HV20: Cinemode Softness (loss of image detail) Pic

[Nathan Shane]

Hey everyone, I've been growing increasingly frustrated with the lack of sharpness in the image details with the HV20 - using Cinemode all the time, so I thought I'd do a quick check to see if Cinemode was softening the image and boy was I suprised. Check out the comparison pics I posted.

http://www.vettaville.com/canon_hv20...e_softness.htm

[Wes Vasher]

Originally Posted by Nathan Shane

Hey everyone, I've been growing increasingly frustrated with the lack of sharpness in the image details with the HV20 - using Cinemode all the time, so I thought I'd do a quick check to see if Cinemode was softening the image and boy was I suprised. Check out the comparison pics I posted.

http://www.vettaville.com/canon_hv20...e_softness.htm

Easy fix. Don't use Cinemode. The lack of sharpening is part of that mode. I believe you can slightly sharpen it using the filter in camera but it doens't do much.

[Joseph H. Moore]

Nathan,
I spent some time yesterday doing similar tests and reached a very different conclusion. I found no loss of actual detail from cine-mode.

http://www.12south.com/video/Joe_Close-up_24P.jpg

My settings:
24P
CINE
All image adjustments OFF
Image stabilization OFF
manual focus

[David Parks]

I have a question. Is Cinemode the 24p conjoined setting or is it a seperate mode that plays with gamma not completely tied to 23.967 fps? (Don't have my HV20 yet) otherwise can you shoot 24p without Cinemode?

Nathan, are both your examples shot at 24p?? What did you have for shutter settings? This affects sharpness in a big way.

The reason I bring it up is that there will be a sharpness difference between
24 fps, 30 fps, and 60 fps. A 60p camera like the JVC HD 200 shooting 60p will look sharper than 24p. So it makes sense to compare like frame rates.
1080/60i derived isn't the same as 1080/24p to 60i. Apples and oranges.

Anyway we really should be using a resolution chart in order to be more accurate. Shots of people's faces are not in my opinion the best baseline test. I'm getting my HV 20 in a couple of days and I'll do some rez tests next week in addition to lat tests.

Cheers.

[Ian Holb]

Originally Posted by Nathan Shane

...so I thought I'd do a quick check...

There's your problem.

You simply can't come to a definitive conclusion by doing a quick check. It looks like your focus is a bit off in the first picture. Without seeing your test parameters...

And .GIF isn't exactly the best picture format to present your images. There is a strong amount of .GIF dithering affecting your images.

[Joseph H. Moore]

I have a question. Is Cinemode the 24p conjoined setting or is it a seperate mode that plays with gamma not completely tied to 23.967 fps? (Don't have my HV20 yet) otherwise can you shoot 24p without Cinemode?

Cinemode is a settings "program" that is not directly reated to 24P recording "mode." It does a bunch of things, including (but maybe more):
- A flat gamma curve
- Tries to stick to 48th sec shutter
- Tries to open aperture before using gain.

A 60p camera like the JVC HD 200 shooting 60p will look sharper than 24p.

That simply isn't correct. There is more temporal resolution, but the frame rate has no direct affect on the sharpness of any one given frame. (Though it is definitely true that the HV20 takes longer to focus in 24P mode.)

[Nathan Shane]

Just to tell you how I tested...I hooked the HV20 up to the computer via firewire and then started HDVsplit set to show "fullframe" on the pc monitor. Then you just toggle the camera between cinemode and av (or any other setting) and you can clearly see on your monitor how soft the image goes in cinemode, it has nothing to do with the image being out of focus in the first picture - that "out of focus softness" is the direct result of the Cinemode setting and nothing more. It's an easy test and anyone can do themselves. I just used shift+prtscn to capture the desktop and then cropped the images in PhotoImpact.

GIF, JPEG, or PNG would not make any significant difference in showing the softness of cinemode, both images were captured identically and saved with the exact same GIF settings, so all things are still equal in this equation. Yes, there is obviously some dithering taking place in the GIFs I posted, but the overall softness of Cinemode is what I'm showing. If you have another file format you would like me to post, I can do that.

So just to be clear, these are not frame grabs taken from tape, they are live images coming into the pc via firewire and captured.

