PDA

View Full Version : Got my 5d Back



morgan_moore
11-01-2009, 09:08 AM
In this thread (http://www.dvxuser.com/V6/showthread.php?t=188719) Im a comparing the 7d to the EX1, irrelevent really because the EX1 does not offer large sensor, small form factor etc, but my general opinion is that the 7d is wooly I used the EX1 as 'a reference'

I have now got my 5d back and managed to do a quick side by side..

http://www.halfinchrails.com/hir/pics/7dgrabs/7d_sign.jpg

and

http://www.halfinchrails.com/hir/pics/7dgrabs/5d_sign.jpg

My methodology was..

70-200 canon F4 mounted on sticks, composed 5d image at 200mm, repalced 5d with 7d and zoomed back until the composition was similar in terms of FOV

IMO the 5d is acceptable sharp while the 7d is not

Maybe I have a duff 7d

Maybe all 7ds are not as good as 5ds

Maybe the 5d which blurs that background saves some 'codec' for representing the sharp part of the image better

Maybe im seeing smoke and mirrors

But im defo getting rid of the 7d and keeping the 5

S

Kellar42
11-01-2009, 10:10 AM
While I believe that the 5D's overall image quality is going to be proven in some way to be superior to the 7D's, I don't think this is the test to do it!

As you mention in your post, you've got the 5D setup to blur the background, which does change things. Immediately it fails to be a real side by side shot of the same image. This also seems to be one of the more decent 7D grabs I've seen, and there appears to be some kind of color distortion or Moire on the sign in the 5D's grab...hardly an accurate or definitive test in the 5D's favor.

morgan_moore
11-01-2009, 10:17 AM
A typical scene not a test will have a background and a foreground

As one fills the scene with sharpness like a test chart they may equalise, but I dont film test charts

The ability to blur the BG with the 5d may help de-stress the codec

The 5d is in this case just moving to moiree/alise the sign, mainly I dont film signs I film people

But my general impression is that the 5d is more pleasant and the 7d looks wooly at my typical viewing resolution which is 3ft back from my 20 inch monitor, playing fullscreen

S

morgan_moore
11-01-2009, 10:18 AM
They were both at F4 BTW

It is a real situation IMO

morgan_moore
11-01-2009, 10:20 AM
BTW I dont want the 5d to be better, I want to film at 25p right now

I own both (again) so there is no bias

S

mico
11-01-2009, 07:28 PM
Maybe theres something to the 5D aliasing more and the 7D less, so you get more perceived sharpness on the 5D. I guess the test that should be performed is one where there are alot of telephone lines twittering and you pick whether you like more twittering with more sharpness or less sharp with less twittering.

Kholi
11-01-2009, 08:28 PM
This test just doesn't work, for several reasons. THere are too many variables at stake to just conclude, from a sign and one zoom lens, that it's a camera vs camera thing.

It needs to be a lot more scientific than what you're setting up here.

Kholi
11-01-2009, 08:32 PM
Maybe theres something to the 5D aliasing more and the 7D less, so you get more perceived sharpness on the 5D. I guess the test that should be performed is one where there are alot of telephone lines twittering and you pick whether you like more twittering with more sharpness or less sharp with less twittering.
I also do believe that this could be a great possibility.

polr
11-01-2009, 09:29 PM
I've got a 5D until tomorrow morning, so here's a quick shootout of it next to my 7D. Using Morgan's same technique but with a Canon 24-105 L. Identical settings on both cams (Neutral, sharpness 1, contrast @ 0). Pardon the less-than-inspiring subject..

Video: http://www.megaupload.com/?d=9TTDF9YL
Still: http://dl.getdropbox.com/u/225478/5D7D.jpg

In the still (extracted from the video) the 5D seems a tiny hair sharper, and does more so in the video, although aliasing seems up as well. Perhaps Mico is right..

