Aliasing
 

While this has been discussed in different threads aliasing is still the number one issue for me. I don’t own the 5DMK2 so my conclusions relay on posted material. First I thought it is only a problem of wrong downscaling for the web (vimeo, flash, QT). But then I started to look for aliasing in raw footage and it is already there. Also it can be found in various footage, so I don’t think everyone is doing something wrong while downscaling. I’m engineer my self and understand the problem Canon was facing with the 5DMkII. They had to design the OLPF (optical low pass filter aka anti aliasing filter) for the 21 Mega pixel resolution of the Sensor rather than for the 2MP used for HDTV. For properly sampled video they had 4 options (IMO):

1. Read out the whole 21 mega pixel 30 times per second (630MP/S) and downscale it to 1080p30 using e.g. bilinear filtering. Even if the CMOS sensor could be read out that fast (I highly doubt that) the resulting data stream would most probably be more than the 5DMKII chipset can handle.

2. Read out only the center of the CMOS sensor (aka region of interest or windowing). While this is very easy to implement it has a lot of drawbacks. First of all it introduced a crop factor like on APSC Cameras. But in this case it will multiply the focal length of your lens not only by 1.5 as on APSC sensors but by 10 (yes then). While this is great news for wild life shooters, it is absolutely not usable for anything else but filming a close up of the moon with your 70mm lens.
Compared to the full readout the sensitivity will drop by at least 3 stops.
Also you will loose much of the shallow DOF. It would be slightly better than a 2/3” sensor.

3. There are CMOS sensors which can combine adjacent pixels (aka binning) and read it as one single large pixel. In the case of the 21MP sensor of the 5dmk2 a 3x3 pixel binning would reduce the amount of data by a factor of 9 to a moderate 70 MP/S. This data rate could be easily handled by the cameras chipset. Unfortunately the 5dMk2 sensor seems not to support binning.

4. Most CMOS sensors can skip rows and pixels to reduce the resolution. This is normally used for preview functions where not the whole resolution is needed. It seams the Canon engineers use this method to reduce the amount of data for the Video mode. The problem is; when skipping e.g. every second pixel the OLPF should be twice as strong (it now should blur an infinite small point in the image to a disc of the diameter equal to the distance of 3 sensor pixels). While such an OLPF would produce clean video the still resolution of the sensor would drop from 21MP to 5MP (not a good thing for a professional DSLR). So Canon decided to go for sharp 21MP stills over properly anti aliased video.

Until either the chipset and imaging sensors are fast enough to handle 30 full sensor readouts per second or the sensor supports pixel binning there will always be a tradeoff between still resolution and proper anti aliased video.

Once captured it is not possible to completely remove the aliasing in post. Or at least I do not see how this should work. The only solution I can imagine is to add some sort of optical diffusion filter.

This is good news Josh. Can it be that simple? This should be investigated more deeply.

Solution 2. is what RED is doing when shooting 2k, quite annoying as you point out.

If you're right (Canon using Solution 4.) that's disappointing (but that's what I guess from user reports too :/) since we could have 10 times better image quality (and noise-free video @6400 ISO) if bilinear downsampling was used...

Do you have some link to raw footage showing aliasing artifacts ?

Have a look to Video Samples 2 from Digital Photography Review
The shop door in the background shows some aliasing. Yes I know its shot with a preproduction model but I don’t think there has been any significant improvement to the production model.

Dan Chungs Uganda video on Vimeo shows aliasing at 2:30 on the woman’s shirt. Dan can you confirm this aliasing is also in the raw data?
P.S. if I was told to give an interview in front of a 5DMk2 I’d wear a nice black and white checked suit :-)

Red doesn't alternate lines, so it's not as much of a problem for them. We don't know if Canon is skipping lines to reduce data or the read/reset isn't fast enough.

I'm not sure if they are just row skipping (1/3rd the pixels) or row+col skipping (1/9th?). The effective "fill factor" must be very low, perhaps less than 10%.

