AF-100 Wishlist/Suggestion Discussion Thread

Status
Not open for further replies.
Bassman2003 said:
I would believe the EX-1 was at f1.9.
Click to expand...

It looks like quite a bit of zoom. The EX1 lens is still f1.9? You should compare to f4.8 in this case. But if it was f2.2 at that focal length, it would be f5.5. F4 should be equivalent to f1.6, a faster zoom on the EX3.

EDIT: All that's for 4/3 though. I just read this was a 7D test, so greater difference. In order to match the 7D at f4 you need an f1.2 on the EX1 which is not possible. The f1.9 is f6.5 or so.

EDIT2: If you try to match the DOF by stopping down the 7D, instead of trying two arbitrary apertures, you will see that the equivalent aperture is correct and the DOF of AF-100 is predictable.
 
Last edited:
I will run the test again but wanted to start with the numbers your were speaking of EX-1 wide open and the larger sensor at f4 (even though I know the 7D is a bit larger of a sensor).

If you have any suggestion on how I should setup the tests let me know. I kept the distance to the subject the same as well as match the frame.

I will play with the 7D a bit as not offend anybody with the image.

I will start at f6.3 on the 7D and go up from there to find a match.
 
Bassman2003 said:
I will run the test again but wanted to start with the numbers your were speaking of EX-1 wide open and the larger sensor at f4 (even though I know the 7D is a bit larger of a sensor).

If you have any suggestion on how I should setup the tests let me know. I kept the distance to the subject the same as well as match the frame.

I will play with the 7D a bit as not offend anybody with the image.

I will start at f6.3 on the 7D and go up from there to find a match.
Click to expand...

Just do the same shot from the same position with the D7 at different apertures. Exactly as you did it. The f6.5 should match the f1.9. On the GH1/AF-100 it will be f4.8 or something like that. On the 5D II it will be about f10. The equivalent aperture ratios are identical to the equivalent focal length ratio which are equal to the crop factor between the two cameras.

You won't have to go to smaller apertures:) F6.3 will be quite close.
 
Bassman2003 said:
I would believe the EX-1 was at f1.9.

About the warning, please read my post! This is about the first time I have clicked the thing over to video mode. I know it looked soft, that is why I put the disclaimer! Give a guy a break for a simple DOF test.

Blame the camera not me. I have the 70-200 2.8L in front of it. I could take the black crush away (I did a bit in my NLE) but that would not help with the lack of detail.
Click to expand...

I am not blaming you at all! Sorry if it came out that way. I think your tests were just fine, and very helpful, so thanks. It was clearly a jab at the 7D. Of course, with time, you will learn to avoid such situations, but clearly, using a 7D for filmmaking is very limiting. And of course, no person can make up for the clear lack of detail. Or non-lack of false detail/aliasing.

f/1.9 should match f/6.5, yes. You will find that filming above f/6.5 at day without a significant ND would be rare. Look forward to the test.
 
Last edited:
Here are two new shots.

I tried some new settings on the 7D and the black crush is better but the image was overexposed as I am not setup for video yet. The LCD needs to be calibrated for live video exposure (or I need to be) so I worked a bit on the image. Please don't judge me on the 7d image. Another quick test in between stuff.

Anyway, f6.3 is the winner to the EX at f1.9. I used a 50mm lens on the 7D to lessen the tele on the EX-1 a bit.

-side note- the 7D is going to take a lot of work to get ready for primetime. The still method of white balance is just not convenient but dailing in the number from the EX-1 does not seem to work very well either.

So sling the arrows!

EX - then - 7D
 

Attachments

  • New EX DOF.jpg
    New EX DOF.jpg
    248.7 KB · Views: 0
  • New 7D DOF.jpg
    New 7D DOF.jpg
    238.7 KB · Views: 0
Thanks Bassman. Looks like f/6.3 is a good match. I am guessing it would be something like f/5.6 on the AF100 to the EX1's f/1.9. Any particular reason for using much Gain on the EX1?
 
looks to me you with EX1 you were focused on the plant on the left, but not on the head- face looks soft and out of focus, while with 7D you focused on the head. you can even see aliasing on the thumb, before opening it!
 
Subhadip said:
Thanks Bassman. Looks like f/6.3 is a good match. I am guessing it would be something like f/5.6 on the AF100 to the EX1's f/1.9. Any particular reason for using much Gain on the EX1?
Click to expand...

Sorry once again for the ratty 7D image. I have the LCD setup for still use and being outdoors did not help.

The EX-1 did not have any gain and I zoomed in and focussed on the eye. The still image was taking from the Edius timeline as a tiff. When I open that tiff file in Photoshop I have it set to convert the sRGB for my still work. So the image gets killed in translation. I then had to pump the levels up in photoshop to match the video exposure. The image does not have as much noise on my production monitor, which showed the exposure as being a tad under but that was where two ND's and wide open fell. It was close enough for the DOF test. I guess I need to look at how Edius spits out the stills and maybe using a different color space inside of Photoshop for video stills.

