Circle of Confusion?

Can anyone tell me what this is?

Is it something to do with the focusing not being 100% on a subject or small area on the subject where there is a blur?

I am reading an article on wiki and I am still confused, excuse the pun. Is this more apparent on certain genres of photography like fashion and portraits or all genres of photography?

Would it be kind of someone to show me an example of this?

There have been very many confusing descriptions of the photographic term "circle of confusion", but this video titled Circle of Confusion Explained and Illustrated is about the best explanation I have seen on the web.

SkipD wrote in post #16323811
There have been very many confusing descriptions of the photographic term "circle of confusion", but this video titled Circle of Confusion Explained and Illustrated is about the best explanation I have seen on the web.

Thanks, I think that video has cleared that up for me 100%

Rich, the "circle of confusion" thing goes a bit further when considering depth of field. In calculating depth of field, the primary consideration is how large a point of light which has been photographed could look on a standard sized print which is viewed at a standard distance by a person with a certain level of visual acuity and still be considered a "point" by the viewer. This, too is a circle of confusion. Then, the amount of magnification between the in-camera image (think a film negative or slide, but this applies to a digital sensor as well) and the standard sized print (8x10 inches or so) is considered to determine what the largest "acceptable" in-camera circle of confusion is when considering a point to be "in focus".

If I've confused you, please let me know and we'll put it all together....

SkipD wrote in post #16323909
Rich, the "circle of confusion" thing goes a bit further when considering depth of field. In calculating depth of field, the primary consideration is how large a point of light which has been photographed could look on a standard sized print which is viewed at a standard distance by a person with a certain level of visual acuity and still be considered a "point" by the viewer. This, too is a circle of confusion. Then, the amount of magnification between the in-camera image (think a film negative or slide, but this applies to a digital sensor as well) and the standard sized print (8x10 inches or so) is considered to determine what the largest "acceptable" in-camera circle of confusion is when considering a point to be "in focus".

If I've confused you, please let me know and we'll put it all together....

lol, I was in happy bliss when you gave the link to the video, hahaha

It has confused me again

RichSoansPhotos wrote in post #16323915
lol, I was in happy bliss when you gave the link to the video, hahaha

It has confused me again

What do you seem to be misunderstanding?

The CoC that people normally refer to (0.03mm, etc...) is more formally the circle of confusion criterion. That is to say it is the threshold of the acceptable size of the circle of confusion or blur circle.

As Skip said, this assumes a standard print size, viewing conditions and visual acuity of the viewer.

With an 8x10 print for an 8x10 film, a typical CoC criterion would be 0.2mm. We're basically saying that at a standard viewing distance a circle that is 0.2mm in diameter will look like a point to your average (or arguably below average) person.

With an 8x10 print for a 4x5 film, the CoC criterion is 0.1mm because you have to enlarge the image 2 times and thus something that is 0.1mm in diameter on the film will become 0.2mm on the print. With smaller formats, the enlargement is even more and thus the smaller CoC criterions.

e-k

And there lies the challenge to making a lens. You have to funnel RGB light waves to a pin point. Sometimes R and G lightwaves cooperate and B won't which makes the circle larger and effects lens image quality.

Primes are easier to funnel light to a pin point. Zooms are harder as this needs to be done at various focal lengths. This is why primes are typically sharper than zooms and why zooms tend to be sharper at one end than the other.

RichSoansPhotos wrote in post #16323915
lol, I was in happy bliss when you gave the link to the video, hahaha

It has confused me again

Ok, baby steps then:

1. From a given distance, a person of 20/20 vision has a set capacity to resolve detail on print.

2. By taking advantage of our limit to see detail from the distance that we expect a person to look at the print from, we can determine how out of focus the image can be before it becomes noticeable.

3. We take the above concept into account when taking the photo itself, placing focus on a point in space that can encompass as much in it's depth of field as possible, details that seem out of focus on close inspection will actually appear sharp in print, when viewed from afar.

4. Large photographs can get away with using the same amount of resolution as a small image, so long as we expect the person viewing the image to look at it from further away.

So, in it's essence, COC is all about how out of focus your image can be before it becomes visible to the eye, whether on screen or print.

A good example of COC in play is when viewing small images online, you can shoot from any camera and lens, but if you resize them all to 1024 pixels on a side, you can make the picture from a $500 camera look as good as that from a $50,000 camera. Therefore, expensive cameras and lenses are all about output size, if you aren't printing large (or aren't printing at all), you don't need more than a dozen MP worth of resolution.

