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

I'm not sure I see the merits of a DR test that tests the range by using different isos. Many cameras have very different ways of implementing what they label as "iso" sensitivity modes. If you are looking to do a minimal test that will be somewhat generally indicative of the potential performance of a camera as possible, it is probably best to test low iso modes first and probably a particularly low contrast gamma/matrix mode Ideally I'd want to test every mode to really get a sense of what the camera is doing and what the numbers all mean in your iso and scene file modes.

I do plan to do a simultaneous test of the af100 and gh2 alongside my 7D when I get the opportunity probably in a week or two.
 
Further compounding it, is the notion of using the iris for testing -- because the t-stop may not match the f-stop, and therefore you may not actually be changing the amount of light entering the camera by a genuine f-stop. I've used shutter speed in the past for this, as it should ideally be a perfect mathematical halving of the amount of light entering the camera. Even so, if you're using something other than a backlit chart, you'll need a light meter to be able to determine exactly what your chart is showing you. Don't just assume that it's an f-stop between bars, it may not be.

Like I said, it's just too prone to error. The only way to really truly know is to use a backlit precision video chart.
 
nyvz said:
I'm not sure I see the merits of a DR test that tests the range by using different isos. Many cameras have very different ways of implementing what they label as "iso" sensitivity modes. If you are looking to do a minimal test that will be somewhat generally indicative of the potential performance of a camera as possible, it is probably best to test low iso modes first and probably a particularly low contrast gamma/matrix mode Ideally I'd want to test every mode to really get a sense of what the camera is doing and what the numbers all mean in your iso and scene file modes.

I do plan to do a simultaneous test of the af100 and gh2 alongside my 7D when I get the opportunity probably in a week or two.
Click to expand...

I think you misunderstood the post above. One is a test of 5d video dynamic range by provideocoalition to see how much it is. It seem from his test that it is more in the 10+ stop than the number of 8.5 that is brandied every now and then. His test is very well documented and perhaps the best way to test scientifically dynamic range.

The second on is the Dxomark result of the 7d. If you look at the dynamic range result the highest is at the lower iso and falls from the 200 iso mark. So I was responding to someone that was saying that a test at 800 iso was ok to show its dynamic range. I was saying that you can' t demonstrate the highest DR of the 7d at this high iso level as proved by the Dxomark test.
 
Can I just ask of anyone using ISO ( ASA) to test dynamic range in stops...

Are you factoring ISO/ ASA and f-stops correctly?

eg.

100 asa to 200 ASA is one stop.

100 to 400 is two stops

100 to 800 is 3 stops

100 to 1600 is 4 stops

etc...

Now can I ask... even when anyone is aware of the dynamic range of this camera... what are you going to do any differently???

Just curious.

and just for reference ---- best film dynamic range is 14 stops
 
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...

Thanks for the tip, I'll try that out.
 
Shooter said:
Can I just ask of anyone using ISO ( ASA) to test dynamic range in stops...

Are you factoring ISO/ ASA and f-stops correctly?

eg.

100 asa to 200 ASA is one stop.

100 to 400 is two stops

100 to 800 is 3 stops

100 to 1600 is 4 stops

etc...

Now can I ask... even when anyone is aware of the dynamic range of this camera... what are you going to do any differently???

Just curious.

and just for reference ---- best film dynamic range is 14 stops
Click to expand...

here's my answer to what I will do differently when I know the d ynamic range of this camera:

I will decide to either purchase it or not to purchase it.
Because if the dynamic range is poor then I will know this is not the camera for me.
 
Homunculus said:
here's my answer to what I will do differently when I know the d ynamic range of this camera:

I will decide to either purchase it or not to purchase it.
Because if the dynamic range is poor then I will know this is not the camera for me.
Click to expand...

It aint poor.
 
Not poor at all. At the end of the day, all the tests say the same thing. Don't let the camera blow out on Cine-D gamma. Keep it under 100IRE and the image that is left over is really juicy with good lighting. From what I've seen, the camera really excels when treated right, but also works the other way just as badly. The camera doesn't like broccoli. So let's not feed it broccoli!
 
PaPa said:
Not poor at all. At the end of the day, all the tests say the same thing. Don't let the camera blow out on Cine-D gamma. Keep it under 100IRE and the image that is left over is really juicy with good lighting. From what I've seen, the camera really excels when treated right, but also works the other way just as badly. The camera doesn't like broccoli. So let's not feed it broccoli!
Click to expand...
i agree on overcast day the images often look down right amazing and filmic. but how fast it clips is utterly troubling
 
I did a simple DR test on my AF100. Set up a shot of a white card at just clipping and then used ND filters to bring it down. I saw somewhat more than 10 stops, but not 11. Certainly not as scientific as a proper backlit chart, but I figured ND filters (Schneider and Formatt) would be more accurate than the aperture (no t-stops on my lenses).
 
I'm going to have some fun with DRS +1 and see how that affects posterization, sharpness and general noise level.
 
PaPa said:
I'm going to have some fun with DRS +1 and see how that affects posterization, sharpness and general noise level.
Click to expand...

Would be interesting to see. I also want to figure out if there is anything I can set in camera to reduce that aliasing on the light a little. Did you try any other settings to see if maybe some of the algorithms in the camera were contributing to sharpening that edge difference and causing the steps?
 
