Canon 7D vs Pasonic AF100 - Skew, DR, etc..

PaPa said:
Bingo.

going to do some "tests" on neutral gray with same lenses on both cameras.
Click to expand...

By the way, nice house :)

Have you tried using some of the other curve settings as well? That's one thing I need to try. I think Cine-V squashes down the high end too much for me liking, but it expands the lower end the least, which I do like :) I so want to create my own curve!
 
Cory Braun said:
I disagree with this. From my INITIAL testing, I have noticed that the AF100 has more noise than the Canon cameras. Again, this is from my initial testing. I still have many more settings and options to explore to find the ideal settings for each situation, but as of now, especially out of the box for both cameras, I have noticed more noise with the AF100 than the Canons.
Click to expand...

This is true, and what I noticed out of the box. But, I don't think it's actual real unaffected noise you are seeing, as I crank the detail setting down, all that noise goes away. When I compare the relative sharpness of both images, out of the box, the AF100 was WAY sharper, and I think the sharpening of the noise to that point is hurting the image, in my opinion. Try pulling down the detail and using something like Cine-V with the Flou setting. ON mine, it is SUPER clean at ISO400. But, it's too clean. It exposes banding at that level.
 
astigmatic said:
I have a 7d and all that I've seen so far have confimed my choice to switch over to the af-100. Mine is coming at the end of the month. When I see the highlights clipping but the shadow being preserved I just think no big deal. I'll just make sure I set exposure correctly to preserve the highlights and let the shadows go to black. No big deal.
Click to expand...

You know... you can bring up the black on a 7D quite a bit. There's almost always two stops hidden in what looks like clipped blacks. There's really not any extra noise in there either, noise likes to hide in the mid-tones, which should almost always be brought down a bit.
 
PaPa said:
Very legit questions and concerns. For what it's worth, dealing with highlights in a current manner is a price I'm willing to pay for better color, shadows, sync sound, less aliasing, skew and moire.

But yeah, still waiting to see this puppy shine.
Click to expand...

The GH2 takes care of some of those issues and for a lot less money.
 
grimepoch said:
This is true, and what I noticed out of the box. But, I don't think it's actual real unaffected noise you are seeing, as I crank the detail setting down, all that noise goes away. When I compare the relative sharpness of both images, out of the box, the AF100 was WAY sharper, and I think the sharpening of the noise to that point is hurting the image, in my opinion. Try pulling down the detail and using something like Cine-V with the Flou setting. ON mine, it is SUPER clean at ISO400. But, it's too clean. It exposes banding at that level.
Click to expand...

what IS banding? you said it is exposed if the image is too clean so it seems from what you're saying some sort of inherent part of an image that is only hidden by low level noise and such. I just still can't understand what the heck it is? Is it parts of the sensor that are visible or something like that? or is that FPN I'm talking about. and why do canon cameras not produce that same banding/FPN that the lumix cams do in low light?
 
Homunculus said:
what IS banding? you said it is exposed if the image is too clean so it seems from what you're saying some sort of inherent part of an image that is only hidden by low level noise and such. I just still can't understand what the heck it is? Is it parts of the sensor that are visible or something like that? or is that FPN I'm talking about. and why do canon cameras not produce that same banding/FPN that the lumix cams do in low light?
Click to expand...

Banding is basically when the image doesn't have enough information to create a smooth transitions of all the values between two areas. Banding happens because of low bit-depth and/or low data rate. AF and 7D are both 8-bit color space, but you see more banding on the lamp shot with the AF vs the 7D because the 7D has TWICE the data rate (48mbs vs. 24mbs). AVCHD is still H.264, it's just a different implementation of it, and even if it is a little more efficient (as some claim it to be)... there's no way it's TWICE as efficient. That would be a revolutionary break through.
 
astigmatic said:
I have a 7d and all that I've seen so far have confimed my choice to switch over to the af-100. Mine is coming at the end of the month. When I see the highlights clipping but the shadow being preserved I just think no big deal. I'll just make sure I set exposure correctly to preserve the highlights and let the shadows go to black. No big deal.
Click to expand...


Just a point here. Setting an exposure correctly to "preserve the high lights" does not necessarily mean the shadows will go to black.
 
Homunculus said:
what IS banding? you said it is exposed if the image is too clean so it seems from what you're saying some sort of inherent part of an image that is only hidden by low level noise and such. I just still can't understand what the heck it is? Is it parts of the sensor that are visible or something like that? or is that FPN I'm talking about. and why do canon cameras not produce that same banding/FPN that the lumix cams do in low light?
Click to expand...

extremehalo.jpg


This is an extreme example where i exaggerated the contrast to emphasize. Obviously here it's happening to a great extent, but would be much much much less in the actual video. It is however there. This is what i have been refering to when using the term "banding". Perhaps I am not using it in the right way, but just to clear up any confusion, I thought i would make it simple.
 
This is banding.

This is banding.

Homunculus said:
what IS banding? you said it is exposed if the image is too clean so it seems from what you're saying some sort of inherent part of an image that is only hidden by low level noise and such. I just still can't understand what the heck it is? Is it parts of the sensor that are visible or something like that? or is that FPN I'm talking about. and why do canon cameras not produce that same banding/FPN that the lumix cams do in low light?
Click to expand...

This is banding; shot with the 5DII. Note the lines in her face and on the gradient background.

