xf-300 vs sony EX1R sensor size?

[Nick Wilcox-Brown]

I have borrowed an XF300 for a special project and got the chance to make a very quick comparison with the EX1 at normal gain settings. My colleague's exposure seems to be a fraction lower than mine, so I have lifted the EX gamma by 15% for the split screen shot. Aperture was f5.6 on both cameras, data rate at maximum and colour at defaults. From memory the Sony gain was minimum and XF300 was at -3dB.

The XF300 was very noticeably sharper at 1080 and there is none of the colour smearing that is obvious on the EX1 in the leaves and the grass behind the car.

Canon XF300 v Sony EX1 on Vimeo

I am very happy to upload the MXF file from the Canon for anyone interested, but my colleague requested that the EX clip is not put online.

[Nick Wilcox-Brown]

Better than adding bulky downloads of static clips - here are the two full screen grabs at max quality jpeg for comparison:

The EX1 is the darker clip, the XF300 lighter - both are untouched grabs from the original camera files.

Nick.

[Steve Phillipps]

Wow, that's pretty impressive.
I wonder how much is due to Detail settings? At f5.6 you'd expect both lenses to be pretty decent so that shouldn't make a massive difference I wouldn't have thought.
I'm quite surprised at how much better it does look.
Steve

[Nick Wilcox-Brown]

Steve,

The camera was used on default settings (no CP set), so detail was at the default of zero (range -8 to 8). Time has been short, so I have not been able to investigate all the CP settings, but I understand that as a departure from previous cameras, some of the settings represent 'standards' removing some of the need to create custom looks.

Nick.

[Tim Polster]

The road shows a lot more detail with the Canon. Would be interesting to see the same scene at wide open or tele or both and the EX with a Nano & 50mbps.

I would also be interested to see a DOF comparison between the two just to know for sure what difference there may be.

[Jose Dominguez]

I don't know if it is my monitor, but to me the EX1 colors look more realistic. the slide darker color of the car looks better to me.

[Buck Forester]

When I was a newbie I used to love these kinds of tests, but now that I have an EX1 and see how one can make any shot appear with all the settings, these tests are basically useless. I don't say that to be critical of the test-maker, I'm sure in 'most' cases they mean well, while in others it's biased based on settings.

An example of an absurdly biased test is the first one that Tans Mark linked to (nothing against Tans, he was just linking to it, not his tests). It showed Canon as a lovely, bright daylight image of some bushes compared to a dull, darker scene of the EX1. C'mon, one would have to have no idea whatsoever about cameras to actually be influenced by such a basic thing as shooting leaves in any easy focal range and easy lighting and having such a huge difference. I could make my kids toy camera look better than the EX1 was shown on that clip.

The one showing the car and the 'less sharp' EX1 is not accurate either. I can make my EX1 razor sharp throughout the entire image at that distance, or make the background look a little muddled depending on my settings.

I'm not biased either, I absolutely love Canon and am a Canon supporter and I have tens of thousands of dollars invested in Canon still photography bodies and lenses. I also have the Canon 5D Mark II, awesome camera for both stills and vids.

For those who think the 1/3" vs. 1/2" chips aren't really that big of a factor, well for many of us it really is huge. I would have seriously considered this new Canon video camera if they put 1/2" sensors in it. Anything 1/3" is not on the table with me, I've been spoiled by 1/2" with my EX1 and I'm not going back, especially in that price range. I hope Canon puts the MK II sensor in a full video camera body, unencumbered, and I'm all over it (or at least will be a contender with Scarlet for me). Going from 1/3" to 1/2" is pretty huge, and going from 1/2" to 2/3" is pretty huge. And full frame is pretty huge. These things do matter to many of us and I'm a little surprised that Canon, in this price range, with their amazing chip making capabilities, with a brand spanking new camera, went with 1/3" chips when the market and all the buzz is trending otherwise.

Just my thoughts... people smarter than me may feel differently.

[Tim Polster]

Buck, I agree with you.

