Guess the ISO

[timmyquest]

http://www.antiwall.com/3200.jpg

I'm finding this 300D hack very nice indeed

[sGu]

since the file name is 3200, i'd go for ISO 3200 ... :roll:

[timmyquest]

since the file name is 3200, i'd go for ISO 3200 ... :roll:

And considering the content of the post is talking about the 300D hack, which adds ISO 3200... :roll: :roll: :roll:

[sGu]

this game isn't as fun as i thought it would be ... :P

[timmyquest]

Doesnt look like ISO 3200 though, at least not ot me. A little in his hair, but the face apears to be...dare i say, noiseless

[sGu]

that is pretty amazing at ISO 3200, then i guess the shutter speed should be fairly quick, how fast was it?

[Jesper]

When you resize a photo to a size for the web, the noise becomes much less visible. I bet that in a 100% crop (1 pixel on screen = 1 pixel in the original image) the noise is a lot more visible, especially in the darker areas.

[maderito]

Timmy,

I always have to watch out for your efforts to prove the impossible.
Check this article (http://blanik.colorado.edu/~rtezaur/photo/other/raw/) and graph (http://blanik.colorado.edu/~rtezaur/photo/other/raw/sn.jpg) for a technical explanation of where signal to noise ratios overlap across different ISO's:

Here's what I see in your image after opening up the shadows. Nice try. :) :)

http://display.lifepics.com/imgdisp.asp?filespec=%60foxhx2cuxmwdogx%5D9%3C%3D8 %3B9dOjeOrmlfqwkxf5svk%13385%3C89%1264%3A4%3A2

[drisley]

If you bump up the gamma on almost any ISO like that you are going to get posturization in the shadows.
A properly exposed ISO1600 shot is virtually noise free. I would expect very good results from ISO3200 if properly exposed too.

[maderito]

If you bump up the gamma on almost any ISO like that you are going to get posturization in the shadows.
A properly exposed ISO1600 shot is virtually noise free. I would expect very good results from ISO3200 if properly exposed too.

I think this is a question of definition. At all ISO's, the noise in the shadows is higher than in the highlights. It may be less noticeable in the shadows - as in Timmy's original posting - since the dark noise merges more readily with dark signal. But it is still there and thus makes it very difficult to "open the shadows." I certainly agree that a well-exposed image confines the noise to regions where it is less noticeable visually and thus less disturbing.

I would not say that a well exposed ISO 1600 on the 10D is virtually noise free. Depending on the degree of magnification, it very well may have visually acceptable noise.

The graph (http://blanik.colorado.edu/~rtezaur/photo/other/raw/sn.jpg) I mentioned earlier is based on 10D RAW data and clearly shows that the S/N ratio at ISO 1600 approaches 1 in the shadow region (log of exposure = -2, where 0 equals the metered exposure).

[Conk]

since the file name is 3200, i'd go for ISO 3200 ... :roll:

And considering the content of the post is talking about the 300D hack, which adds ISO 3200... :roll: :roll: :roll:

That's how I would have guessed it. :lol:

[timmyquest]

LOL, you guys are....i'll keep it to myself.

This is ISO 3200...yes there is going to be noise. It's the amount of noise that is suprising.

You adjusted the levels to maximize the noise...thats nice, would you ever do that to a final photo?

Didnt think so.

www.antiwall.com/fullsize.jpg <-as the filename suggests

[drisley]

Yup, very impressive!

I took a lot of ISO1600 shots this weekend, and not only is the noise acceptable, it's basically not visible even at 100% magnification. And it's most definately not visible on prints I've made, even up to 8x10.
Anyone that complains that these new Canon DSLR's are noisy at high ISO's have been spoiled. Just try shooting high ISO film and getting anything close :)

ISO1600, NO Noise Reduction, 100% Crop
http://www.mts.net/~lftbrain/iso1600crop.jpg

[Jim_T]

I find I see noise in images taken at ISO 1600 when I view them at 100% on my monitor.. Of course there is more at ISO 3200

If I resample them for web use and shrink them from 3072x2048 down to 1024x 683, they are as clear as a bell.. I don't have to do any additional 'anti noise' processing.

As mentioned, they also look great when printed up to 8x10.. (I've never printed larger so I don't know how they'd look.

