We made a short sharpness test between EX1 and EX1R . The EX 1 is sharper(noticeable in the corners) than EX1r but a bit darker. Here are the test pictures made in same conditions both camera with same settings: 1/50, f1,9, 0db, PP off

But the "why" is critical.
Could it be the defaults have changed? I doubt they made hardware changes that would make the image softer or more light sensitive.

Another question is would the firmware upgrade on the EX-1 to 1.20 impact such defaults. What firmware was the EX-1 at?

That's pretty extreme, IMHO. Were both cameras recently back-focused on an identical object from the same distance? Just guessing it might be a focus issue?

Only in the conditions were the same , the EX1 test is from a few months ago. The firmware on EX1 was 1.1 . We chosen the most obvious difference the upper left corner, the difference on the other corners is not so striking. We made also a test with continuous focus movement but we have not analyzed the result yet.

I find it extremly difficult to reproduce the same conditions between two cameras. Today, at the evening before I will send my EX1 (FW 1.11) to the new owner (I'll keep the EX1R FW 1.00) after reading this thread, I tried to do a comparison too.

Its even difficult to find the same Z-stop. Z58 is not always exactly the same. You can get closer to Z57 or to Z59. Although I tried several times you see that I was not able to find exactly the same size of the picture.
I had to use Z58-stop because being closer I wasn't sure if I could find the best focus. So I was around 1m away from the test image, normal yellowish indoor-light, filming in 720/50p, PP, shutter, steady-shot, WA-conversion all off, -3db, 3200K.

I did several tests, always trying to find the best focus with peaking, exporting a frame with ClipBrowser. Afterwards I tried to find the best and sharpest still from the EX1, then the best from the EX1R.

Although this is not scientific and probably I forgot something (I've resetted both cameras before tests) I would conclude it like this:

the EX1 is slightly better and I beat myself paying 1800 Euro for changing the EX1 to the EX1R. Its not that bad as I thought after the images by Zoltan, but in all tests the EX1 was somehow sharper - and brighter. The EX1R has a yellowish tint and together with the slightly better sharpness the EX1 looks clearer and better, in all my tests. The lower right corner of my EX1R seems to be worst, maybe because of the not perfect lighting in this low light scenario.

It still may be a matter of Sony simply changing the defaults on the EX1R camera.

It would be interesting to see how the EX1 behaved after upgrading the firmware to 1.20.

It may also be a result of something that improved the IR contamination issue.

You'd really need to test the EX1 and EX1R with both having the new current firmware.
Then you would need to see if modifying a PP of EX1 to match the default of EX1R.

I wouldn't say the above means the EX1 is better in any respect as it may be that Sony just changes the default settings.

This would mean that Picture Profiles made specifically for the EX1 won't precisely work with the EX1R and viceversa, right? If this is the case, then it should be noted!

I guess "PP off" means out of the box PP, right? Maybe you could look through the PP menu and see if there are any differences. If there are no differences... well, that's just plain trouble for picture profile's compatibility.

Odd , the first person to post about EX1R here said it was a bit faster than the EX-1.

It does stand to reason that there may be some filtration added to mitigate the IR problem and that could affect color & speed.

The extra IR filtration on the EX3 compared to an EX1 makes a very slight color difference to the pictures from the two cameras and the EX1R is slightly different again. You can't cut down on the red light reaching the sensor without some small effect on the final image.

Was the sharpness setting the same on both cameras?

In my case: none, PP off.
BTW - I did another test with the EX1R - result is the same but the "16:9" in the lower right corner is little sharper, so I am not afraid anymore that this corner of my EX1R is worse than the others.

Interesting. I have been actually having the impression my EX1 looks softer since I upgraded it to V1.20 even though I know this is probably not possible.

Michael, this is why I brought up the importance of knowing the firmware version. It's not that anything got softer though. It's quite possible that the defaults are different now. That "0" (or STD 3 for example) means something different. Of course this would mean PP are not directly comparable when copied. It may well have been the result of Sony hearing "too sharp" complaints and didn't want to lose sales as a result. Of course this is all speculative but seeing EX1 1.11 or thereabouts vs EX1 1.20 vs EX1R at "default" settings might tell the story.

After the first shock in this thread I checked again all my test footage and stills I made that evening. I had some 1080 shots too, but in 50i only. I forgot to shoot in 1080p and now the camera is gone to its new owner.

But switching on a good 24'' monitor back and forth between several very similar stills of the EX1 and EX1R I am now relaxed again. The sharpness/resolution-capatibilities between the 2 cams are so close, I do not find a real difference anymore. In some frames maybe, yes, but changing to a another frame with little different focus the fine lines within the stars and the numbers seem to be exactly the same.

