Upcoming hands-on comparison of F3, FS100 and AF100 from Philip Bloom

[David C. Williams]

Originally Posted by Mark David Williams

I measured the lense diameters a while back, can't remember the exact figures now and your right the 9.5mm would have vignetting and would have to reframe on that lens However the 25mm and 16mm, will be okay the 12mm may need a little reframing. Not good when the starting resolution is only 680 lines.

Measuring the lens will do you no good. You need to measure the projected image diameter set at the proper depth of focus. The light does not exit parallel, it converges to the focal point.

[Mark David Williams]

Zone plates show both the F3 and FS100 to have the same vertical resolution, but the FS100 having more horizontal aliasing. Out of band moire is very similar for both.

Alister
What do you measure the resolution at?

Mark

[Alister Chapman]

FS100 MTF50 was about 730 LW/PH. The results are earlier in the thread somewhere.

Here are the MTF 50 plots for the Letus on EX1 based on Marks frame grabs.

Vertical = measured 433 LW/PH, corrected 639 LW/PH
Horizontal = measured 503 LW/PH corrected 675 LW/PH

Well short of the 925 LW/PH of the EX without the adapter.

[Mark David Williams]

Hi Alister

I don't think you should use MY framegrabs to measure resolution as I downloaded it from the internet and printed it out. The idea was to give the figures for the the EX1 chart and for the EX1 with letus and that way you can judge roughly the difference between them.

Could you measure the resolution on my frame grab for the EX1 only as well? Or have you done this?

Mark

[Mark David Williams]

Alister

Did you test both pictures thinking they were both taken with the letus? Looking at the frame grabs it appears as if you have although I'm unfamiliar with this process.

Mark

[Alister Chapman]

The great thing about the Imatest process is that it does not use the trumpets or other traditional resolution measuring parts of the chart. All it needs a dark to light diagonal line. A very poorly printed chart will degrade the results, but most inkjet or laser printed charts will give a reasonably accurate result.

One plot is horizontal the other vertical. Both done from the letus chart.

[Mark David Williams]

Thanks Alister

I think it would only be fair to do the same test on the EX1 picture as we dont know what resolution is lost through my unscientific frame grabs.

Mark

[Alister Chapman]

I'll try and run the bare EX charts over the weekend.

[Brian Drysdale]

Originally Posted by Alister Chapman

FS100 MTF50 was about 730 LW/PH. The results are earlier in the thread somewhere.

Would a higher quality lens make much difference this figure?

[Alister Chapman]

I don't think so. The same lens gives well over 850 LW/PH on the F3. I've tried various lenses and get the similar numbers on the F3 and FS100 including Zeiss Ultras on the F3.

[Chris Barcellos]

Originally Posted by Mark David Williams

Here is a link to a trailer I made using the EX1 and Letus
Vampire Gang Origins (2011) - IMDb

Mark:

Re your earlier question about reliability, my earlier experiences with the Letus are based on a borrowed unit that was supposed to have been little used, and my own first original Letus35 (no flip) in which I had motor wire break off after a short time. I put up with it because it was the only game in town, and it was a pain.

Your trailer post above would be more helpful if it was shown in higher definition. But I think it still exhibits typical Letus softness. I am not saying it is objectionable, even -- it actually looks great in a lot of ways.
And that is what makes this debate about lines of resolution a bit ridiculous. What I care about is how my audience receives my work, not how many lines I get up on the screen. That is why the 5D, even with its lower end numbers and other disabilities is still playing in the game.

[Steve Mullen]

David, the FS100 is most likely a prosumer NEX. The consumer NEX VG10 is a 1920x1080 camera using a 14MP chip.

I really doubt there would be any reason to even think the FS100 was using the F3 chip to get a 1280x720 resolution. There would no logical reason to do this and then upscale to 1920x1080.

Plus, both the F3 and FS100 are 800ISO cameras. So, the "magic" that is used to get the extra sensitivity is being used by both cameras -- even more reason to believe the sensor (which we know to be the same) and the DSP operate in an identical manner.

However, one option never mentioned is that although the same DSP may used in the F3 and FS100, the power differences came from the clock-rate the chip is run at. To process 4:4:4 10-bit data means the chip needs to run much faster than when processing 4:2:2 8-bit data. You need to understand that the de-bayer -- done on 3.7Mpixels, can directly create YUV at either 4:2:2 8-bit (for HDMI output) or 4:4:4 10-bit (for HD-SDIi output). For 4:2:0 encoding, both 4:2:2 and 4:4:4 are decimated to 4:2:0.

[Mark David Williams]

Hi Chris


Some pics attached for you to have a look at.

Steve don't forget the FS100 is using 8bit signal processsing so that's going to cut down power consumption too.

I can't see any softness in the Letus only that the edge sharpening is TURNED off on the EX1?

[Alister Chapman]

I though Steve that you were suggesting that the "magic" was happening at the sensor level with pixels groups being read together off the chip as opposed to singly to overcome all the normal readout speed and heat issues associated with trying to read every pixel with DSLR type pixel counts? In which case the sensitivity would not change significantly if the debayer was less sophisticated.

If the sensor is read the way David is suggesting then the sensitivity would be the very similar to the F3.

I'm not convinced by either theory. I think the differences are just down to a less sophisticated DSP in the FS100.

[Alister Chapman]

I'd suggest you might want to reduce the detail correction levels or raise the frequency your using Mark to get rid of the thick black line around the guy's nose and the girls dress. I don't like the smooth lack of texture on the guys skin (the girl is better), nor the blown out skin tones. The images lack any real contrast only crushed blacks and blown whites, but very little in between. The girls hair is just a solid black mass on the right side, no texture or detail of any sort at all, makes it look like crushed video. Sorry, not to my liking, but that's just my opinion and others will of course differ.