Other observations I've made having both cameras for a couple hours:

-White balance seems more accurate, or pleasing, on the 5D (in this footage, they were both in Tungsten WB mode). Notice the slight greenish-yellowish hue in the 7D footage. The WB was consistently slightly off on the 7D in all tested modes (Tungsten, Daylight and Manual WB)

-Noise is slightly lower on the 5D, especially at higher ISOs, but when HTP is active, the "fixed pattern" noise is more noticeable than on the 7D.

-Dynamic range seems identical to me (notice the identical highlights on the teapot), but depending on whether or not you use HTP, the 7Ds higher, but more pleasing noise might give it an edge in that department.

morgan_moore
11-01-2009, 10:01 PM
Maybe theres something to the 5D aliasing more and the 7D less, so you get more perceived sharpness on the 5D. I guess the test that should be performed is one where there are alot of telephone lines twittering and you pick whether you like more twittering with more sharpness or less sharp with less twittering.

Or you use the 5d to focus on your subject and avoid twittering in my case sticking to the EX1 for wide shots

morgan_moore
11-01-2009, 10:06 PM
This test just doesn't work, for several reasons. THere are too many variables at stake to just conclude, from a sign and one zoom lens, that it's a camera vs camera thing.

It needs to be a lot more scientific than what you're setting up here.

I dont see whats wrong with that, these are shots that are possible with my lenses

I think the complex background may 'eat' the codec and a 7d gives you a complex background

I could have lost the background shooting with my 400 2.8 - but that is not a lens that one can carry

I had space in this situation to back off from the subject but one cannot do that when shooting interiors

practical situations are the only thing that matters to me

S

Kholi
11-01-2009, 10:14 PM
I do understand that, Morgan. But here's another variable that may very well matter: A zoom rarely every performs the same at all focal lengths. I'm not sure which end your 70-200 is best at, but it's a great possibility that-- as well as the other posters 24-105-- the end you're using it on for the 5D to match the 7D is the better end.

The likely hood of what Mico posted is great, that's what I said in your other thread as well.

If I get a chance this week I'll also try it.

polr
11-01-2009, 10:30 PM
Doesn't the fact that the 7D have a cropped sensor, and therefore will always use a smaller part of any lens's surface mean that a 100% foolproof test can never be done?

It is my belief, however, that considering the 5D/7D's relatively low video resolution (compared to their still modes) that lens sharpness variation, especially on good glass, is probably a non-issue.

morgan_moore
11-01-2009, 10:31 PM
You can only buy what lenses are out there

Once you go wide it is a pretty limited selection unless you hit proper T2 cine lenses

To be honest ive done some side to sides full frame and the differences are closer - but Im doing tests that really break both cameras (newsprint from a distance)

Ill try and conceive a scene that is optimal - maybe something tonal like a bowl of fruit

It is the whole package, from sensitivity of focus (harder on 7d) through to ability to monitor that actually bring quality

also it may not be optimal to turn the sharpening off on the 7d - maybe it is better to sharpen before compression ?

vids like the scrambling bikes look great but oversharpened

neither camera does what you can see on the screen in zoom mode, so a external 1080 out would help too (with a recorder)

S

ydgmdlu
11-01-2009, 11:25 PM
White balance seems more accurate, or pleasing, on the 5D (in this footage, they were both in Tungsten WB mode). Notice the slight greenish-yellowish hue in the 7D footage. The WB was consistently slightly off on the 7D in all tested modes (Tungsten, Daylight and Manual WB)
This might be due to a possibly weaker infrared filter on the 7D than on the 5D. IR pollution was discussed on a thread in the main forum...

morgan_moore
11-02-2009, 12:26 AM
Tonal Test,

www.halfinchrails.com/hir/pics/7dgrabs/tone.jpg

Taken 5d and 7d images overlayed and a 5d still reference, in the middle Ive given the curves an extreme tweak, at the bottom Ive added noise (B+W, Gausian distribution)

the scene is the back of a beauty dish sidelit with a flicker free flouro, it has a mildly stippled surface seen lrearly in the 5d (still) image

S

mico
11-02-2009, 05:47 AM
Or you use the 5d to focus on your subject and avoid twittering in my case sticking to the EX1 for wide shots

Thats what I've been doing with the 7D and EX1.