Report: F3 looks almost the same a EX3

[Kenn Christenson]

Coming soon - "Test Charts - the Movie" LOL! Was just expecting a little more resolution from a camera which is roughly 3x the price of my current camera - don't think that's too much to ask for.

[Jacques Mersereau]

The proof is in the pudding. Test charts are great for getting specs, but IMHO, what does the image look like projected 20 ft tall and 35 ft wide?
I haven't seen the F3 this way, but the EX3 looks pretty darn good for what it costs.
The vimeo F3 content also looks very good despite the compression.
Right now, the F3 is a real contender in its class.
RED may answer and deliver a camera at this price point that is even better, but we've all been waiting for a long time (no offense intended).
RED Scarlet vs SONY F3 is much like DAVID vs GOLIATH without David at this point.
It would be great to see what happens in April.
SI is also out there working away I am sure.

[Kenn Christenson]

I agree. The EX series cameras are excellent - in their class. Don't think much out there beats them for resolution.

Thing is - I've got clients who can see the difference in resolution between what I used to shoot (HVX200 & HPX500) and what I shoot with, now (EX1.) I'd hate to spend all that money - to go backward in what I consider a major component to any image.

[Steve Kalle]

Originally Posted by Kenn Christenson

Any resolution tests/comparisons between the F3/EX1 and maybe the AF100? I've heard both the F3 and the AF100 top out at about 800 lines/vertical. Been considering the F3 as a replacement for my EX1- until I heard about the lower resolution. Is this something Sony can firmware fix? (my guess would be no - but have to ask, because I like this camera, otherwise.)

Yes, Alister did some tests and yes, the F3 is a bit lower than an EX1/3. HOWEVER, the amount of noise in the EX1/3 is exponentially greater than the F3, which is why the F3 is the only camera on my list to buy. Look at it this way - use a noise removal program (Neat Video) on the EX1 video and you will end up with reduced resolution and less noise.

[Jacques Mersereau]

Agreed Kenn. I think there are a lot of buyers out there who, with their next camera purchase, want to move beyond the 'acceptable to very good' stratosphere and into the orbit of REAL PRO.
THE RED ONE has made it into ORBIT. Yes, yes yes, most of us know all the pluses and minuses,
yet the FACT is REAL MOVIE MAKERS are winning OSCARS with it. RED ONE IS THERE.
So the question is who will deliver similar capability/hardware for a lower price point?

Did I forget to mention Canon? Well, talk about waiting. We've been waiting for Canon for years and years.
The 5D doesn't count. That was, for Canon corporate HQ, a mistake imho - but ONE THEY SHOULD NOW EMBRACE.

[Alister Chapman]

I got about 950H 850V resolution at MTF50. That's not a bad set of numbers for a 1080 camera. My charts and test procedure are not perfect, possibly with a better chart you would see marginally higher numbers. What I can say though is that visually on a monitor it appears at least as sharp as my EX's, if not sharper. This appears to be because there is no noise to blur or soften the image.

[Steve Kalle]

Resolution isn't the most important factor; so, try not to get too hung up on it. Remember that many people have been praising the 5D & 7D; yet, they can resolve barely more than an SD camera (500ish TVL). Btw, I am certainly NOT one of those praising those 2 cameras.

PS I'm on Alister's side about peoples' infatuation with Red's 4k/5k as 'resolution' when 4k only means the number of pixels. And yet, even Red misleads people by stating '5k resolution' on their website.

[Peter Moretti]

Originally Posted by Alister Chapman

I got about 950H 850V resolution at MTF50. That's not a bad set of numbers for a 1080 camera. My charts and test procedure are not perfect, possibly with a better chart you would see marginally higher numbers. What I can say though is that visually on a monitor it appears at least as sharp as my EX's, if not sharper. This appears to be because there is no noise to blur or soften the image.

Alister, I may be wrong, but I thought all cameras have a Nyquist band limiting filter just before the output going to the A/D converter. I believe that menas an absolute limit of about 1K tv lines for 1080 cameras. But that's not at MTF50. So the MTF50 number is going to be lower.

If that's correct, I believe getting significantly higher rez than an EX from any 1080P camera probably ain't gonna happen--please excuse my American, LOL.

[Alister Chapman]