I have been watching this camera as I like to keep up on the market and I wonder about its market position. Is Canon trying to get new camera owners or existing camera owners to switch?

I would think there are a lot of people who already own a video camera in this segment (somebody who would pay $8,000 for a video camera). I am in this category and there is not not enough reason to make a switch.

To be honest, if this series had two things I would consider selling everything and switching in order to have matched cameras. - 1/2" chips and a JVC style mini-shoulder mount model (fixed lens is fine). Having a larger shoulder option along with a smaller version gives options. For some reason which we can only speculate they did not or will not go there.

For those who say 1/3" is fine and there is no difference, well I can only point to history and the pecking order that has always been there. There is a reason why the big money uses the bigger sensors. It is just too convenient for the manufacturers to say its all different now. I can't agree. If it looks great at 1/3", then it will look better at 1/2". Why not "go there", take over the space and remove all doubt?

With the rise of the vDSLRs and 4/3s video cameras like the upcoming Panasonic model, these 1/3" chip cameras will look limited in comparison going forward, by price and performance.

But I am just one situation though...

Anyway, looks like a wonderful camera, just not quite enough for the price to get my interest.

[Doug Jensen]

Buck and Tim, you're right on. You have summed it up perfectly.

Thanks.

[Peter Moretti]

Originally Posted by Buck Forester

...

For those who think the 1/3" vs. 1/2" chips aren't really that big of a factor, well for many of us it really is huge. I would have seriously considered this new Canon video camera if they put 1/2" sensors in it. Anything 1/3" is not on the table with me, I've been spoiled by 1/2" with my EX1 and I'm not going back, especially in that price range. I hope Canon puts the MK II sensor in a full video camera body, unencumbered, and I'm all over it (or at least will be a contender with Scarlet for me). Going from 1/3" to 1/2" is pretty huge, and going from 1/2" to 2/3" is pretty huge. And full frame is pretty huge. These things do matter to many of us and I'm a little surprised that Canon, in this price range, with their amazing chip making capabilities, with a brand spanking new camera, went with 1/3" chips when the market and all the buzz is trending otherwise.

Just my thoughts... people smarter than me may feel differently.

Buck, I also think that 1/2 is a much bigger advantage over 1/3 than people are admitting. And if this camera were 1/2, I'd buy it tomorrow.

So Sony has the advantage when it comes to sensor size. BUT the Canon has the advantage when it comes to codec. 4:2:2 with a higher data rate is significant.

The only way I can explain the results of the test above (other than the EX being poorly set) is that the detail in the image overwhelmed the XDCAM-EX 35mbps codec. I'd like to see a real head-to-head between the two cameras using production models and more extensive testing.

But if the comparison shown above really does show both cameras at their best (a rather big "If"), then the Canon has significantly higher resolution in real world compositions (as opposed to zone plate charts that don't stress the codec nearly as much).

So it may come down to what's more important to you: higher resolution or shallower DoF?

[David Heath]

Originally Posted by Brian Woods

In regards to Canon pulling some other magic out of the bag, I will say this: When I was playing with the XF305 at Cinegear, I zoomed all the way in (18x), pointed at some high contrast vertical lines, and whip panned away at various speeds. I couldn't see any rolling cmos skew in the LCD.

That's an interesting observation, but I don't think it proves anything conclusively. It's conceivable that the LCD was being scanned in such a way as to balance out the rolling shutter effect. To prove the matter. you'd really need to connect the camera and another camera for comparison to the same external monitor and do the same test to each.

Originally Posted by Tim Polster

To be honest, if this series had two things I would consider selling everything and switching in order to have matched cameras. - 1/2" chips and a JVC style mini-shoulder mount model (fixed lens is fine).

Exactly my thoughts as well. I commend Canon for the choice of codec, and also the use of Compact Flash. Maybe there are technical issues surrounding the use of 1/2" chips which would just make it a bridge too far, but I don't see any real excuse for the form factor. We're talking about an expensive camera here by 1/3" standards, there's just no excuse for it having the form and ergonomics of a consumr camera. JVC have shown it can be done, Canon should have done something similar.

