Decision Time: Sony SR12 or Canon HF10?

[Robin Lobel]

I agree that we'll never agree ;)

One thing anyway: why do you say stills are not good for an A/B ? Videos are just 60 stills per second, and CCI performance scores are 90% based on stills. The colors, resolution, low-light performances are all based on stills processed through Imatest. So I'm not sure frame grabs are useless...

[Dave Rosky]

Originally Posted by Ken Ross

Dave, why not look at the HF10 or HF100? If you're concerned about the hard drive, they should suit your fine.

Ken, Actually, I am planning to look at both of them (the SD and HF). Since my need isn't extremely urgent, I may even wait a month or so to see if a CX9 materializes if I can't get comfortable with one of these two.

Currently, the advantages I see for the SD9 for backcountry use are that the OIS appears to be significantly better (I say "appears" because this information is still mainly just from reviews), and it is a bit lighter and more compact. The 5.1 microphone is a nice icing on the cake, especially for capturing outdoor ambiance.

However, I hate making decisions based just on reviews and downloaded clips, so I am planning to check them out. OIS is high on my list of importance because I don't want to whittle the camera weight down to a pound or so and then have to carry a 7 or 8 pound video tripod into the wilderness to get steady footage. I'd even like to avoid the extra 2 pounds or so of a monopod.

So, I'm staying open minded until I try everything - even though the HF100's OIS is not as good, if it's good enough when I try it out, that may make swing the decision in that direction.

[Ken Ross]

Originally Posted by Robin Lobel

I agree that we'll never agree ;)

One thing anyway: why do you say stills are not good for an A/B ? Videos are just 60 stills per second, and CCI performance scores are 90% based on stills. The colors, resolution, low-light performances are all based on stills processed through Imatest. So I'm not sure frame grabs are useless...

Where do you get the idea that CCI bases their performance ratings 90% on stills? Can you show me that in writing? Since they do not assign a 'weighting' to any category, I'm very curious how you came to this number. Their overall evaluation is a subjective one and takes in to account crazy things like how the actual video looks.

By the way, since your so high on these 'stills', why don't you take a look at some of the highly esteemed camcorders such as the HV30 and those infamous 'stills' depicting resolution. Please take a look at the HV30 resolution test contained within the HF10 review and tell me how 'wonderful' that still looks! You will see that frame laden with noise. Do you really think the actual video from an HF30 has that amount of noise? Of course not. Why? Because these stills are NOT indicative of video quality.

Now, although there are '60 stills per second' as you say, extracting one of them in the same quality as seeing the '60 stills' in motion, are two different things. I will never convince of you that, I understand that and I'm done trying. Robin, just get your HF10 and be happy as I'm sure you will. I'm sure it's a great cam. In the interim I'll actually test both cams and see which really looks best on my own HDTV with honest to goodness video. What a nutty idea huh? :)

You and I wll get nowhere, on that we can agree. You may have the last word, but I'm pretty much done responding to this subject.

[Ken Ross]

Originally Posted by Dave Rosky

However, I hate making decisions based just on reviews and downloaded clips, so I am planning to check them out.

And you are a very smart guy for doing so. Even posted clips that we download could have been from cams that were improperly adjusted. Many of these clips are from video newbies, not the best source to make a determination.

As I've said and you agree, there is nothing like actually testing these cams yourself with real world material. It sure beats 'knowing' which is best without even seeing it! :)

[Robin Lobel]

"feeling" is not something you can evaluate mathematically.
How can you say a captured color is better than another, a resolution better than another, or low-light performance without running mathematical tests on it ?

That's why CCI use Imatest, as stated in the low-light paragraph of HF10:

"The second part of the low light test involves shooting an X-Rite Color Checker chart at an even 60 lux, then exporting frames to Imatest imaging software for analysis on color accuracy, noise, and saturation."

And you can be pretty sure they used Imatest for "video resolution" and "video performance" as well, since they use color/resolution charts from Imatest. How do you think they can make notations like "21.94" ?? (HF10 resolution score). BTW, from the very beginning that's the only thing I said: HF10 seems to have better resolution than SR11. Dot. I never told about color accuracy, and low-light performance, other than the latter looks pretty similar so far. I don't have thoses charts for both HF10 and SR11. But I can judge on video resolution when I see to 2 pictures, with details on A I can't see on B.

Remember, it all started on post #6 when I said "And as far as I know, the pictures from HF10 are slightly sharper than SD12" and you asked (#8) "Robin, how did you arrive at that conclusion?".
So that's how I arrived at that conclusion.