Oh yes.. I find any sharpening at.. Any method.. really brings out the grain when you get over ISO 400. That's the only thing I find annoying about high ISO's

[drisley]

You must be doing something wrong.
I get NO grain at all at ISO800 when sharpened, and only SLIGHT grain at ISO1600. ISO400 is smooth as silk. But,when I shoot I make sure I properly expose the images, or slightly overexpose, and I always shoot RAW.

The picture posted above is a sharpened ISO1600 shot at 100% crop. I dont see any noise.

Here is a double sharpened (once in C1 at conversion and once again in PS) ISO800 shot at 100% crop. Do you see any noise?
http://www.mts.net/~lftbrain/iso800crop.jpg

Here is the full image:
http://www.mts.net/~lftbrain/rebel/share/fitness01.jpg

[Mark Kemp]

Do you see any noise?


Nobody would dare make any noise in case she got upset :lol:

[c0ntr0lz]

just one sec
i go to set mine to 3200 and it says H, I'm guessing for High
when i take a shot it's the nastiest thing i've ever seen

[maderito]

First, Drisley's shots are wonderful. I've seen and admired them before.

To suggest that ISO 1600 is virtually noise free is simply misleading. Under some conditions, you can get great shots with minimally visible noise. I've taken hundreds of shots under such conditions (see these examples (http://www.pbase.com/maderito/multicultural_talent_show)). Noise will be minimized when shots are perfectly exposed, as already noted. Also, high contrast shots like the one's shown by Drisley and in my examples also yield good results since the shadows are dark and there is a lot of tonal variation that is pleasing to the eye.

But the noise is always there, and under less than optimal conditions it will be unacceptable. Increasing the gamma in an image with levels or curves "unmasks" the noise. This is not an artifact - you are simply making visible real image data that was visually merged in the shadow areas. If your shot doesn't require a significant gamma adjustment - great! No one will see the noise (as in Timmy's shot).

You also lose exposure latitude at high ISO (maybe by a stop or so). This makes shooting and post processing even more demanding. Drisley is to be congratulated for overcoming the odds. But anyone who has spent a full day of shooting at exclusively ISO 1600 (whether on purpose or by mistake :( ) can testify to the many difficulties encountered. Thus we all shoot at low ISO's whenever possible.

[drisley]

Thanks for the compliment Maderito. :)

Yes, I have about 100 shots from that show at ISO1600, and the noise is just not noticable. And that was in a HOT theatre, and heat supposedly increases noise. I do agree that high contrast images produce lower noise. But, a few weeks ago I shot some diving pictures at ISO1600, with lower contrast, and again the noise was minimal.
Maderito, do you shoot Raw, and use C1 Pro for conversion?

I do find that C1 Pro, when set at one notch BELOW it's default noise suppression, gives amazing results at high ISO's.
Perhaps this is why I get such good results. I think I've tried Breeze and PS CS, and found they had alot more noise. I'm sure jpg would too.

All I know is that to me, I have no hesitation using any ISO on this camera. Bravo Canon.


Nobody would dare make any noise in case she got upset :lol:
:lol: :lol:

[Johnnynf]

Timmy and drisley,

If I may ask, were any of these images run through any sort of noise reduction software such as Noise Ninja or Neat Image?

[timmyquest]

Timmy and drisley,

If I may ask, were any of these images run through any sort of noise reduction software such as Noise Ninja or Neat Image?

I adjusted the levels, that is it.

[drisley]

Timmy and drisley,

If I may ask, were any of these images run through any sort of noise reduction software such as Noise Ninja or Neat Image?

Nope, none at all. Just processed RAW straight thru C1 Pro at default settings.
As a matter of fact, I actually LOWERED the default noise suppression settings of C1 Pro (as mentioned above). I find the noise so minimal, I see no need to use the higher default. 8)

[maderito]

Thanks for the compliment Maderito. :)
Maderito, do you shoot Raw, and use C1 Pro for conversion?

I do find that C1 Pro, when set at one notch BELOW it's default noise suppression, gives amazing results at high ISO's.
Perhaps this is why I get such good results. I think I've tried Breeze and PS CS, and found they had alot more noise. I'm sure jpg would too.


I shoot mostly RAW and use Adobe Camera Raw with no noise reduction. When necessary, I use Neat Image (highly recommended!) for to clean up noise.

C1 has a reputation for excellent noise reduction algorithms. Perhaps that explains, in part, your good results.

[Laziferous]

Honestly, from results I've seen posted throughout this forum, it seems to me, that the 300D seems to be better at keeping a handle on noise than the 10D.