It's hard to decide what I dislike the most about my new 5d2. Aliasing is up there very high, but I have to balance that against all the other failings:

• Aliasing, moire, stair-stepping, jaggies, etc.
• Lacks even the most rudimentary manual controls
• Sets shutter speed based on the 1/focal length rule
• 30p
• Compression artifacts
• Heavy and ugly noise reduction even in "NR: off" mode, poor demosiac.
• Stops at 4GB instead of writing a new file.
• No audio meter, control, or headphone monitor.
• No live HDMI

It seems as if Canon got just about everything wrong; the 5d2 does very poorly in just about every possible metric for measuring the quality and utility of a video camera, except one: sensor size.

Unfortunately for me, this means I will have to continue lugging around my XH-A1 in addition to the 5d2 until Canon releases a free firmware upgrade (it is to laugh).

I just got Tiffen Soft #1 and Diffusion #3 filters. On a 50mm f/1.8 Nikon, the Diffusion filter definitely removes all aliasing. The Soft filter seems too weak. If I had to guess right now, a Soft #2 would be the target.

I need to do more testing before I'm certain about that though...

Thanks, Jon; I appreciate your testing. I will pick up a Soft #2.

That sounds very interesting, can you put some footage up?
But Tiffen also offers Diffusion #2 and #1 - why would you assume soft #2 would be strong enough?

I wonder... there's already a filter to deal with the d90's aliasing in post - from the plugin developer (Too Much Too Soon Free Plugins for Final Cut Pro and Final Cut Express):

"The D90 rescaler plugin works because the D90 really captures an 800p image and scales it to 720p, and the way it's done makes it possible to scale it back up almost losslessly to 800p and then back down using a better algorithm."

maybe something similar can be worked out for the 5D? If we can figure out how many sensor lines are being skipped it should give us a target resolution to try the same process.

EDIT: hmm, maybe not - taking a 16x9 crop of the full resolution (5616 x 3744 pixels) gives us 5616x3159. Sampling every 3rd horizontal line would give us 1053 so maybe they're not doing any additional downsampling which could be re-done in post as with the d90.

You're right that it's an assumption right now. I haven't had a chance to test with it.

From what I've seen the Soft filters give an overall blur, while the Diffusion filters bounce the highlights around for more of a highlight blur effect. How they do this, I have no idea. Somehow they keep clean details in the blacks, but smear only the brightest aspects of the picture. That makes me surmise that the Soft filter is the better choice for anti-aliasing. But I don't, in fact, know that a #2 is strong enough. But from my quick tests, we can scratch the #1 off the list.

Thanks for that interesting information!
The downside is, that these filters simply are not available in Europe - so I´d have to buy them in USA. I would really have to be sure to buy the right one - so please keep us informed if you find any information or even RAW-Video using one of these filters.

I will try to get my hands on one of these german filters:
-Zeiss Softar
-Heliopan DUTO
-B&W Soft Pro
Unfortunately they are just available in 2 versions.
As B&W is available in USA too, if anyone got experience with them, please post! :-)

Unfortunately, I won't be able to do any tests for a few weeks. My son took the camera with him as he returns to college, where he will shoot the pilot of ~ The Murder of Dirk Snowglobe - A Not Dead Detective Series ~.

How do you arrive to this value?

This is really simple. In my example I was skipping every second pixel (horizontal and vertical). When sampling only one out of four pixels the resolution will drop to one-fourths. This equals to 21MP/4 => 5.25MP ≈ 5MP.

If an image is projected to the sensor which has details small enough to fall between two active pixels, aliasing will occur. The function of the OLPF is to spread such small details so it always hits an active pixel.

See Nyquist Shannon sampling theorem - Wikipedia for details about sampling.

To be clear I don’t say the 5DMKII skips every second pixel. This was just an example. But it is save to say the Canon 5DMKII does some sort of sub sampling and it is also obvious that the OLPF is too weak in this situation.

Yeah but it´s pretty obvious that the 5D2 uses every third pixel of every third row!
That pretty much fits the pixel count 1920x3 = 5760.

Using every second pixel of every second row would bring us only blue and red (or just green) pixels of the bayer patterns!
Sampling every third pixel in every third row brings us another perfect bayer-pattern again!

