First Round FX vs. A1
 

So far..not impressed with the Canon and it looks as though it is no different from their SD cameras with unacceptable noise in low light.

Canon at 0db everything is else in auto:

http://homepage.mac.com/epiphany2002/0dbcanon.wmv

Sony at 0db everything in auto:

http://homepage.mac.com/epiphany2002/0dbsony.wmv

You should have put them on similiar settings, not auto.

The auto mode obviously gained up the canon more than the sony. The Sony is noticable less grainy, because the gain isn't as high.

At 0db, it sure did exhibit a fair amount of noise, but it looks that it was significantly lighter as well.

I bet that if you matched them as far as exposure is concerned they would be more similiar.

on my machine the canon is a lot brighter. But it would be better to test without the zooming.
I think Holly may be right about auto mode (on the canon anyway) messing with gain even if you leave it set at 0.

Also, are those Standard def files? Shot in SD or HD? The grai will be much worse in SD, and also, ify ou down converted SD to HD, the grain will intensify.

I dont see that this is an accurate test, not according to the Night clip shot in the taxi on 0db gain...

^^ most likely shot in HD downrezzed to SD.

to have that much grain in the canon video, the gain must be up, that is a TON of noise for 0db of gain.

Umm..the gain was set manually at 0db for both and everything else was left in auto so the settings were the same.

In fairness, the Sony was done slightly after so at dusk..it probably had gotten slightly darker even though i twas only minutes after the Canon test.

it was not downrezzed and both were in 0 db set manually

Yeah, I also agree with Holly. If you use Auto mode, cameras from different manufacturers will use different algorithms to change shutter, aperture, AND gain according to their engineering team's preferred solution, which may have a different "sweet spot" than another manufacturer's. I don't know about the Sony, but "Green Box" Easy Mode on Canon cameras ignore all manual settings; the actual "A" or Auto mode does allow manual gain control, but shutter and aperture would still be auto, so you don't really know what exposure it is trying to get to or exactly how. Maybe Green Box vs Auto is where the confusion lies, because it sure looks like the Canon was very much gained up.

It looks like the Canon achieved a much brighter image than the cleaner, but very much darker Sony clip. Since the final results differ by several stops in their final exposure, the Auto mode on the two cameras obviously arrived at such different endpoints that unfortunately it is pretty meaningless to compare these clips. For such a test to have validity, you first have to either set each camera manually to default/neutral settings, or optimized manual settings.

For those that may seem a little worried, don't be, these tests are far from a standardized test. It is interesting that anyone would even take them time to do a test and put it on auto. Who uses auto anyway, not to mention for an objective test?

I am utterly convinced that gain was used. I have never seen grainer footage from a camcorder that is fully open with 0db.

Is there anyway for you to do another test with lowlight, and manually open the iris and turn the gain off. Screw auto, it doesn't do the cam justice.

For a lowlight test, I suggest Kaku's footage...

This was not a test:

Both in auto
Non Controlled lighting??! (are you kidding me?! it's a light test!)
HD camera used as SD?
Motion/Zooming shouldn't be necessary and only distracts from test


If it's important to you, I'd suggest re-doing the test making sure to keep everything locked down, cameras, gain, and lighting.

So far I'm extremely pleased with mine, seems to have great night sensitivity.

We all appreciate your work and intentions, but tests need to be controlled to carry weight on these boards, otherwise people just bicker.

Not to threadjack...

Brent, how would say objectively the light light performance is?

a few minutes at dusk can be huge difference...

So far... not impressed with your testing parameters. Try shooting at the same time, same framing, same settings. Please avoid auto mode.

From an amateur that would of course use auto most of the time to start with...this would matter to me..my Sony VX2100 is better as far as noise in auto...
BUT, if the the gain was also set to auto would that make it a better picture, less noise.....full auto?

I have yet to learn about gain, so if I get it I would need auto for a bit.

Also was this SD or HD footage? thanks
Jerry

PS, I thought I had learned in digital SLR's that the more pixels(they are smaller then), that you crammed on a sensor, then you can expect noisy pics, could that be the case here?

HERE'S NOISE WITH GAIN FULL UP

Here's a framegrab from a Canon XHA1 with AGC on, wide open at 1/15th sec on a dark street. I don't know what db level that gain is. If you turn agc off and put it on H gain, it's 12db and looks the same -- which is brighter than the Sony at 18db gain. I also have a frame grab of a Sony Z1U at 18db gain, 1/15th second wide open. The Canon is brighter but they're both noisy.

These are still frames out of fcp, de-interlaced and resized to the correct aspect in photoshop, and uploaded as "7" quality jpegs.

chuck

Thanks for the pics.