[Joseph H. Moore]

Nathan,
I believe you're getting what you're getting. ;-) All I can say is I'm definitely NOT getting the same results. I even shot a grid printed from a 1200 DPI laser printer and was able to resolve it tack sharp. I promise. I was monitoring from the HDMI port on a 32" HD monitor.

[David Parks]

Originally Posted by Joseph H. Moore

That simply isn't correct. There is more temporal resolution, but the frame rate has no direct affect on the sharpness of any one given frame. (Though it is definitely true that the HV20 takes longer to focus in 24P mode.)

Joseph,

That's true. I was thinking temporal resolution per one second of time vs. one frame. But, therein lies the danger in printing a single frame. It doesn't tell the near the whole story. Unless you plan on shooting still frames I guess.
I'm just saying that comparing stills without dsc charts is very unscientific and lends itself to inaccurate info.

Thanks for the info on Cinemode.

[Nathan Shane]

Originally Posted by Joseph H. Moore

Nathan,
I believe you're getting what you're getting. ;-) All I can say is I'm definitely NOT getting the same results. I even shot a grid printed from a 1200 DPI laser printer and was able to resolve it tack sharp. I promise. I was monitoring from the HDMI port on a 32" HD monitor.

Joseph, I agree with you that my results and yours (or others) may vary. Not sure if this can be due to slight tolerance differences that may be inherent in each individual camera - but you would like to think that all things are equal between every single HV20, but perhaps not. It may be that when using Cinemode, the lighting requirements must change for a better quality capture...who knows. But when all things are equal in the testing environment and just switching between Cinemode and Av shows such a drastic difference in image details and softness it still shows what can occur with the HV20...just perhaps not every single HV20. Who knows for certain?!?

I'm certainly no expert professional here but I think there is some logical basis for thinking that if Cinemode is a preset that is supposed to help emulate the "film look" alongside the 24p capture, then it sounds reasonable that Cinemode also softens the image alongside all the other adjustments it is making. Now, how "noticably soft" the Cinemode image will look also depends upon how close you are zoomed in on your subject matter. In other words, you may be using Cinemode and have a fantastically sharp image such as Joseph has posted - but unless you can also show the exact same non-Cinemode image as well, you may not be aware that there can be differences.

I think the real answer would need to come from Canon as to whether or not "image softness" is part of what is programmed into the Cinemode preset...(but I'm convinced it certainly softens the image details). But it can still be all subjective, because one man's sharpness may look like another man's blur. LOL!

[Joseph H. Moore]

David, I agree with the caveats about proper testing, but Nathan's results are so dramatic that something is clearly wrong.

Nathan,
Something's just not right on your camera. I don't know if it is a setting, a different firmware, or if it's an actual malfunction.

I did similar A-B testing under fairly low levels of natural light, and although the other P modes added more edge enhancement and contrast, the level of real detail remained the same.

If cinemode really was that soft, there'd be a lot more upset people, myself included!

[Nathan Shane]

Originally Posted by Joseph H. Moore

David, I agree with the caveats about proper testing, but Nathan's results are so dramatic that something is clearly wrong.

Nathan,
Something's just not right on your camera. I don't know if it is a setting, a different firmware, or if it's an actual malfunction.

I did similar A-B testing under fairly low levels of natural light, and although the other P modes added more edge enhancement and contrast, the level of real detail remained the same.

If cinemode really was that soft, there'd be a lot more upset people, myself included!

Can you do a test using HDVSplit like I did and post some image results? I agree that I'm not liking what I'm seeing with "my HV20" and would be keen to have others do a like test using HDVSplit and capturing like I have.

[Nathan Shane]

Okay, I know what I need to do then...another test, but this time I'll pick non-human subject matter so that all subjects can be static and completely motionless. That would be a much more accurate and representative test I think.

[David Parks]

Nathan,

http://en.wikipedia.org/wiki/Image:Siemens_star.svg

Print the Siemens star from above link. This is a very basic chart (used for back focus) but should at least help you out. Please don't use people's faces only.

Let us know what you find. Cheers.

[Joseph H. Moore]

I happened to shoot most of the modes yesterday. Give me a bit and I'll post some grabs.