[Steve Phillipps]

I think the lack of 1/2" chips can only be down to one of two things: this is already a (surpisingly to a lot of people) expensive camera, using 1/2" chips may just have pushed the price too high for it have a market, not only would the chip and other hardware have cost more the excellent lens would have been more expensive too to cover 1/2". The second possibility is not wanting to rock the boat, as a 1/2" camera with 50 mb/s codec really have caused huge waves with Canon's semi-partners (ie they make a lot of lenses for their cameras) Sony and Panasonic.
I still don't quite see such a big issue with 1/3" vs 1/2", it is after all only 1 stop difference. So just shoot f2 instead of 2.8 or f4 instead of 5.6. At the other end you've got diffraction limits - f4 vs f5.6 theoretically, though in real world terms probably f5.6 vs f8. So the Canon will have a useable aperture range of f1.6 to f5.6, surely that's workable?
As long as the sensitivity and noise are not too far off the 1/2" camera's as in this case then surely it can't be that big a problem. Having said that I don't use either of them so I admit I'm not talking from specific experience here.
Steve

[Chris Hurd]

Originally Posted by David Heath

there's just no excuse for it having the form and ergonomics of a consumr camera.

I don't think there's anything "consumer" about it -- it's not significantly different from the Canon XH, the Panasonic HVX, or the Sony EX1, and those certainly are not consumer camcorders.

My opinion is that their decision to go with 1/3" was based on the strategy of their overall product roadmap. The XF is most likely just the handheld version of a forthcoming shoulder-mount, probably the "XL F" and in order to be called an XL and maintain compatibility with existing XL HD lenses, it must have an XL mount and therefore it must be 1/3" -- in all honesty, who couldn't see that coming?

[David Heath]

Originally Posted by Steve Phillipps

I still don't quite see such a big issue with 1/3" vs 1/2", it is after all only 1 stop difference. So just shoot f2 instead of 2.8 or f4 instead of 5.6.

I'm more used to 2/3" cameras, but even there often find myself juggling with ND and lighting to make sure I'm very close to wide open. So I don't see any move to 1/2" chips as welcome - and 1/3" even less so. I'm sure it's possible to make do with 1/3" and accept interview backgrounds that are distractingly sharp - but if there exists an alternative, why not go for it?

A move to 2/3" may be prohibitively expensive for many, but such as the EX is actually a lot cheaper than the X305.

Originally Posted by Chris Hurd

I don't think there's anything "consumer" about it -- it's not significantly different from the Canon XH, the Panasonic HVX, or the Sony EX1, and those certainly are not consumer camcorders.

What I said was that it had "the form and ergonomics of a consumer camera" and in relation to handheld use I stand by that statement, even if in other respects it's far from "consumer". I've had plenty of good things overall to say about the EX - but handheld ergonomics is certainly not one of them - and I know I'm far from alone in thinking that. The same applies to the Panasonic HVX, which doesn't even have the true manual lens of the EX. Unfortunately a lot of people have grown up knowing nothing other than this type of styling - that doesn't mean it can't be improved on.

Tim Polster brought up the JVC styling as how a camera in this class CAN exhibit pro form factor and good handheld ergonomics. I agree with him - in this respect, it is a mile ahead of the XF305 and the three you name, I'd also add the EX3 as a half way house.

[Steve Phillipps]

I'm used to 2/3" cameras too, and I never seem to have this ND juggling problem. The NDs go in steps of 2 stops on Varicam and the like, so you have 2, 4 and 6 stops of ND. The 2x extender is 2 stops too, so that fits in nicely.
Say you're at f2 and the light changes a bit, stop down to f2.8, no big deal. If it changes more and you don't want to use f4 then stick 1 ND in and go back to f2. I personally find it such a non-problem, and very intuitive to work with.
Steve