This is a simple, very basic shot; it would, by all but the most cavalier standards, be unacceptable.

Good shooting and best regards,

Leo
 

Attachments

  • girlmoiree1.jpg
    girlmoiree1.jpg
    38.7 KB · Views: 0
Ticheli said:
This is banding; shot with the 5DII. Note the lines in her face and on the gradient background.

This is a simple, very basic shot; it would, by all but the most cavalier standards, be unacceptable.

Good shooting and best regards,

Leo
Click to expand...

Interesting. So i guess what i'm seeing is banding.
 
editman said:
I guess we expect this camera to be flawless but as it turns out it's just another tool.
Click to expand...

Exactly.

Im not sure whats going on out there in the world - I dont have my AF100 yet, but apparently us mods work for Panasonic so whats up with that? :P But, I did shoot with it for a day back in November and got fantastic results. Great DR for the price - better than my 5D, sharp with my detail turned down, nice tones etc So why is it so hard for others to get that on a production model?

Papa I dont mean to single you out here, but honestly none of the shots in this thread are shot optimally for the AF100 or for that matter the 7D. My point is, if people want to get great shots take the time with any piece of gear to do your homework to get the best looking shot out of the camera (tool) you are using - seems no one wants to do any work these days for quality results.

These kind of tests are meaningless. Go shoot something worthwhile, and if that takes work to make it happen well then thats what you have to do.
 
Noel Evans said:
Exactly.

Im not sure whats going on out there in the world - I dont have my AF100 yet, but apparently us mods work for Panasonic so whats up with that? :P But, I did shoot with it for a day back in November and got fantastic results. Great DR for the price - better than my 5D, sharp with my detail turned down, nice tones etc So why is it so hard for others to get that on a production model?

Papa I dont mean to single you out here, but honestly none of the shots in this thread are shot optimally for the AF100 or for that matter the 7D. My point is, if people want to get great shots take the time with any piece of gear to do your homework to get the best looking shot out of the camera (tool) you are using - seems no one wants to do any work these days for quality results.

These kind of tests are meaningless. Go shoot something worthwhile, and if that takes work to make it happen well then thats what you have to do.
Click to expand...

Hey Noel, appreciate the response.

I am aware that the shots were not optimally done, that was essentially the point ( for me at least ). I have produced results I am very happy with the camera where things are framed and appropriately lit. For me what's interesting sometimes is where the camera falls apart, so that when it comes time, which happens frequently enough, where I really need to make the image shine, I know what to avoid, especially if time, budget, and gear is a factor. And for what it's worth, I have learned some important aspects by doing these tests, however worthless they may be. Surely you can appreciate that. So for me, this has only been useful and I hope the same thing for others looking at what these tests for what they really are, without taking it any further.

Thanks.
 
Last edited:
ah I see, so that "gradient" effect as a byproduct of 8bit is what most people term banding. in my head i just called them gradients. i guess the other thing which to me is far worse is the fixed pattern noise which the gh13 produces and which I don't think (and hope not) that AF100 produces. that is the vertical streaks that appear in low light video and i am really curious what in the world that is and what cuases it, as opposed to banding which I sort of have a decent scientific notion of how that is caused due to 8bit technology. even the grainy type noise i can live with, it's the FPN that is abominable and ruins everything in low light on gh1
 
That is the banding I am referring to as well. It really stands out if I crank all the noise out of my AF100. However, when I let a little of the noise back in, then I think it looks really really nice.
 
I have some footage that looks just excellent and I have some that are not so great. It's up to me to produce something worthwhile. This camera has some flaws if you look upon it to be flawless. Compared to a DSLR it's about perfect on the other hand.
 
I prefer to use the term "posterization" for what is being described as "banding" as it is more specific. Banding is a proper term as well, but the term can easily be mistaken for the same term used to describe a variety of other image artifacts such as the appearance of certain common patterns of fixed pattern noise, and certain artifacts that resemble digital image sensor smear.
 
Ticheli said:
We chose 800 ASA because the camera goes down only to 200 ASA; had we shot the test at native 400 ASA, we could not have pushed the DR at the bottom.

I'm sure we would enjoy seeing your DR tests.

Good shooting and best regards,

Leo
Click to expand...

I don' t have the pretension to have some test so I rely on people that I can believe in because of their methodology, reputation and work flow to test it. It is the 5d and should be close to the 7D.

http://provideocoalition.com/index....d_how_much_dynamic_range_does_it_have_really/

See here for photographic Raw (not video) how the dynamic range decreases with the ISo. So what is the 7d maximum dynamic range, it is not at 800 iso I can tell you.


http://www.dxomark.com/index.php/en/Camera-Sensor/All-tested-sensors/Canon/EOS-7D
 
Danielvilliers said:
I don' t have the pretension to have some test so I rely on people that I can believe in because of their methodology, reputation and work flow to test it. It is the 5d and should be close to the 7D.

http://provideocoalition.com/index....d_how_much_dynamic_range_does_it_have_really/
Click to expand...

Seems to be nearly the same results as the Zacuto test... 10.5 roughly for the 5D. They rated the 7D even higher at 11. Apperently it's something called a "gapless microlens" that gives the 7D an edge dispite the higher pixel density. Nikon's been using them as well on the D3.
 
You must log in or register to reply here.
Back
Top