The image gets re-compressed on the DVXUser server as well. So this is not a good representation of anything outside of DOF.

It does highlight how important the AF-100 will be after working with the 7D for video! I had to add a fair amount of sharpening in Edius. I added too much but the image is pretty soft straight from the camera even with the sharpening at 3 or 4.

Lets hope the AF-100 has a nice crisp amount of detail.
 
While the sensor for the AF100 is set to be the one from the GH1, I would like, it not too far off topic, to discuss what sensor might be good for the second generation of this kind of camera, mainly the resolution.

Many on this thread agree to go for a video-only camera, as opposed to a camera that tries to take good stills and good video. Naturally, then, it comes to mind that the GH1 sensor's 12 million pixels are too many. What then, is the ideal pixel count?

Given:

1. You must use one sensor, with a Bayer filter, if you want to use already-made still or cinema lenses
2. Debayering loses resolution, about 20%, so if you are shooting for 1920x1080, you need a 2500x1400 Bayer sensor.
3. More pixels cost you. Given a set image sensor size (4/3 in this case), as you add pixels, you shrink them. This reduces dynamic range and light sensitivity.

I think I answererd my own question. The ideal resolution for a video-only 4/3 (or any Bayer) HD camcorder is 2500 x 1400.
 
A 1080p format does not really need to resolve 1080 lines at high MTF. Higher resolutions need downsampling filtering which introduces problems also and limits the MTF. Many $1000 consumers camcorders come with multimegapixel sensors and they still do not resolve 1080 lines.

A 1080p system with 800 lines of resolution looks great and uses the distribution compression datarate very effectively. If you take a 35mm 1080p bluray shot on film and apply a sharp filter on it, you will see that there is nothing above 800lines. No theatrical 35mm film projection resolves more than 800lines either. When did 1080 lines become a requirement for good HD?

If AF100 gets 700lines at a reasonable MTF without serious luma aliasing it will be great.
 
combatentropy said:
While the sensor for the AF100 is set to be the one from the GH1, I would like, it not too far off topic, to discuss what sensor might be good for the second generation of this kind of camera, mainly the resolution.

Many on this thread agree to go for a video-only camera, as opposed to a camera that tries to take good stills and good video. Naturally, then, it comes to mind that the GH1 sensor's 12 million pixels are too many. What then, is the ideal pixel count?

Given:

1. You must use one sensor, with a Bayer filter, if you want to use already-made still or cinema lenses
2. Debayering loses resolution, about 20%, so if you are shooting for 1920x1080, you need a 2500x1400 Bayer sensor.
3. More pixels cost you. Given a set image sensor size (4/3 in this case), as you add pixels, you shrink them. This reduces dynamic range and light sensitivity.

I think I answererd my own question. The ideal resolution for a video-only 4/3 (or any Bayer) HD camcorder is 2500 x 1400.
Click to expand...

My experience is only with the 7D and not the GH1 sensor, but I can say this, the light sensitivity on the 7D is awesome, so I am happy with their pixel size. Where I see the biggest problem with the DSLR sensors is that, if they were to read EVERY line on the sensor, the read/reset would be even worse (rolling shutter time). My desire would be to capture the ENTIRE sensor and use a GOOD downsampler in camera to 1080p. So they need to work on the capture speeds of the lines AND the downsamplers.

The biggest thing about bayer pattern as well is that the density of green is twice that to red and blue, so the finer the resolution of the sensor, the more you can push out aliasing and moire at the final destination resolution. (I say this because no one puts on a OLPF that is strong enough to reduce ALL aliasing and moire that I have seen)
 
Good points, Otis! I myself actually would be fine with a 1920x1080 sensor (like yours ;) Even though it would not yield full 1920x1080 resolution, neither do theatrical release prints, like you said.
 
I thought this thread dealt with a wishlist, not a discussion of technical issues.

I'm wondering what about "the biggest thing about bayer pattern as well is that the density of green is twice that to red and blue, so the finer the resolution of the sensor" has to do with a wish list.

best
 
David S. said:
I thought this thread dealt with a wishlist, not a discussion of technical issues.

I'm wondering what about "the biggest thing about bayer pattern as well is that the density of green is twice that to red and blue, so the finer the resolution of the sensor" has to do with a wish list.

best
Click to expand...

I think it can be indirectly inferred that people are hoping the tuning of the sensor for the AF100 is significant for certain things with regards to the performance on the GH1. And that, ultimately, I wish the AF100 read the whole sensor and down sampled instead of pixel-binning to reduce moire and aliasing further. :)
 
My wishlist has gotten short- the more I shoot with my hacked gh1 the more I cant wait for the af100- I just want it in a pretty fedex box on my doorstep december 1st.
 
Status
Not open for further replies.
Back
Top