Canon 7D Noise?

C

Cpinch73

Suspended / Banned
Messages
90
Name
Chris
Edit My Images
Yes
Yesterday I was down my local rugby club practicing some action shots. The lens I was using was a 70-200 F4 L IS. I know the day was a little overcast but once I got home and looked at the pictures on my pc screen I was really disappointed with how noisy my pictures looked!
I was shooting at 1/1000s to freeze the action. In all I'm not very pleased with how my 7D handles noise. Please can someone tell me if I'm doing something wrong or are any of you disappointed with 7D's noise handling?

Many thanks
Chris
 
You must expose properly, even better expose to the right. Underexposure is a disaster for noise on any camera. The 7D is certainly not worse than other APS-C cameras in this regard, unless you stare at pixels instead of pictures. Please post an unedited sample image with EXIF intact. That might allow a better diagnosis and suggestions going forward.

FWIW here are a couple of sports shot from my 7D at 6400 ISO. Due to miserable lighting I could only get to 1/250 at f/2.8, but because the exposure was correct I think the result looks pretty clean....

20100115_200707_2092_LR.jpg
20100115_210037_2333_LR.jpg
 
Last edited:
I agree I think yr pictures look very good under the circumstances.

I've already edited my pictures with lightroom and reduced the noise with a good outcome.

Can you tell me, did you have your high ISO noise reduction turned on?

I appreciate yr help in this matter.

+ plus what do you mean expose correctly? I was shooting at 1/1000, F4, auto ISO?
 
Last edited:
I also tested one up to 6400 and I was quite impressed with the low noise, especially with the high pixel density and medium sensor.
Are you viewing the images in photoshop at 100% or something else?
 
I was viewing them in Adobe lightroom 3 zoomed in. The players were in black shorts where it showed the most.
 
Cpinch73 said:
I agree I think yr pictures look very good under the circumstances.

I've already edited my pictures with lightroom and reduced the noise with a good outcome.

Can you tell me, did you have your high ISO noise reduction turned on?

I appreciate yr help in this matter.

+ plus what do you mean expose correctly? I was shooting at 1/1000, F4, auto ISO?
Click to expand...

I shoot raw so most in camera processing has no impact on the files I process. FWIW I turn off as much of that in camera trickery as I can, including High ISO NR. I don't want the camera wasting its CPU cycles fussing over stuff I don't want it to do. Even if the raw file is unaffected the camera will still apply the processing to the embedded JPEG preview image.

By "expose correctly" (or to the right) I mean expose well enough that you do not have to brighten the image at all when editing. In my opinion using auto ISO means you have let go of control over the exposure and the camera will be making decisions for you, and quite possibly making them poorly. For example, ISOs 125, 250, 500, 1000, 2000, 4000 are best avoided if you want to minimise noise, but by shooting with auto ISO the camera may well choose them. If you have HTP enabled then that can be a real disaster for noise. I don't ever use it. ALO is also a bit iffy and can contribute a little to noise too. Both HTP and ALO will affect your image and histogram and may lead you into a false sense of security regarding the exposure.

Also, depending on your metering setup, but possibly regardless of that, if you are shooting in tight on players wearing bright/white strips then that can throw off metering and result in underexposure unless you compensate. I prefer to shoot with manual settings in order to avoid such irritations.
 
Last edited:
You were using auto ISO, that's the problem. Check the ISO on the shots you find grainy and let us know what it set itself to.

I wouldn't run auto ISO if I was you, especially if you have ISO 12800 enabled.
 
danmon_81 said:
I have issues with colour balance in a lot of my 7D photos...its doing my head in. I've played with the WB but just cant get it right...Noise seems fine.
Click to expand...

Under what sort of lighting (please don't say fluorescent) and with what WB settings in camera? Do you shoot raw or JPEG? What, if anything, do you do to try to fix WB in your editing software? Do you perform a custom white balance before shooting? Do you include a neutral white/grey sample within your scene from which you can set WB using a colour sampler/dropper?

If you do say "fluorescent" then at what sort of shutter speeds? Please don't say faster than 1/50 (or 1/100 at a pinch).
 