Still grainy, but the light is very low. It looks to be brighter grainier, but with more color than the Sony. All in all I am impressed. The Sony does very well for cramming that many pixels into a 1/3 ccd, and from the looks of it the Canon is right on it's heels.

I am interested in what DB the Canon was at. I know it was on AGC, but typically don't Canon's stop the Auto gain at 12db or 15db and only 18db with user input? I could be wrong but that is what I remember with my Gl2 years ago.

Thanks

The Kaku clips of nighttime streets didnt have anywhere near this much noise. Whats up? It really looks like gain has to up somewhere.

The picture on the left looks overexposed for the conditions; the one on the right is closer, probably. But from the lights it should be pretty dark. It looks to me as if you're trying to bring the darkness up too high with both cameras.

I don't have this camera, but the manual is now available from Canon USA.

http://www.usa.canon.com/consumer/co...&modelid=14061

Maybe people are forgetting to turn off the AGC switch? According to page 60 and 61 of the manual, when in Auto (A) mode, you can control the gain with the H-M-L gain switch but the AGC switch must be off.

I remember a similar uproar when Mikko posted noisy footage from IBC on another site but admittedly he was not familiar with the camera. Kaku's taxi footage definitely does not exhibit this kind of noise. It's pretty clear the camera is not at zero dB in Daniel's footage, in my opinion. Perhaps this is another case of operator unfamiliarity. The discrepancy between Daniel's footage and Kaku's footage is enough to be cautious about making any conclusions about the footage until we get more information.

This all looks/sounds consistent with what we've seen in the past comparing the Sony FX1/Z1U to the Canon XL-H1. The Canons appear to be more prone to graininess unless you're very careful about gain, while the Sony footage looks darker in poor lighting. Neither is ideal.

One thing is a low light test. Another thing is a NO light test. I don't thing anybody can be seriously expecting ANY camera to shoot with that little light without grain. Anyway, I think I read somewhere that there's a grain reduction button somewhere on the canon for extreme situations...don't know how much it will help though.

were all these test shot in the SD mode?

The specs say .4 lus in manual mode with 18db gain.

Wouldn't shooting in auto mode in the low light automatically at the 18db and set the shutter and stuff so that the noise would be minimal?

If not and the SD footage is that noise, is that indicative that the HD footage would be just as noisy?

I am disheartened to say the least. Had hoped for something that would beat the vx 2100 in low light and still have great HD.

Thanks
Jerry.

hey, thanks for posting clips..

..however, this 'comparison' proves nothing about picture quality. all it tells us is that the canon will boost its gain more in auto mode. if you kept the settings the same as they are in these clips but lowered the gain down on the canon, you'd have a picture that more closely matched the sony (perhaps more grain or less, perhaps better, perhaps worse.. dunno, can't tell from the video.).

I think the big question here is if these are SD, HD, or HD downconverted to SD, and if downconverted, how?

A downconversion usually adds grain, and I would assume that the HD mode will not suffer as much from Grain, because there is more pixel information to combat it.

Ok, we've all been rather hard on Daniel's quick look. Let's let that go and await further testing using manual mode and specified settings.

I agree with Pete, and Matthew.

I was not bashing his tests at all. Quite contrer, I was trying to ask what Matthew was askiing and what would be the difference if the gain was in the auto mode too....what difference would that really make if any? I am trying to learn also.

I actually appreciate Daniel's quick test, because that is the conditions I shoot in, only everything would be in the auto mode and was just wondering since he said the gain was manually set to 0, what would it look like set to auto.

As for the clips posted earlier by Kaku, would be nice to know what mode and settings he used for his to compare with.
Jerry

posting a so-called test on "auto" isn't a test at all, unless you also post a similar test using the full manual controls which push the limits of each camera. it's the only fair comparison.

with the degree of image control available in either of these cameras, the only reason to shoot on auto is to either set the levels and work from them as a baseline or because the action is happening so fast that it is the only way to capture the event before it ends. those are very specific conditions. why pay up for all of this image control and then neglect to use it? i don't get it. if you want to shoot out-of-the-box and not have to actually work the extensive controls to improve the image, there are undoubtedly better cameras out there. that's not a newsflash....