*pouts in corner*

[drisley]

Just for the heck of it, I tried that same ISO1600 image processed with C1 with noise suppression off completely.
I dont really see any difference. It's still looks noise free like the picture above.
I do know that Adobe Camera Raw is infamous for much noisier conversions than most other Raw processors. It also leaves a fine "herringbone" pattern on most images that is visible at magnification (so most of the time it isnt noticable).

Now if C1 would fix the red dithering problem on some coversions, it would be perfect.

[drisley]

I just did a test.

I find that C1 Pro images at ISO800 have less noise than Adobe Raw images at ISO400. And that is with C1 Pro noise suppression off (all the way to the left), and Adobe Raw noise reduction enabled!
Detail in both images was similar. So if you shoot high ISO images often, I would definately recommend C1 (well, I would anyway) :wink:

[maderito]

I was tempted to let this rest, but I'll take one more shot - and that's it.

I make the following prediction:

In 6 months you will:

1. Take more shots at ISO less than 1600 than at 1600 or higher

2. Read no reviews of the DRebel, the 10D or their successors claiming that ISO 1600 is virtually noise free.

3. Encounter Phil Askey's reviews of the 10D (http://www.dpreview.com/reviews/canoneos10d/page18.asp) and/or the 1D Mark II (http://www.dpreview.com/reviews/canoneos1dmkii/page18.asp) (which has a similar noise profile) and see that his image crops and noise vs. ISO graphs show levels of luminance and color noise at ISO 1600 that is impossible to ignore.

4. Revisit this thread and worry that you may have overstated your case.

5. Remember and thank me for recommending the 85mm f/1.8 in this thread (http://photography-on-the.net/forum/showthread.php?t=27097&highlight=85mm+bodybuilding ).

6. Capture a lot more great images.
If I'm wrong, I owe you my very own Photoshop Action - "Adding Noise the Hard Way Using Adobe Camera Raw".
:? :) :)

[drisley]

Hey, all I can say is look at the ISO800/1600 pictures posted above.
I have lots more like that. If those are "noisy" images, then I have no problem with it at all.
All I go by is what my pictures look like. I would never even use ISO100 on my G3 before because the noise bothered me.

BTW, yes, thankyou for the 85mm f1.8 recommendation.
On that note, I have a question about that I will put in another thread.
:wink:

I think this thread should be renamed "The Iso Man Cometh" :P

[Tom W]

Honestly, from results I've seen posted throughout this forum, it seems to me, that the 300D seems to be better at keeping a handle on noise than the 10D.

*pouts in corner*

I think that's just due to higher expectations with the 10D than the Rebel. They use the same sensor and a good deal of the same electronics - if someone were to do side-by-side tests with a few samples of each, I suspect that the average results would show equal noise at any given ISO setting with the two camara types.

Also, don't pout - there's no crying in photography! ;)

[neil_r]

Tim, I know you started a tread once called "ISO 3200 is crap". Seems you have changed your mind. :D

N

[Andy_T]

I got another one ...

http://gallery.photo.net/photo/1551783-lg.jpg

so, any guesses as to the ISO?

Best regards,
Andy

[Laziferous]

Also, don't pout - there's no crying in photography! ;)
I wasn't crying. I was pouting. That's only pre-crying :wink: :lol:

[maderito]

Looks like a low compression JPEG file masquerading as an high ISO image.

The pretty red head looks like she has a case of measels!

The exposure isn't that bad considering the backlighting. No flash.

I'll guess ISO 400.

[timmyquest]

Tim, I know you started a tread once called "ISO 3200 is crap". Seems you have changed your mind. :D

N

It was actually 1600, and the reason for the noise was because they were underexposed, which seems to be the largest cause for noise.

[Tom W]

Tim, I know you started a tread once called "ISO 3200 is crap". Seems you have changed your mind. :D

N

It was actually 1600, and the reason for the noise was because they were underexposed, which seems to be the largest cause for noise.

Yes, underexposed and shadow areas will exhibit noise more readily than well-exposed or black areas (though the black sky will eventually fill up with noise as well given enough exposure time) at higher ISO settings.

Drisley is right to expose to the right on his indoor non-flash images. The benefit of low noise seems to outweigh a slight tendency to blow out the occasional highlight.

This would be an area where you might want to experiment a bit to find the right balance of high exposure vs. noise.

BTW, use fill flash next time you take a shot of a lovely young lady with a strong background light - I hate to see such beauty masked by digital noise like that. ;)

[Andy_T]

Looks like a low compression JPEG file masquerading as an high ISO image.