Cheers tdodd
I'll turn them bad boys off soon as
Just got my 7d and getting to grips with it
 
Unfortunately I can't put the originals on here as I've already edited them before joining Talk photography, but there's been some good advice here so I'll take it all on board.

Should I set my ISO to go up in 1:3 or 1:1 intervals or will this not matter where noise is concerned?
 
Cpinch73 said:
Unfortunately I can't put the originals on here as I've already edited them before joining Talk photography, but there's been some good advice here so I'll take it all on board.

Should I set my ISO to go up in 1:3 or 1:1 intervals or will this not matter where noise is concerned?
Click to expand...

But editing in Lightroom does not destroy or alter the original file, so you should be able to return to the original if you wish.

As for ISO, I only use the full stop ISOs - 100, 200, 400 etc.. It's really personal preference, and that's mine, but the reason I avoid the intermediate ISO settings is because they aren't real ISOs. The camera fakes the data by internally underexposing and brightening or overexposing and dimming the images. I don't want any of that going on. I'd rather reserve that privilege for myself. It gives more flexibility in the results I can obtain. But that is me being pretty anal about squeezing the max performance I can from my gear. For other people, especially those who shoot to JPEG, these things may be a convenience they wish to embrace.
 
Tdodd does that mean you have yr camera iso's set at 1:1 or 1:3? I wasn't sure by yr answers.

Sorry to keep asking lol
 
Mine are set to 1:1 (if that's the right way to express it:thinking:) and I always select ISO manually. But like I say, that what works for me. Others may prefer to do it the other way.
 
Thanks tdodd. I have read that yr iso settings give the best results.
 
This was a shot at 3200, ok on grass but not a rugby player :lol:

You certainly don't want to underexpose if you want to avoid noise at high ISO's

IMG_4016.jpg
 
Last edited:
Artyman that's a really crisp picture!

Bet he can play better than my club did the day I was taking pics :lol:
 
Convert raws in DPP rather than Photoshop / ACR
 
tdodd said:
I shoot raw so most in camera processing has no impact on the files I process. FWIW I turn off as much of that in camera trickery as I can, including High ISO NR. I don't want the camera wasting its CPU cycles fussing over stuff I don't want it to do. Even if the raw file is unaffected the camera will still apply the processing to the embedded JPEG preview image.

By "expose correctly" (or to the right) I mean expose well enough that you do not have to brighten the image at all when editing. In my opinion using auto ISO means you have let go of control over the exposure and the camera will be making decisions for you, and quite possibly making them poorly. For example, ISOs 125, 250, 500, 1000, 2000, 4000 are best avoided if you want to minimise noise, but by shooting with auto ISO the camera may well choose them. If you have HTP enabled then that can be a real disaster for noise. I don't ever use it. ALO is also a bit iffy and can contribute a little to noise too. Both HTP and ALO will affect your image and histogram and may lead you into a false sense of security regarding the exposure.

Also, depending on your metering setup, but possibly regardless of that, if you are shooting in tight on players wearing bright/white strips then that can throw off metering and result in underexposure unless you compensate. I prefer to shoot with manual settings in order to avoid such irritations.
Click to expand...

Tim,

Why do you say to avoid these particular ISO settings to minimise noise ?

SimonTug Wilson
 
Tugster said:
Tim,

Why do you say to avoid these particular ISO settings to minimise noise ?

SimonTug Wilson
Click to expand...


Could be something to do with them being more software-based equivalents rather than actual ISO settings which go in full stops. A few people have mentioned that... will await TD's answer :)
 
My understanding of the operation of the 7D and all other Canon cameras apart from the 1 series bodies is that the camera fakes these intermediate ISO values by exposing at the full stop values 1/3 stop below those listed and then digitally manipulating the image data to "pretend" that the file was exposed in the correct fashion. The problem with this approach, and HTP, which is even worse, is that it causes underexposure. The correction in camera adds (very slightly) to noise and may introduce a little posterisation. These are not things you want to happen. IMHO you want proper ISOs that don't digitally massage the image data. That is something which I think would be better left to the individual to fine tune at the PC.

