Thanks Samir. I have come to realize this about fb (make it your friend). I would like to find a way to provide the kids with their photos but still have people directed towards my site to order. Do you have advice on how best to do that? Enable my links? Also, what size do you think I should limit to? Would you go all the way to medium? I have now gotten the watermark thing down so bit by bit
What I do is enable links, and reduce the size down to medium. This is of course, with watermarks. My watermark is a url (Huntsville Car Scene.com), so people know where the photo is from.
btw what camera body were you shooting with? also can you post a larger size of the photo? presently I don't see much noise.
2) there is some evidence that shooting "in between" iso values may cause additional noise. Stick to 100,200,400,800,1600 when you can
Happened to chance upon this and same time chance upon an article that talk about using non-std ISO setting.
Its too technical for me to digest and in my opinion I think the subtle difference in noise at non-std ISO setting should not be noticeable.
Here's the article
The problem of which you speak only exists in Canon DSLRs, AFAIK, and is different for different models.
There are two basic sources of most noise; shot noise, which has to do with the number of photons captured, and read noise, which is noise generated reading the sensor and digitizing its signal. The shot noise part is more dominant in higher tonal ranges, and low ISOs. Shot noise varies in a camera ONLY with exposure. So, with the camera set to Av-priority, and an EC of zero, there will be a smooth progression of shot noise going from ISO 100 to 125 to 160 and to 200 (1/6 stop each increment); the trade-off is simple.
Read noise comes from camera electronics, and dominates low tonal ranges and higher ISOs. The way Canon generates its 1/3 stop ISOs is quite flaky. What they do is, they use the same amplification upon reading the sensor for 3 ISOs like 160, 200, and 320, and they adjust the exposure as they should, but multiply the values to bring them back to "normal". IOW, ISO 160 is exposed for ISO 160, but recorded at the same gain as ISO 200, and then the hot values are multiplied by 4/5 (so that middle grey is the same value at both ISOs, according to metering). ISO 250 works the opposite; it is exposed for 250, with the same gain as 200, but the results are multiplied by 5/4 to normalize them. So far, so good, but on Canon DSLRs, the read noise is the same strength at ISOs 100 and 200, so when ISO 125 is attained by underexposing 100 and multiplying by 5/4, it has more read noise than ISO 100, and when 160 is obtained from 200, it has less read noise than 200, and ISO 125 winds up with 1.6x as much read noise as ISO 160, even though the former is a lower ISO.
Will you see the read noise? Yes, if you use the shadows; no, if you don't. If you use Highlight Tone Priority, you will be much more likely to see the read noise from the 250/500/100 group.
This is for the APS-C models, and the 5D2 (not sure about the new 1Dmk4), For the older 1D* and 1Ds* models, all the intermediate ISOs are achieved through underexposure. Non of this matters above ISO 1250, because the main ISOs have read noise nearly proportional to ISO, so these glitches do not appear. The 125/250/500/1000/2000 group do, however, have 1/4 stop less DR because of clipping caused by the 5/4 multiplication, on the APS-Cs and the 5D2. The 1Ds lose DR at all intermediate stops.
Happened to chance upon this and same time chance upon an article that talk about using non-std ISO setting.
Its too technical for me to digest and in my opinion I think the subtle difference in noise at non-std ISO setting should not be noticeable.
Here's the article
The problem of which you speak only exists in Canon DSLRs, AFAIK, and is different for different models.
.
.
.
snip
.
.
That was a good read. Could you, perchance, include the source link? That would pretty much finish out the post as well as provide proper attribtution.
Know probably a day late and a doller short in this post. Just bought a canon 10D. Noise is ridiculous @ ISO 500 and above. I am using a L lense. Now with the next model Camera 20D, noise is not that bad @ ISO 800. I also have a 5D MKII, noise is not bad up until ISO 3200 thats when it is noticable.
Newer camera bodies are probably better for noise reduction. Newer sensors can handle higher ISOs.
Old canon 20D photo @ ISO 400 @ 2.8, 1/8 second pushing it no noise, this is a pretty old Canon too!!!How flaky can that get?
Comments
Want faster uploading? Vote for FTP!
Happened to chance upon this and same time chance upon an article that talk about using non-std ISO setting.
Its too technical for me to digest and in my opinion I think the subtle difference in noise at non-std ISO setting should not be noticeable.
Here's the article
The problem of which you speak only exists in Canon DSLRs, AFAIK, and is different for different models.
There are two basic sources of most noise; shot noise, which has to do with the number of photons captured, and read noise, which is noise generated reading the sensor and digitizing its signal. The shot noise part is more dominant in higher tonal ranges, and low ISOs. Shot noise varies in a camera ONLY with exposure. So, with the camera set to Av-priority, and an EC of zero, there will be a smooth progression of shot noise going from ISO 100 to 125 to 160 and to 200 (1/6 stop each increment); the trade-off is simple.
Read noise comes from camera electronics, and dominates low tonal ranges and higher ISOs. The way Canon generates its 1/3 stop ISOs is quite flaky. What they do is, they use the same amplification upon reading the sensor for 3 ISOs like 160, 200, and 320, and they adjust the exposure as they should, but multiply the values to bring them back to "normal". IOW, ISO 160 is exposed for ISO 160, but recorded at the same gain as ISO 200, and then the hot values are multiplied by 4/5 (so that middle grey is the same value at both ISOs, according to metering). ISO 250 works the opposite; it is exposed for 250, with the same gain as 200, but the results are multiplied by 5/4 to normalize them. So far, so good, but on Canon DSLRs, the read noise is the same strength at ISOs 100 and 200, so when ISO 125 is attained by underexposing 100 and multiplying by 5/4, it has more read noise than ISO 100, and when 160 is obtained from 200, it has less read noise than 200, and ISO 125 winds up with 1.6x as much read noise as ISO 160, even though the former is a lower ISO.
Will you see the read noise? Yes, if you use the shadows; no, if you don't. If you use Highlight Tone Priority, you will be much more likely to see the read noise from the 250/500/100 group.
This is for the APS-C models, and the 5D2 (not sure about the new 1Dmk4), For the older 1D* and 1Ds* models, all the intermediate ISOs are achieved through underexposure. Non of this matters above ISO 1250, because the main ISOs have read noise nearly proportional to ISO, so these glitches do not appear. The 125/250/500/1000/2000 group do, however, have 1/4 stop less DR because of clipping caused by the 5/4 multiplication, on the APS-Cs and the 5D2. The 1Ds lose DR at all intermediate stops.
My Photos
Thoughts on photographing a wedding, How to post a picture, AF Microadjustments?, Light Scoop
Equipment List - Check my profile
Newer camera bodies are probably better for noise reduction. Newer sensors can handle higher ISOs.
Old canon 20D photo @ ISO 400 @ 2.8, 1/8 second pushing it no noise, this is a pretty old Canon too!!!How flaky can that get?