Now, I agree that the ultimate test for a video camera would be to take an average of 5 frames (with a non-moving camcorder) before processing the picture through Imatest: this way it would simulate persistence of vision, which you can evaluate around 1/12sec, and would therefore give a scientific way to estimate noise response, that you could rate with 2 decimals. But it would not change video resolution score.

BTW, another HF10/SR12 comparison has been published on the internet:
http://plusd.itmedia.co.jp/lifestyle...news078_4.html
http://plusd.itmedia.co.jp/lifestyle...news007_3.html
Now I see what you mean with the "haze" feeling on SR12.

[Ken Ross]

Robin, if you can't understand the concept of a subjective evaluation then this simply proves that you and I are miles apart. Robin, evaluating a camera isn't just about numbers, it's about the entire VIDEO that you see on screen. It's about how the image strikes you. Is it natural, does it look processed, etc. etc. Your means of determining which camera is best is cold and impersonal and does not reflect if the 'sum of the parts is greater than the whole'. It does not reflect how the overall picture looks. Yes, certain aspects of camera performance can lend itself to numbers, but the entire picture is not something that can be put into a simple quantifiable score. And you still haven't proved your 90% rationale. I know you 'think' you did, but you didn't.

I am well aware of CCI's 'methodical' procedures for certain aspects of testing, but in the end it's how does the VIDEO look? Do you see any numbers for exposure latitude that indicates a methodical evaluation of this area? Do you see any numbers for noise that reflects a methodical evaluation for this? Do you see any numbers for color accurcy in different lighting conditions (indoor incadescent, flourescent, cloudy day, sunny day etc.) that reflect a methodical evaluation? I could go on Robin, but I think you see my point...then again I seriously doubt you do.

Robin, what also amazes me is that you so conveniently ignore that I've done hands-on testing with the SR12 vs the HV20. I've done many many many A/Bs with both cameras on a real, honest to goodness HDTV! These A/Bs were done side by side, same weather, same focal length. My goodness, what a crazy loony idea...right Robin? You just throw that out. Yes my friend you and I will never agree on how cameras should be evaluated.

Hey kiddo, you stick with your cold numbers and I'll stick with how the video looks on an actual HDTV. Oh and by the way, using my procedure I might actually find out down the road I prefer the HF10...but it will be as the result of watching moving video on my HDTV, not still frame blowups. It will be done in a TRUE A/B fashion, with both cams side by side, at the same time of day, under the same weather conditions at the same focal length.

Just as a final note Robin, your "A/Bs" are not at all true A/Bs. Do you understand how to do a true A/B? These pictures were shot on different days, under very obviously different weather conditions with no attempt by the tester to shoot at the same focal length. The shots using the SR12 WERE shot on an obviously hazy day. And no Robin, that is NOT what I mean by 'hazy' with the issue I've seen on the SR12. If it were, the camera would be gone by now. You are totally unobjective on this subject. Your third "A/B" is almost hysterical. Do you REALLY think those shots show anything as to how these two cams perform relative to each other? About the only thing I can tell from these wonderful A/Bs is that Canon still hasn't solved the 'magenta sky' issue. As much as I love my HV20, I'll never understand why Canon can't produce a consistently blue sky. In some of those shots the magenta issue is really bad. The other thing I note is that in some shots the contrast on the HF10 is really hyped. Hopefully the video doesn't look that way because if so, shadow detail will suffer badly. But whatever floats your boat. Yes you and I will never agree on how to compare cameras.

I think we've beaten this to death and to go on further is just pointless and must be boring to other readers here. I think this is now a waste of Chris's bandwidth. Sorry Chris.

[Tony Parenti]

Those HF10/SR12 shots should not be used for comparison. There's so much more contrast in the HF10 shots.

[Ken Ross]

Originally Posted by Tony Parenti

Those HF10/SR12 shots should not be used for comparison. There's so much more contrast in the HF10 shots.

You're right Tony and I hope that it's not indicative of the HF10's video. Tony what's interesting is, if you look at the last series of "A/B's" (I hate to use that phrase since these are NOT A/Bs), you'll actually see MORE detail in the Sony shot. I believe what's happening here is the Canon shot has so much contrast it's actually obliterating detail. If you look immediately to the left of the red part of the Pagoda, you'll a gray vertical metal piece of the roof that appears to have 'dimples'. On the Canon shot it's very tough to make out any detail, yet with the SR12 you can. It's either due to the extreme contrast or the obviously different time of day it was shot. I also dislike the Canon's color in this shot, but again I doubt you'll see this issue to this extreme in actual video. The third shot is really amusing in that the gold emblem is totally in the shadow in the SR12 and in bright sunlight in the HF10. This is an A/B? C'mon guys.