Again, this argument applies more strongly for raw shooters than JPEG shooters, but I think everyone ought to be aware of the pitfalls. At low ISOs or small image reproduction sizes in all honesty I doubt it matters much, but if you want to maximise the potential of your camera then I suggest you think carefully before using intermediate ISOs on a Canon camera. I don't know about other brands.

Discussion here - http://forums.dpreview.com/forums/readflat.asp?forum=1019&message=37575043&changemode=1
 
Last edited:
I think your right in that 1D series have an additional amplification stage for producing the 1/3 increments as well as the 'proper' ISO numbers.

I think some of the intermediate ISO settings -1/3 from the next true ISO in one direction whilst other intermediate ISO settings +1/3 from the next true ISO in the other direction.

From memory if you're unsure how intermediate ISO choices are being handled - your EXIF data will show the 1/3 adjustment when you may not have expected it.
 
Sorry to hijack, but do Nikon have the same iso policy in their cameras?
 
I don't know. I've read stuff, just in forums, that is contradictory. As I don't shoot Nikon I have not delved into the truth. You could try Googling for something like fake Nikon ISO and see what comes up. The equivalent search for Canon turns up plenty on the subject.

What Phil says about the 1 series second stage amplifier is, as I understand it, also true, but apparently the second stage amp is not that great and may as well also be ignored. Whether the 1D4 has remedied that trait I don't know. I only have a 1D3. Anyway, FWIW I only use full stop ISOs on that camera too. It just maintains consistency between my bodies. I know one click for ISO always gives me one stop of exposure adjustment. Again, this might be hair splitting, but if something can be done better, like producing cleaner files with more dynamic range, why not do it better? I think these features might be included for convenience rather than for technical excellence. I do wish Canon would simply implement proper ISOs throughout the range. I don't understand why they do not. I'd happily use them if they worked properly.
 
Mike.P said:
Hmmm, my K-5 has 1/3 1/2 or 1 so I will stick it on 1 and see if it makes any difference (Had it on 1/3).

Interesting info, thanks for that.
Click to expand...

Yeah, but Pentax is a different game again. I have no idea whether that fakes intermediate ISOs or not. If they are real ISOs then you may as well make full use of them. The K5 is meant to have a very tidy sensor, noise wise, I thought. Part of a new breed of sensors with incredibly low read noise and zero banding or pattern noise, much like the D7000, and much unlike many of Canon's offerings (yes, 5D2, I'm looking at you).

Anyway, unless you know that your sensor has fakes, or unless you have any concerns about your IQ, why change simply because Canon cameras have an iffy ISO implementation?
 
tdodd said:
Anyway, unless you know that your sensor has fakes, or unless you have any concerns about your IQ, why change simply because Canon cameras have an iffy ISO implementation?
Click to expand...

Because after reading your post I went and did a little searching of my own and found this: ISO 160 isn't a native ISO value and is actually ISO 200 pulled back 1/3 of a stop.

Therefore I would presume its the same all the way through the range and I'm all for a little less noise at high ISO if possible ;)
 
OK, then Pentax are cheeky perishers too. :)

However, exposing at 200 ISO and pulling back to 160 in camera will not increase noise; it will reduce it, a little. The danger is, if you like to expose to the extreme right, as I often try to, that the initial capture by the sensor will be at 200 ISO, but your shutter speed, aperture and metering will be as though you were shooting at 160. In theory you may find a bit of highlight clipping that you would have avoided by shooting at a proper ISO instead of a fake. But because the file is pulled back in camera it may conceal the damage. It's not noise to worry about at such ISOs, but reduced dynamic range.

The ones to watch out for with respect to noise are the ISOs such as 125 for Canon, which is shot at 100 ISO, but metered and exposed for 125 ISO, which means a 1/3 stop underexposure. Again the in camera shenanigins will probably conceal the error from you, at least as far as the histogram and clipping warnings are concerned. There have been reports on several Canon cameras saying that 125 ISO is actually noisier than both 160 and 200. On that basis it is an utterly stupid ISO to even have available. If you want a slow shutter speed an choose a low ISO to do it you as well stick with 100 ISO. If you want a faster shutter speed, but still a low or noise free image then go straight up to 200 ISO. I just don't see the point in a pushed 125 ISO.
 
I've put mine on 1:1 and noticed a positive difference already.