StephenM said:
I did a simple DR test on my AF100. Set up a shot of a white card at just clipping and then used ND filters to bring it down. I saw somewhat more than 10 stops, but not 11. Certainly not as scientific as a proper backlit chart, but I figured ND filters (Schneider and Formatt) would be more accurate than the aperture (no t-stops on my lenses).
Click to expand...

My initial test on the most precise testing instrument for this purpose, showed 10.5 stops. So if you just found between 10 and 11 on a production model, then... that sounds just like it should.
 
grimepoch said:
Would be interesting to see. I also want to figure out if there is anything I can set in camera to reduce that aliasing on the light a little. Did you try any other settings to see if maybe some of the algorithms in the camera were contributing to sharpening that edge difference and causing the steps?
Click to expand...

Hey, no, but i noticed the aliasing as well.
I believe i turned my settings down to -7 detail and -7Vdetail with coring only at +1. trying to see if those things will make a difference in the posterization i'm seeing as well as the general detail.

I have to say though, I'm a little bit of a fan of DR+1, be able to protect highlights a little better as well as still getting good detail in the darker range is looking real good, with a very acceptable level of noise IMO.

I basically pointed the camera at the white lamp, used my marshall monitor to get the blowout point right along the top edges ( where my false color BARELY gives me red ) and then used that to see on both cameras, at that exposure, how much detail they could get in comparison. Interesting test, but again, please this isn't scientific, just a run n gun inquiry on certain image details.

Then i played the videos again, but this time they were both given the same S gamma curve in post, and then color matched to be as similar as i could get them with very very minimal color management.

Don't use this to compare detail and sharpness though, I believe one of the lenses got a slight bump and may be every so slightly off on focal planes. Detail indication is still very visible however.

On the AF100 i noticed that blowouts tend to go to 110 IRE on my waveform while the 7D clipped at 100. So i did adjust the curve slightly on the AF100 for white cilpping so that they were more similar in value.

Will post it when it's done.
 
Just finished doing another skew experiment between the two. AF had an 85mm and the 7D was probably near the 100mm mark, however much match the viewing angle. Will post soon.
 
PaPa said:
Just finished doing another skew experiment between the two. AF had an 85mm and the 7D was probably near the 100mm mark, however much match the viewing angle. Will post soon.
Click to expand...

I am going to make time this weekend to do my skew test, which is more of a scientific method so we'll know. I also have the skew measurements made by someone else on the 7D so I can test that my results are consistent with his results. Basically, making multiple measurements and then looking at the deviation in them, which for him was around 3% I think max, it should really let us know what to expect.

For those curios, the best thing about this method is the non-reliance on focal length. You can pick one that gives you enough slant to make it easier to read. Basically you take a straight line on the wall and you pan past it. Then, you can measure how far anyone one row on the image to the next image changes in location horizontally to how many lines a line moves from the top scan line to the bottom, also in pixels. This delta along with the framerate allows you to calculate the read/reset time very nicely. Also, you use a fast shutter to eliminate any motion blur, but, you always measure from the same side of the line. Very cool method.

As for the aliasing, I am also going to setup a scene with barn doors like another user and then use 1:1 on my DP1X and mess with the settings will observing that edge. Want to see if I can figure out what is causing it. Will let you know.
 
Here it is. Just to recap:

I have to say though, I'm a little bit of a fan of DR+1, be able to protect highlights a little better as well as still getting good detail in the darker range is looking real good, with a very acceptable level of noise IMO.

I basically pointed the camera at the white lamp, used my marshall monitor to get the blowout point right along the top edges ( where my false color BARELY gives me red ) and then used that to see on both cameras, at that exposure, how much detail they could get in comparison. Interesting test, but again, please this isn't scientific, just a run n gun inquiry on certain image details.

Then i played the videos again, but this time they were both given the same S gamma curve in post, and then color matched to be as similar as i could get them with very very minimal color management.

Don't use this to compare detail and sharpness though, I believe one of the lenses got a slight bump and may be every so slightly off on focal planes. Detail indication is still very visible however.

On the AF100 i noticed that blowouts tend to go to 110 IRE on my waveform while the 7D clipped at 100. So i did adjust the curve slightly on the AF100 for white cilpping so that they were more similar in value.
Click to expand...

 
Last edited:
I am assuming the top video is the AF100? :) One thing I notice, is the shutter speeds different? The clarity of the LCD on the stove and the strobing of the television are different between the two. Seems like that 7D has a little bit more motion blur to it.

That said, wow the AF100 is much more detailed almost everywhere. Especially that thing on the wall in the laundry room. It's interesting how fast the youtube compressor goes to crap, but when in the bathroom the second time, the box of tissues between the two is quite different.

Also, when you go by the darker couch, check out the moire in the bottom clip! Someone was sitting on that couch with oil on their pants :P

This is exactly what I wanted to see in your test, exposed to the same level of hightlights, and to me, I think the AF100 clearly looks better in like 85% of the video, the other 15% being hard to tell a difference. I also watched this on a DELL LCD that amplifies all the horrible qualities of any camera (it is amazing at that) and the AF100 footage in the darks held out REALLY well. Same settings as you were using before? (that you posted)
 
You must log in or register to reply here.
Back
Top