From the images posted at Reduser, I think they use every pixel on every third row. There doesn't seem to be aliasing of vertical lines. It's horizontal lines that cause the problem.

This could be implemented with a simple low pass filter before the A/D as the lines are read, providing inexpensive horizontal filtering with a potentially ideal aliasing/resolution result in that dimension. In the vertical dimension, it's another story. By skipping every third row, we get aliasing issues.

Oh well. This isn't an issue if we shoot sunrises. Zooming in on herringbone suits could be another story...

I can´t believe that they would be reading out 7MP 30times per second - that would be a very high pixel rate.
And it would be a pity - because if they were able to read 3 out of 9 pixels in 30 fps it was too close to 4 of 9 pixels which would provide us a totally new possibility:
http://www.gamsig.com/geheim/Bayer2.gif

I think the solution is to find the right soft/diffusion filter that reduces aliasing without reducing the resolution by too much. Frankly, a solid 720p image is probably enough for most of us. Vimeo is 720p. Unless we're making Blu-ray discs or showing at a film festival, I'd think most of us would be happy with aliasing-free 720p. (That and 24 fps and Exposure Simulation that actually sticks in video mode...)

Yes Jon adding a low pass filter before ADC will help reducing horizontal aliasing. But if canon uses a usual Bayer pattern this method will average blue and green or green and red pixels. I’m not sure if this can be corrected by a clever de-Bayer algorithm. May be Canon managed to only average pixels form the same color. If not, I think the color separation will be a problem.

As for read out speed: In burst mode the 1DmkIII can capture 5 images per second. If the 5DmkII has the same sensor, we should be safe to say that the read speed is at least 105MP/s.
If the mechanical shutter or the DSP / CPU are the limiting factors, then the sensor may be read even faster.

Reading every pixel at every third line would result in 210MP/s which is very fast. At the moment I don’t really care how canon exactly reads the pixels. It is more important to know how we can work around this issue. I think some sort of optical filter will be the only clean solution. I don’t have access to a 5Dmk2 so I have to wait for test results.

Exactly what I think. This will be - in the short term - the way to go. A filter might even add some organic bokeh, similar to 35mm adapters. Not everybody's choice but some will like it.

I hope we will find out about the best filter ASAP - and one thing is positive:
Even with the sharpness parameter of the 5D2 all down, filterless video looks tack-sharp.
So with the right soft-filter in the "sweet-spot" between aliasing and softness, the in-camera sharpening should be able to compensate for the filter.

We could also sharpen in post, but considering the AVC-encoder of the 5D2 the results might even be better with moderate in camera-presharpening.

a cheap solution for the antialiasing problem could be black Tulle. like this:
Tulle Netting Black Wedding Favor Netting Spools 6" wide 75' / 25 yards for $2.88
just put one or two layers of black Tulle in front of the lens.

professional people photographers use this to get soft portraits...

Yes this could be the solution. Very smart.

My personal moire showcase ;-)
 

To keep the search for the best workaround alive,
here´s my personal showcase for "5D2 moire":

ALIASING on the Canon 5D mark II on Vimeo
Aliasing is way uglier in the original 1080p-file that´s downloadable at the bottom of the page.

While shooting portrait-style with wide-open bokeh nearly 98% of the footage is decent.
As I took a winter walk, only 4 of 20 scenes looked ok.
If you got roofs or cars in the frame and you´re doing a wideangle shot, chances for aliasing are very high.

Andreas: If you shoot in urban environments you will almost always have hard edge ore fine repeating detail where the 5DMk2 will fail. May be I am hypersensitive for image artifacts but what can I do. I can’t stand the rainbow effect of DLP Beamers. I do not like the sharpening artifacts of cheap digital cameras. I’m less than unhappy with rolling shutter artifacts of my HV20. I do not like the motion artifacts of 100Hz / 120Hz TV sets……
All of this technology fails in one ore the other situation. Once spotted it attracts my eye all the time.

I was curious about the Tulle method Ralph mentioned. But then I got nervous about affecting the bokeh. So I did this quick test with my HV20 and a tee sieve. This method fails unless you use a very fine mesh.