I understand Meryem.
But as i am amateur, just like with my DSLR's, I shoot auto out of the box for a time until I have time to learn the camera and what it does and learn about the different settings and how they affect the image...I shot auto with the DSLR for 6 mos, on things that I definitely did not want to miss or screw up, and in between worked with the different things until I knew what Tv and Av were and how to use them and so forth. There is nothing wrong with using a camera or corder in the auto mode until you are comfortable with it and have learned all the ins and outs. It has taken me 4 years to understand and fully appreciate and understand almost everything about the DSLR's, but in the meantime I shot auto on shots that were there and I wanted to make sure they were right and not take a chance on working on trying to set image settings that I had not become familiar enough with. Same with the camcorder. I am new at it totally from a manual settings viewpoint, so until I get it down I will use auto mode on those need to get it right scenes. And I don't think there is anything wrong with that or admitting that I don't know anything yet about manual settings on a camcorder. I never oncetried manual settings on my Vx in 3 years, I am ashamed to say as I spent more time on the DSLR learning it. I am no pro. But I was wanting to see how it did in full auto, because I will be using full auto while I learn the ins and outs of adjusting the image settings and so forth.
Jerry

Thanks for posting this Chuck ... did you try the "black stretch" on the Z1? I have heard that this improves the lowlight performance.

The point has been beaten to death here but if you are not a skilled operator, then you should probably look at another camera. The Canon cameras offer a wide range of looks, adjustments, etc. that are almost unparalled in the class. They are not point and shoot cameras. Sony does a better job in their auto modes but there are far less options in the manual modes making it MUCH easier for their auto modes to adjust. On Sony's VERY high end cameras like the F900, the image out of the "box" is very flat, milky, with pastel colors.

Like the XLH and XL2, there are an abundance of adjustments to help with low light, gain, etc. Coring, NR, setup level, master pedestal, etc. etc. all can help mask noise and help with low light. The auto mode will not engage any of these.

Bottom line, the Canon is capable of images the Sony simply cannot re-create, the reverse is not true...



ash =o)

Jerry,

No problem, everybody has to start somewhere. Welcome to the hobby, I can tell you this this is an awesome hobby ! It is relatively easy to comprehend but impossible to master (as film is art). Before long, you would want a stabilizer, a mixer, recorder, matte boxes, 35mm adapter ...etc.

Welcome aboard ! let us know if we can help in any way.

Daniel, I checked out rest of your gallery (only saw the demo last time). Your work is unbelievable imo and gives videographers who are getting into the business a goal to work for. BRAVO!

Thanks Michael. Those edits on my site are mostly done onsite the day off and shown at the reception. (ie Same Day Edits or SDEs)

jerry, i understand where you are coming from...but i simply wouldn't call these comparisons "tests" because that's misleading.

i think it is fine to post "auto" footage of a camera, i did it myself when i bought an FX-1 and it arrived in the mail minutes before i took it on a trip into the desert, where i had no time to familiarize myself with the camera. but i just wouldn't call it a test, that's my main beef. "auto" really only shows one of the things that *can* be done with a camera, not what *can't* be done....i wouldn't draw any conclusions about the camera's quality on that basis, which is the initial grounds of the discussion...

Thanks Jason, much appreciated...I started with the Optura 100, then the vx2100 but never took time to learn the manual aspects..Now that i have the still camera down mostly I can spend time on video manual operations, I hope and hope they correlate in many ways so that the learning curve is not so steep.

Meryem, thanks again....i can understand what you are saying. My biggest question for the original poster was he said that all was auto but the gain was manually set to 0db. and my curiousity was what would it look like with it all, the gain also in auto, if that makes sense.

I meant to mention also, something i did in another post, coming from the stil digital camera side, there is a highly regarded program Neat Image that helps decrease the noise in images. I think that they now have a video version of that software, they have been around awhile and you can even make profiles for diff cameras and iso's and situations, so would imagine the video version would do the same.
I think they have a trial version you can try.
It would be nice to see these clips run thru a noise decreaser program like neat image and see what it does to it...it is new I think so i havent even tried it for my vx footage. but if anyone wants to give it a shot www.neatimage.com I haven't read a lot about the video version so don't know if it is limited to sd or both hd and sd.
Jerry

Is it possible or has anyone compared Katu's settings and so forth to Daniel's settings on his clips. Maybe before condemning Daniel, we could do that and find a clue there of why the difference. Maybe Katu was not auto but in the manual mode? just a thought
jerry

Kaku posted his settings, go back and reread.

My recollection is that most or all of his night time clips had zero gain.

Nobody is condemning anybody.

Daniel's clips were done in Auto mode, and that's why many folks have simply pointed out that that introduces many unknowns into the equation. That does put significant bounds on what judgments you can reasonably make about the camera's low light performance.

Here's the link to the post on Kaku's night highway shots, with settings:
http://www.dvinfo.net/conf/showpost....29&postcount=1

All were at 0 Gain, factory default on the custom presets, and specified shutter, aperture, and frame rate.