The pretty red head looks like she has a case of measels!

The exposure isn't that bad considering the backlighting. No flash.

I'll guess ISO 400.

Maderito,

actually, that's not a low compression file masquerading as high ISO, but rather it is a crop from a larger picture.
As you observed, it was shot in a dark room with a lighted wndow in the back and handheld (I was doing a presentation and using the introduction to snap some pictures)

Also, it was the best that the 'remove ISO noise' setting of BreezeBrowser could come up with. I wanted to preserve the shot because I think the pose of the redhead is nice, but I couldn't get a better picture.

All that said, you're still right... it's ISO 400 on a G2.

Best regards,
Andy

[maderito]

Looks like a low compression JPEG file masquerading as an high ISO image.

The pretty red head looks like she has a case of measels!

The exposure isn't that bad considering the backlighting. No flash.

I'll guess ISO 400.

Maderito,

actually, that's not a low compression file masquerading as high ISO, but rather it is a crop from a larger picture.
As you observed, it was shot in a dark room with a lighted wndow in the back and handheld (I was doing a presentation and using the introduction to snap some pictures)

Also, it was the best that the 'remove ISO noise' setting of BreezeBrowser could come up with. I wanted to preserve the shot because I think the pose of the redhead is nice, but I couldn't get a better picture.

All that said, you're still right... it's ISO 400 on a G2.

Best regards,
Andy
Thanks Andy. It's been a couple of days since you posted your question...and I was losing some serious sleep. :twisted:

Looks like I was 3 for 4 (pretty redhead, backlit, ISO 400). And I get a bonus point for being the only forum member to post an answer. :D

I actually tried to reproduce your image by saving one of my JPEGs with *zero quality* compression in Photoshop. (I hadn't even realized you could do that until now. ) Didn't look too bad. :shock:

So I saved it again - same quality setting. Finally, the results started looking pretty awful but didn't duplicate what I saw in your image. This testing took place after I had already posted my answer. So I was prepared for the big let down when you finally made your way back to this thread.

BTW - your image had minimal noise. The winner is Breezbrowser. :)

[Andy_T]

Thanks Andy. It's been a couple of days since you posted your question...and I was losing some serious sleep. :twisted:


I am truly sorry.
I had not foreseen that last week I actually had to do some work at the office, so I couldn't spend my time posting at the Canon Digital Photography forum, as usual :wink:


Looks like I was 3 for 4 (pretty redhead, backlit, ISO 400). And I get a bonus point for being the only forum member to post an answer. :D


You have definitely earned that point ... :wink:

Best regards,
Andy

[angst911]

c1 pro? what is that?

[maderito]

c1 pro? what is that?
Capture 1 Pro (http://www.c1dslr.com/Products/C1_PRO_new.aspx), Phase One's RAW conversion software application for Canon and other digital cameras.

C1 Pro is not yet a household word but ...

[angst911]

ok... interesting product... 599 bucks though.... how do the SE and LE versions compare?

[maderito]

ok... interesting product... 599 bucks though.... how do the SE and LE versions compare?
I'm not a C1 user. I believe most use the C1 SE version which costs $249 - or $350 less than the Pro version - depending on how you look at it. :)

Hopefully someone else is still following this thread and can answer your actual question - SE vs. LE?

[Stapler123]

I Have been wondering the same thing, I think I can swing the LE version, but not the pro, They have free trials on their website, so I'm gonna get my download on

[dn7elson]

ok... interesting product... 599 bucks though.... how do the SE and LE versions compare?

http://www.rawworkflow.com/CaptureOne/Marketing/C1V35_comparison.pdf

[theoldmoose]

They are about to release a big bug fix package, v3.5.1. The 3.5 release was rushed out under deadline, and contains a number of very annoying (and in some cases just downright disastrous) bugs.

I'm still holding at v1.3.1, until they get the bugs fixed. Unfortunately, when they posted v3.5, they removed the download links for the earlier release. So, folks that don't have a copy yet, can only evaluate the 3.5 version.

[dn7elson]

I'm still holding at v1.3.1, until they get the bugs fixed.

I'm in the same situation. I've downloaded the 3.5a version, just to have it, but haven't bothered to install it. V1.3.1 is just fine for now until the bugs are out of 3.5x.

[samdring]

I believe most use the C1 SE version which costs $249

Wow - something one can buy cheaper in UK than US. Is this a first?