-Street light in focus no filter
-Street lights out of focus no filter
-Street lights out of focus Tee sieve right before lens (4mm)

Has someone jet tested the diffusion filters? Footage?

This thread is too important to let it sink into oblivion. There's still no satisfying answer of how to reduce the sometimes annoying aliasing of the 5D. (Not only here on dvinfo but on other sites, too) Besides the struggle for better manual controls and 30p to 24p conversion, this is the 3rd most worrysome issue with the 5D. I would love to know that there's a solution before buying the 5D next month.

A very crude method to use in post... just apply a low amount of blur onto your footage in your NLE.... untill the aliasing reduces to an acceptable level.... then sharpen the footage, if required.

This doesn't work for every shot, but it can do wonders for some footage... and once down to 720p, the sharpness is still very good, without anywhere near the same amount of aliasing.

Worth keeping in mind untill a proper solution is found.

compressor, with its amazing array of frame controls, offers anti-aliasing as an option. While it doest totally deal with aliasing issues with various cameras, it does make problems much less noticeable.

frame controls: ON

Anti-aliasing: Full strength (100) for the 5dm2

Filtering in post can remove stair steps on slight diagonals, such as car roofs, but it won't fix aliasing on repeating patters, such as fine patterned shirts, herringbones, and tiled roofs.

Any time that fine patterns cause larger patches of light and dark, your sunk.

By definition, aliasing is the mistranslation of high frequencies into lower frequencies - sometimes VERY low. DC low. That kind of aliasing is there forever. You need an optical filter and (optionally) oversampling to remove aliased patches.

Down this thread it was mentioned that a Tiffen soft #3 would completely quit the aliasing but was too strong to maintain image shaprness. The same poster claimed that al Tiffen soft # 1 was too weak. His guess was, that the #2 would be the perfect solution.

I wonder if anybody tried that one or a similar filter with good results.

If we find a filter that does the job we have resolved a big problem with few effort. I personally could life even with an image downsampled to 720p.

Conclusions
 

So far there are 3 solutions
1. Tiffen Soft #2 filter in front of lens
2. Tulle ultra fine mesh
3. Keep the object slightly out of focus so the lack of focus compensates for the aliasing

2. does not seem to work consistently..i.e. It gives an effect but it shows the mesh on closeup, especially over lights.

So far, no one has confirmed Soft #2 filter.

Does anyone wish to add something ?

I would personally prefer moire to any of these sharpness reducing solutions. Of course we would have to monitor for moire on every shot and re-setup shots with really bad moire, but that's the way life is.

For example, if a building in the background has a pattern of bricks showing bad moire, then we change DOF, we change focal length, or switch locations until it looks better.

The is no way Canon can fix the moire in the 5D2 without a serious hw fix, and that's that.

When you mean changing DOF and focal length, you are simply creating softness in the image. Isn't that the same as the out-of-focus option mentioned above ?

There are two ways to fix the moire - a lower frequency optical filter or a digital filter that reads all of the samples.

The digital filter approach would be awesome, but it would have to deal with three times the data rate coming off the sensor. Maybe we'll see this in a $10k+ video camera. That would compete with the FF35 Scarlet.

The other approach would be an optical low pass filter with 1/3 the resolution of the current filter. A well matched screw-on filter would do the same. The problem is that we'd get a 720p (or so) result, rather than 1080p. Red faces this same problem, which is why their lowest resolution is 3K. A 3K Scarlet should be able to give a true 1080p result without aliasing. That's one of the beauties of RAW. It moves the demosaicing and the digital low pass filter from the camera to the PC.

Still optimistic that a simple filter will be the workaround for now - just hope that the whole footage then doesn't look like softporn...

Or like a closeup of a 1940s starlet. "All right, Mr. DeMille, I'm ready for my close-up."

That's why we should look at soft filters, rather than, diffusers. I don't know how they do it, but diffusers seem to smear bright light farther than soft. A true optical low pass filter will limit the spatial frequencies, regardless of brightness.