMK2 for people shots and the GH2 for deep depth of field landscape shots. Pity though that the Unified and Vanilla patches don't work so well in HBR 25p mode.
Anyone know of patches that are stable with HBR 25p?

[Andrew Reid]

James told me the top OLPF is easy-ish. But the 2nd is bundled in with the self-cleaning unit. My advice is wait for him to do a guide of some kind. Looks very promising.

[cameraboy]

yes ...
this is only for people who know what are they doing...
u can damage camera easy...
[url=http://www.lifepixel.com/tutorials/infrared-diy-tutorials/panasonic-lumix-gh2-ir]http://www.lifepixel.com/tutorials/infrared-diy-tutorials/panasonic-lumix-gh2-ir[/url]

[Alex Mand]

This is not for the faint-hearted...

Really open a new MKIII and take off the OLPF is a good challenge, also u need to replace with some sort of UV filter or something... Rocket science for me but resolution increase seems massive! But need more details on moire and other stuffs.

People are really trying to push foward MKIII res. Very good indeed.

[cameraboy]

canon change internal design with every new camera ...
old one like 350d had only one olpf ir uv filter ...
then 5d mark II 550d and 7D had two with one glued on sensor ...
5D mark III seems different design ...
we should wait  ...

[jcs]

Andrew- can you guys shoot a resolution chart? Curious how no OLPF and no post sharpening compares to OLPF and post sharpening (resolves about 890 lines (horizontal resolution) post sharpened).

Also, I see lots of ALL-I looking compression artifacts (macroblock quantization). Can you try IPB (my tests show stores more information, less noise, and less artifacts (+ ~3x less diskspace).

Finally, you can get much better post sharpening results by turning off in camera sharpen: this footage has halos.

Note that even with the compression artifacts and lower light/contrast on the OLPF image, it looks excellent post-sharpened compared to no OLPF (not clear if its worth removing vs. post sharpening. A line chart test would be helpful).

[Andrew Reid]

What have you done to that frame grab to make it look so bad?  ??? There aren't any halos on the original footage so just curious to what you did to introduce them... What is the point? There's even one on the text I added, that wasn't there before (Photoshop).

In my opinion the footage looks hands down better. Not sure if in-camera sharpening was on or off in this test, something to bear in mind too.

[cameraboy]

u can not sharper no aa image same as one with aa ....
use smaller radius value ...
no aa image has better pixel separation and too much sharpening produce halos ...
use less , now u got more REAL resolution...

[cameraboy]

looking at the image 5d mark iii looks like very close with c300 in resolution...
maybe c300 use only one aa ...
i told before that i think c300 sensor is just s35 crop of 5d mak III....
when u resize c300 to FF u got 22.3 it match in the pixel....

[Andrew Reid]

There's nothing similar what-so-ever between the 5D Mark III's sensor and C300. One is a 8MP sensor with custom bayer design for true 1080p and the other is a classic bayer 22MP for photography.

[jcs]

[quote author=Andrew Reid - EOSHD link=topic=501.msg3293#msg3293 date=1333225032]
What have you done to that frame grab to make it look so bad?[/quote]

You're right, that version doesn't look very good: tried to match brightness. Sharpening was in Photoshop: USM.

This version is much better, done in PPro CS5.5 using convolution sharpen and my GPU noise effect. To my eyes, the processed OLPF-present image looks pretty solid. There are regions which look better in the OLPF-removed image (front face brick areas, etc.), however I'm not sure those aren't related to using ALL-I (was James using ALL-I or IPB)?

[cameraboy]

custom Bayer is marketing  gimmick ...
its just pixel binning without debayering ....
same as af100 is different sensor from gh1 ...
c300 area is 2.8 smaller  than 5d mark III...
so 22.3/2.8=8 ...
they got same pixel pitch...
same tech ...

[brice]

[quote author=jcs link=topic=501.msg3298#msg3298 date=1333232048]
[quote author=Andrew Reid - EOSHD link=topic=501.msg3293#msg3293 date=1333225032]
What have you done to that frame grab to make it look so bad?[/quote]


This version is much better, done in PPro CS5.5 using convolution sharpen and my GPU noise effect.
[/quote]


Wait, what? Much better? Sorry, but this looks horrible.

[jcs]

[quote author=brice link=topic=501.msg3301#msg3301 date=1333234858]Wait, what? Much better? Sorry, but this looks horrible.[/quote]

That's cool. All of this is art. I like it (I oversharpened it to make a point: could have been matched to the other image for a more pleasing film look). What part(s) don't you like? The lighting and color between the images was different from the start. I processed it to show high frequencies present in the WITH OLPF dataset (image frame).

It's not clear yet if removing the OLPF will result in any more information being stored after processing and compression. A resolution chart would show with reasonable certainty if resolution has been improved past ~890 lines (measured from multiple sources with resolution charts).

[FastFisher]

I thought we all knew it wasn't the AA filter that makes the 5d3 footage soft.
It's the "Precise Canon Engineering" that makes it soft.
It's so obvious they made it blurry just right amount.
So somebody needs to hack it from the image processing side not messing with the AA filter.
Or just use other options available right now and save your time/money/effort.
Me as a Canon user this is rather sad of course.

[TC]

I am rather astonished by this.  Agree with FastFisher and this would be better tackled from the software (firmware) side.  The width of a physical pixel on the 5d III is 6.25 microns.  So the anti-aliasing filter will introduce a blur on this length scale to prevent aliasing in stills.  Probably a blur of around 3-5 microns.  The pixel pitch for HD video on a full frame sensor is 19 microns.  So the anti-aliasing filter should really not be having a big impact on the video image. 

However, I have just looked at James's video and the footage with the filter removed is some of the cleanest and sharpest HD video I have ever seen.  Amazing. 

[brice]

[quote author=jcs link=topic=501.msg3302#msg3302 date=1333235769]
[quote author=brice link=topic=501.msg3301#msg3301 date=1333234858]Wait, what? Much better? Sorry, but this looks horrible.[/quote]

That's cool. All of this is art. I like it (I oversharpened it to make a point: could have been matched to the other image for a more pleasing film look). What part(s) don't you like? The lighting and color between the images was different from the start. I processed it to show high frequencies present in the WITH OLPF dataset (image frame).

It's not clear yet if removing the OLPF will result in any more information being stored after processing and compression. A resolution chart would show with reasonable certainty if resolution has been improved past ~890 lines (measured from multiple sources with resolution charts).
[/quote]

Sorry for being harsh, just not my cup of tea. But I am wondering what all was done to the screenshot? Given that the text was embedded in the file you worked with, I am wondering why the text looks like it has a slight Gaussian Blur applied?