When measuring the resolution of a well designed video camera, you never want to see resolution that is significantly higher than HALF of the sensors resolution. Why is this? Why don't I get 1920 x1080 resolution from an EX1, which we know has 1920 x1080 pixels, why is the measured resolution often around half what you would expect?
There should be an optical low pass filter in front of the sensor in a well designed video camera that prevents anything above approx half of the sensors native resolution getting to the sensor. This filter will not have an instantaneous cut off, instead attenuating fine detail at ever increasing amounts centered somewhere around the Nyquist limit for the sensor. The Nyquist limit is normally half of the pixel count with a 3 chip camera or somewhat less than this for a bayer sensor. As a result measured resolution gradually tails off somewhere around Nyquist or half of the expected sensor resolution, but why is this?
It is theoretically possible for a sensor to resolve an image at it's full pixel resolution. If you could line up the black and white lines on a test chart perfectly with the pixels on a 1920 x 1080 sensor then you could resolve 1920 x 1080 lines. But what happens when those lines no longer line up absolutely perfectly with the pixels? lets imagine that each line is offset by exactly half a pixel, what would you see? Well each pixel would see half of the black line and half white line. So each pixel would see 50% white, 50% black and the output from that pixel would be mid grey. With the adjacent pixels all seeing the same thing they would all output mid grey. So by panning the image by half a pixel, instead of now seeing 1920x1080 black and white lines all we see is a totally grey frame. As you continued to shift the chart relative to the pixels, say by panning across it, it would flicker between pin sharp lines and grey. If the camera was not perfectly aligned with the chart some of the image would appear grey or different shades of grey depending on the exact pixel to chart alignment while other parts may show distinct black and white lines. This is aliasing and it's not nice to look at and can in effect reduce the resolution of the final image to zero. So to counter this you deliberately reduce the system resolution (lens + sensor) to half the pixel resolution so that it is impossible for any one pixel to only see one object. By blurring the image across two pixels you ensure that aliasing wont occur. It should also be noted that the same thing can happen with a display or monitor, so trying to show a 1920x1080 image on a 1920x1080 monitor can have the same effect.
When I did my recent F3 resolution tests I used a term called the MTF or modulation transfer function, which is a measure of the contrast between adjacent pixels, so MTF 50 is where there is a 50% of maximum contrast difference between the black and white lines on the test chart.
When visually observing a resolution chart you can see where the lines on the chart can no longer be distinguished from one another, this is the resolution vanishing point and is typically somewhere around MTF15 to MTF5, ie. the contrast between the black and white lines becomes so low that you can no longer distinguish one from the other. But the problem with this is that as you are looking for the point where you can no longer see any difference, you are attempting to measure the invisible so it is prone to gross inaccuracies. In addition the contrast at MTF10 or the vanishing point between black and white will be very, very low, so in a real world image you would often struggle to ever see fine detail at MTF10 unless it was strong black and white edges.
So for resolution tests a more consistent result can be obtained by measuring the point at which the contrast between the black and white lines on the chart reduces to 50% of maximum, or MTF50 (as resolution decreases so too does contrast). So while MTF50 does not determine the ultimate resolution of the system, it gives a very reliable performance indicator that is repeatable and consistent from test to test. What it will tell you is how sharp one camera will appear to be compared to the next.
As Nyquist is half the pixel resolution of the system, for a 1920 sensor anything over 960 LP/ph will potentially aliase, so we don't want resolution above this. You don't want to see a higher number than this as it has the potential for problems as the extinction resolution must be higher than this and thus there must be the risk of aliasing. This where seeing the MTF curve helps, as it's important to see how quickly the resolution is attenuated past MTF50.
With Bayer pattern sensors it's even more problematic due to the reduced pixel count for the R and B samples compared to G.
The resolution of the EX1 and F3 is excellent for a 1080 camera, cameras that boast higher than 960 LP/ph will have aliasing issues, indeed the EX1/EX3 can aliase in some situations as does the F3. These cameras are right at the limits of what will allow for a good, sharp image at 1920x1080.

[Peter Moretti]

Agree with everything you've said. But I believe that in addition to the OLPF there is a hard cutoff filter just before the A/D converter. This prevents sending a signal that could alias in the digital signal processing b/c OLPF's aren't perfect, and in the case of a Bayer sensor, the OLPF will have to allow some red and blue aliasing.

A result of this hard cut filter is an absolute limit to what total resolution can be.


P.S. Well I should say I agree with the tenor of everything you're saying. I believe you're not using LP/ph properly. If you mean line pairs per pixel height, Nyquist is 1080/2 = 540. But your explanation of the role of the OLPF jibes with my understanding, FWTW.

[Alister Chapman]

MTF plots for the camera don't show a hard cut off.

[Peter Moretti]

Alister,

I could try to explain this away as dithering, interpolation, sharpening, codec distortion, electronics' noise, setting the A/D frequency higher than twice the optical Nyquist frequency of the sensor, the fact that the sensor's Nyquist rate is theoretical in that it assumes perfect fill, but I'm just not sure.

That said, I'm almost positive that the A/D includes what could be thought of as an electronic LPF, otherwise you'd have digital aliasing that will have no picture value whatsoever. The net effect of this "ELPF" is that there is an exact governor in the system for resolution coming off the sensor.

[Glen Vandermolen]

So, a camera that shoots as good as an EX3, but has a cleaner image, less noise, better control over depth of field, a much greater variety of lens selection and is better in low light?

Ding, ding, ding! We have a winner!!!

[David Heath]

Originally Posted by Peter Moretti

I believe that in addition to the OLPF there is a hard cutoff filter just before the A/D converter. This prevents sending a signal that could alias in the digital signal processing b/c OLPF's aren't perfect, .........

I don't think so - by the time the image hits the sensor it's too late, it's the image on the chip and the sensor pattern that burn the aliases in. Put into a system a higher frequency than the Nyquist limit, and the output will wrap round the Nyquist frequency.

To take Alisters example further, then think of putting in a signal that's even finer than the sensor pattern. Let's say 960x1.5 line pairs across the horizontal (so 1440 white lines and 1440 black). Move across horizontally, and pixel by pixel what you'll see will be :

Pixel 1 - first 66% white
Pixel 2 - last 66% white
Pixel 3 - last 33% white
Pixel 4 - first 33% white
Pixel 5 - first 66% white etc etc

Which is a light grey line every two pixels, followed by a dark grey line for the next two. Across the sensor width, you'll see 480 line pairs.

A frequency 1.5x the Nyquist limit has been turned into an alias 0.5x the Nyquist limit. And no filtering etc that comes after the sensor can distinguish between this alias and a genuine pattern of 480 line pairs.

[Peter Moretti]

David,

I believe I understand what you're saying and it does make sense. Essentially the sensor's photosite grid itself acts as a "hard filter," so resolution above Nyquist gets broken down into a sub-Nyquist image.

I appreciate you and Alister's input, but I have to dig into this whole issue of a digital filter more, b/c I could swear there is more than one Nyquist filter in a camera design.

BTW, how do you explain Imatest numbers above Nyquist (which I realize in no way proves the existence of an "ELPF")?

http://www.dvinfo.net/forum/sony-xdc...g-moire-2.html

Thanks.