What remains obvious is, that in my tests the EX1R is in any case more yellowish and seem to be not that clear as the EX1. But maybe this was even more neutral and "better", since my tests took place in a real low light condition with warm soda colored tungsten lights. There is no possibility for me to repeat my tests in brighter situations, maybe under neon light and with 1080p.
So I will remember to decrease WB for 100 to 300 Kelvin for being closer to the "expected" results till I am used to the new settings of the EX1R.
(And maybe I really forgot to reset a setting in the EX1R, although ClipBrowser shows identical values.)

Thanks especially to Craig for calming me down...

It's quite possible that the defaults are different now. ^^ It may well have been the result of Sony hearing "too sharp" complaints and didn't want to lose sales as a result. Of course this is all speculative but seeing EX1 1.11 or thereabouts vs EX1 1.20 vs EX1R at "default" settings might tell the story.

Seasoned to my taste, I usually set my EX1 ver 1.11 to about -30 detail. The same look on my new EX1R is about -5. My EX1 rung like a bell at detail 0, the new EX1R, hardly at all. Much less over enhancement on my EX1R, ver 1.0 and 1.10.

Pick two different test cameras and you may get different results.

After the first shock in this thread ...

no kidding, if it looked that bad i would have sent mine back.

So how do i get me ex3 to look sharp as my Canon 7D, i have to say, that with the sharpness turned all the way down on the 7D shooting video it still looks so good, so instead of tying to get the 7D to look like the Ex3 (as i have read in the 7D area) - i would like to make a Ex3 setting sharp , any one got a good PP, or setting.
Rob.

Robert in my side by side tests the EX-1 ate the Canon's lunch when it came to resolution. No contest whatsoever.

Robert in my side by side tests the EX-1 ate the Canon's lunch when it came to resolution. No contest whatsoever.

Indeed, I think he might need to clean the lens. An EX resolves @ 1000 lines, and the 7D @ 600. I wonder what he's viewing the image on?

Probably mistaking aliasing for resolution. The Canons are actually very soft compared to an EX, in reality they have little more resolution than a good SD camera but aliasing can make the pictures deceptively sharp.

To look at this from another angle. what's interesting is to see how two modern cameras (the EX1 and EX1R) made a few years apart can be so very close in performance. In film days (and I mean movies and stills) lens technology (design, manufacture, coating, assembly and inspection) meant two cameras with consecutive serial numbers could exhibit pretty startling resolution differences on screen.

Generally our modern camcorders are designed to be used at full aperture, as the tiny chips we employ dissuade us from stopping down into diffraction. In film days we learnt never to shoot till we'd stopped down at least two stops in an effort to hide some of the losses caused by element miss-centering and zoom lens mechanical failings.

I'm constantly amazed at how cheap 1080 domestic camcorders happily fill 50" screens and even when obviously used wide open, remain sharp right into the corners. This says a huge amount about component and production tolerances. Aspherical surfaces have helped in this respect, but even so, a 10x zoom today uses the same number of elements that a 10x zoom used 40 years ago.

tom.

Below are some frame grabs from an EX and a Canon 550D/T2i. the difference is pretty clear.

but even so, a 10x zoom today uses the same number of elements that a 10x zoom used 40 years ago.

tom.

That's not always true. Cameras like the EX use a much simpler varifocal zoom design with fewer elements than the parfocal lenses that were essential 40 years ago. Parfocal Zoom lenses used to have to stay in focus throughout the zoom range and this needs some very fancy design. Many modern video cameras now use lenses that don't track focus, but use an electronic look up table to adjust the focus as you zoom. This make the lens much cheaper to produce as it needs fewer elements.

5D Mark II 24-105 vs. EX1R

Ex1R has a very good idea, we used identical diaphragm 8 and 1 / 100 exposure

you shouldn't use f8 to test anything on an EX-1 as it will have already started losing sharpness due to the small f stop- sad but true . Best to test at 5.6 or more open. EX-1 looks terrible at small f stops.

I honestly do not see visible differences, we shot with the EX1 and diaphragm 8 .. probleme.Am had no major problems from the diaphragm 10 up :-) but only once.
usually use maximum aperture until 8, but now I EX1R that lens is not so good.are quite large differences in image corners,1.10 software does bring something new to the improvement of focus? may be a mistake by Sony technicians, should I try a new calibration to target?

Docea,
Don't know how you tested or what kind of monitor your using . If you wish to maintain that the EX-1 or Ex1r resolution doesn't deteriorate after f8 then be my guest, but it has been tested by many.