Thanks again Tdodd
 
Mike.P said:
Because after reading your post I went and did a little searching of my own and found this: ISO 160 isn't a native ISO value and is actually ISO 200 pulled back 1/3 of a stop.

Therefore I would presume its the same all the way through the range and I'm all for a little less noise at high ISO if possible ;)
Click to expand...

So yep - it does it the same way.

You may find some of the intermediates use the next true ISO 'down the way' pushed a 1/3rd rather than your example of 160 using the next true ISO 'up the way' (200) being pulled back 1/3rd.
 
tdodd said:
Under what sort of lighting (please don't say fluorescent) and with what WB settings in camera? Do you shoot raw or JPEG? What, if anything, do you do to try to fix WB in your editing software? Do you perform a custom white balance before shooting? Do you include a neutral white/grey sample within your scene from which you can set WB using a colour sampler/dropper?

If you do say "fluorescent" then at what sort of shutter speeds? Please don't say faster than 1/50 (or 1/100 at a pinch).
Click to expand...

It tends to be more on the indoor shots non fluorescent...
I shoot RAW...
At the min im spending alot of time in PS adjusting the WB using the adjust tool (if there's a neutral grey on the pic) or playing with the WB settings to get it right.
I dont always do a Custom WB pre shoot (especially at weddings when time is precious)...
I try to get a sample WB...
I always shoot (handheld) as fast a poss usually a minimum of 1/50 with IS.
 
Dan, it's hard to say why you have a problem. Fluorescent lighting at high shutter speeds would have been the easy answer - they're a problem.

If you're talking about indoor shooting then presumably that's with tungsten light in the room. Is that combining with daylight coming through windows or flash? If flash, are you bouncing or using it direct? If bouncing, what sort of colour surfaces? Are they tainting the colour of the light for better or worse? If direct, are you geling the flash to close the gap in colour temperature between the flash (blueish) and the tungsten (yellowish)? If you have mixed temperature light sources then you may well struggle to get a balance that suits both. Either match the colour with gels or set things up so that one light source overpowers the other. Failing that, convert to black and white.

If none of that covers your problem then maybe you'll need to post a sample image, before any adjustments, just so we can see what you're dealing with.

Of course, if you have a series of images that need equal adjustment then it should take but a moment to apply the same adjustment throughout the set. There is no need to adjust each one individually, unless the light is all over the place from one shot to the next.
 
Cpinch73 said:
I've put mine on 1:1 and noticed a positive difference already.

Thanks again Tdodd
Click to expand...

I too have now changed to 1:1

Cheers mate :D

They didn't mention that little fact in the manual!
 
tdodd said:
Dan, it's hard to say why you have a problem. Fluorescent lighting at high shutter speeds would have been the easy answer - they're a problem.

If you're talking about indoor shooting then presumably that's with tungsten light in the room. Is that combining with daylight coming through windows or flash? If flash, are you bouncing or using it direct? If bouncing, what sort of colour surfaces? Are they tainting the colour of the light for better or worse? If direct, are you geling the flash to close the gap in colour temperature between the flash (blueish) and the tungsten (yellowish)? If you have mixed temperature light sources then you may well struggle to get a balance that suits both. Either match the colour with gels or set things up so that one light source overpowers the other. Failing that, convert to black and white.

If none of that covers your problem then maybe you'll need to post a sample image, before any adjustments, just so we can see what you're dealing with.

Of course, if you have a series of images that need equal adjustment then it should take but a moment to apply the same adjustment throughout the set. There is no need to adjust each one individually, unless the light is all over the place from one shot to the next.
Click to expand...

Generally when using flash (bounced) its ok....
Its hard to explain and it may just be me!!....i'll start a thread on it and post some shots....cheers for the advise!!!!
 
Are you shooting raw or JPEG? If JPEG then you may struggle to sort WB out in post if you don't at least get close at the time of shutter release. If raw then I'd think you could get close enough for an appealing image without too much trouble, unless you have really extreme lighting which starves one RGB channel of light so much that there is nothing there to be mended. An example would be underwater shots at depth with no flash, where there simply is no red light to speak of.
 
You must log in or register to reply here.
Back
Top