In my own testing between the SR12 and HV20 I've found shots where I could see a bit more detail with the SR12 and others where I could see a bit more with the HV20. But if anyone thinks that these linked shots reflect the difference you'll see with the actual video, they're very very mistaken.

But again, to profess these are "A/B's" is disingenuous at best. ;)

[Tony Parenti]

What is the point in linking to Japanese comparisons... who the heck here can read Japanese?? It's completely useless.

[Robin Lobel]

I admit you can't make a straightforward A/B comparison on this one, therefore I removed the links. But while being shot under different weather condition at different time, it shows anyhow that HF10 and SR12 don't have the same render. I personally prefer HF10's contrasted cine-look, while SR12 exhibit a more neutral video-look. But I won't compare technical details anymore here as we both know how it will ends...

Tony: the point is to compare how thoses camcorder renders; as said, even under different weather condition you still feel how one and the other goes. Don't you ?
Click the first link, then the second. Common, even different weather conditions don't make such a contrast on every picture. The average of all clearly shows more contrasted and warm colors on HF10 (like it or not).

[Ken Ross]

Robin, not to be a hard head here, but 'contrasted' and 'cine-look' don't mix. A cine-look, as you call it, is a lower contrast look as opposed to the higher contrast of video.

That contrasted look works against the HF10 as it renders LESS detail in one of the shots you posted. However with that said, I've never seen footage from any Canon camera with as much contrast as those stills show. It looks to me like someone was playing some games with those stills. However, it surely does make my point, you can't use these to determine video quality.

[Robin Lobel]

http://www.dv.com/dv/magazine/2006/gamma1.jpg
(source: http://www.dv.com/features/features_...leId=174900673)
can't say it's a like contrast curve (last one) but it's like giving contrast to the middle-range, pushing shadows and highlight without clipping them.

Anyway I agree that it seems a bit overcontrasted here, maybe that guy set HF10's contrast setting a bit high while shoting.

[Ken Ross]

Robin, if ever you've used one of the 'cinema' settings on a video camera, you'd see it gives the overall video a reduced contrast look. Some like it, some don't. Personally for me I prefer the 'you are there', live look of video. Also, your agreement that the contrast looks high and may be the result of a setting by the user, shows what I've been talking about. The only way to really determine which cam suits you best is to do your own A/B on your own HDTV. That way there's no guesswork as to whether someone was setting a camera incorrectly or deliberately...you are in control.

[Tony Parenti]

One thing about the SR11 I didn't like is that there's zero control over picture parameters, ie brightness, contrast, sharpness and color.

[Dave Blackhurst]

I dunno, but the first Canon shot made me want to break into "purple haze"...
there seemed to be a distinct color shift on my monitors, and an unnatural one at that...

Tony - you CAN "translate" japanese pages via Google or Babelfish - the trans-(ob)-literations are always comical, but you can sort of figure some of the stuff out.

There's no way those series of shots could be use for comparison - far too much difference in the times and days -it's tough enough to A/B under controlled conditions (same day, time, comparable settings, identical shots) and get reliable results. These two sets of pictures may as well have been "night and day".

And heaven forbid if the reviewer has a bias... I read certain "reviews" of Sony product and shake my head, as I have the camera in hand and can't replicate the supposed "issues", or worse yet can absolutely refute the claims of the review...

IMO you need to spend more than a few hours with a camera (which is probably more than the average "review" gets) to get to know it's features and quirks and set it up to your preference (and if you like a tiny joystick and hate touchscreen, NOTHING will alter your perception... er, bias). I wouldn't pop a new camera out of the box and leave everything "factory" - first thing I do is scroll through the menus and poke at everything!


My experience with shooting the HC7 and HV20 side by side was much what I expect will be found in A/B with the latest generation - the Canon blacks were darker and more contrasty, which looked good against the somewhat noisier blacks of the Sony, but details were being lost. Reds were VERY problematic on the Canon, tending to blow out and "dance". Colors were more accurate with the Sony, although they could be adjusted with the Canon. BOTH cameras were great, but I went Sony... build quality and sound made the difference for me - the PQ was quite acceptable either way!

I'm seeing similar issues from the above pages, but... until one has an HF in hand, it's all navel gazing! I'll have an SR11 shortly, and looking forward to seeing what it can do. I doubt I'll be disappointed, even if I later discover the HF in some opinions is a bit "better"...