The EX is one of the sharpest prosumer cameras about, most people turn off or reduce Detail setting to at least -20. It might help if you posted a screen shot of your EX and Canon side by side to illustrate the problem.

At a guess, your using very high f-stops, causing diffraction limiting. Any 1/2" sensor over F5.6 will begin to soften, worse as your stop down. The larger the sensor, like your 7D, the further you can stop down before limiting.

As well as agreeing that an EX needs to be kept more open than f8 to avoid diffraction effects from softening the picture I would again add that you really need to be sure about what you are looking at.

The Canon DSLRs don't have an effective anti alias filter and the way the sensor is read means that the image is seriously under sampled, as a result the pictures are full of alias artifacts which can be mistaken for detail. For example, what should be a smooth edge may appear to be a very sharp, instant light to dark edge as the sensor sampling isn't high enough to resolve the smoothness of the edge. This isn't real detail, it's fake detail. It can look OK, but any camera movement will make the edge flicker as it jumps from one pixel to the next and straight lines take on a stair stepped appearance. This is aliasing and most cameras normally have a filter to soften the image a little so that stair stepping etc doesn't occur (that's why an EX only resolves 1000 TVL despite having 1920 sensors). The best way to see the difference in resolution between an EX and Canon DSLR (other than a chart) is to look at how they both handle subtle textures and things like wide shots of grass or foliage. In the frame grabs I posted earlier look at difference between the texture of the in focus flower blooms, it's a night and day difference. The other thing to take into account is contrast, a high contrast image will appear to be sharper than a low contrast image, so unless exposure and contrast are the same making a subjective assessment is difficult.

Sharpness and resolution are too different things. The resolution is the amount of fine detail that can be resolved or seen. Sharpness is how quickly the image goes from light to dark or how pronounced edges are. You can increase the sharpness of an EX (by turning up the detail level) and create a very sharp image with high contrast edges but this will not increase the resolution of the image, all it does is make the small details more pronounced, perhaps un-naturally so. Taken to extremes it will introduce aliasing.

alister-chapman, its a personl taste i likr the look of the 7d, so i am going to play with the detail,
Ta,

This is aliasing and most cameras normally have a filter to soften the image a little so that stair stepping etc doesn't occur (that's why an EX only resolves 1000 TVL despite having 1920 sensors).

Point of clarification...TV lines are the lines that can be counted within a circle whose diameter is equal to the vertical frame dimension. So the nyquist limit for either horizontal or vertical is just 1080 TV lines. If Imatest is used to report the MTF50 transfer, the resolution number is measured at the point where the contrast between the black lines and the white is 50%. But if you extend the reading out to MTF40,30,20%...there is still measurable resolution all the way to the nyquist limit, it just happens at reduced contrast. For the EX1/EX3/PMW350 the high MTF resolution is not being mistaken for aliasing. The key is the smoothness and shape of the transfer curve itself. If it has squiggly amplitudes that are resonating, that's aliasing. But if it's smooth all the way to the nyquist frequency limit, it's resolving all the detail the format permits.

One of the keys of understanding thus, is that it is not the absolute number that matters most always, more informative is the shape and smoothness of the transfer curve, and the flatness which defines the rate at which contrast decays as frequency is increased.

Two cameras could share equal MTF50 numbers of 1000 TV lines, but the one whose contrast decays at a lower rate will be perceived to have the greater transparency, better reality.

Sharpening (edge enhancement) on the other hand, primarily creates the illusion of sharpness by drawing outlines around shapes. This generally occurs in the frequency range of 200 to 500 TV lines. It's not possible to draw a line less than one pixel wide, therefore details more than 520 lines cannot be enhanced. The real high frequency information cannot be improved, and often what is mistaken for real resolution is the edge enhancement outlines that are being applied at much lower frequencies, particularly needed if the camera has a steep rolloff in contrast ratio.

Regardless of the subjective preference for the 7D (or the 5DMkII that I own), they don't do full raster sampling with video, and the MTF50 resolution is no better than 550-650 lines. Heavy amounts of edge enhancement are applied to the lower frequencies to assist the perception of detail.

There is another type of aliasing (stairstepping) separate from edge enhancement related artifacts, caused by the temporal difference of motion between the frames. Sampling every 3rd line as happens with the Canon DSLR's only makes this worse. It cannot be eliminated by reducing applied sharpening.

What mitigates many of the ugly aliasing artifacts in the 5D/7D is the shallow depth of field permits a larger percentage of the frame's subject matter to remain out of focus, smoothing the jaggies, while drawing attention to the intended subject which is the unique benefit of DSLR video.