The COC that the video linked to above talks about is a different, but still relevant topic. There, the COC of the sensor determines two things:

1. The aperture at which diffraction sets in.

2. How out of focus or blurry an image can be before it becomes visible at 100% magnification.

Let's take a couple of sensors, one with 18mp that has fewer and larger pixels and a 36mp sensor, and they are the same size. Now, in order for a sensor to create detail, you need two pixels, because detail is created from contrast - one white pixel, one black.
Since the pixels of the 18mp sensor are much bigger, the point of light that falls on them can be larger (blurrier or more out of focus) before the sensor registers it as smeared detail. Specifically, the point of light has to be wider than two pixels across in order for it to be visible.

When you stop down a lens, the smallest size that any point of light can be increases, and this is called diffraction; softness from diffraction sets in when the minimum point size becomes larger than two pixels on the sensor. For an 18mp sensor this can easily be as late as f/14, while for a 36mp sensor, this can be as early as f/6.3. If we had a sensor of infinite resolution, it would continue to be diffraction limited even with f/1.2 lenses wide open!

Important! - The COC of the detail that the sensor records has nothing to do with how the image will look when printed, if you were to shoot two images with the 18 and 36mp cameras at f/14 and print them at around 13x19", they would look exactly alike! This is because the large amount of diffraction softness in the 36mp shot would be lost to the limited amount of detail our eyes can see on such a small print at a distance.

Time for a breather.

You might have noticed that I said sensor COC determines how blurry your image can be before it seems blurry. If we take the shaking if your hands, or any other vibrations as linear displacement of the sensor, a lower resolution sensor would need to see a much greater level of vibrations than a high resolution sensor before recording it as blur. To put it into practice, the 18mp sensor may be ok with the 1/FL rule for sharp hand held shots (ex: 1/50th for 50mm lens), but a 36mp sensor may need twice the shutter speed to make sure the shot is sharp at the pixel level.

So in addition to the fact that you can judge the amount of detail you can expect your sensor to resolve, and the eye of the person viewing a print, you can also expect the paper itself to limit the amount of detail that can be reproduced. This can be different even with different types and finishes of paper.

So in closing, photographers since long ago have used COC to shoot their images already having in mind the size they would want their images to be, taking into account format size, aperture, and the amount of enlargement they could get away with.

Kolor-Pikker wrote in post #16324161
The COC that the video linked to above talks about is a different, but still relevant topic. There, the COC of the sensor determines two things:

1. The aperture at which diffraction sets in.

2. How out of focus or blurry an image can be before it becomes visible at 100% magnification.

You are confusing some concepts here. The video does describe the normal definition of COC and is why they reference target area and state that different media have different COC criterion.

The pixel density on a sensor (or grain size on film for that matter), sets a lower limit on the COC criterion but this limit is typically less than the standard COC criterion (for example an 8MP Canon 350D has one pixel every 0.007mm). When you say that the blur circle is not visible when a digital image is viewed at 100%, this is the limit you are looking at. That is, it is the same as saying you want your COC criterion to be 0.007mm.

Yes, diffraction also puts a lower bound on the COC criterion you can use and it is something to be aware of when using small apertures, but unless it exceeds your chosen COC criterion, then you needn't worry about it too much.

e-k

E-K wrote in post #16324314
You are confusing some concepts here. The video does describe the normal definition of COC and is why they reference target area and state that different media have different COC criterion.

It may have gone by too quickly, but I thought they only showed the COC for different film and sensor formats, not output format. Edit: yep, I watched it again and they only talk about sensor COC and not output COC, which I believe to be more important for making actual photos rather than pixel peeping.

The only time sensor COC directly affects output quality is in video, where the capture format is typically of same or similar resolution to the output, more so exaggerated by the large enlargement factor on screen.

This second part I address later on, which probably means you didn't read the whole thing, I even marked it with Important!:

Kolor-Pikker wrote in post #16324466
It may have gone by too quickly, but I thought they only showed the COC for different film and sensor formats, not output format. Edit: yep, I watched it again and they only talk about sensor COC and not output COC, which I believe to be more important for making actual photos rather than pixel peeping.

It's partially a terminology thing. In common parlance the COC criterion is the COC. In this case, it always refers to the diameter of the blur circle at the point of capture. In the video he states "the measurement where a point of light grows to a circle that you could actually see it in a final image is the circle of confusion". So yes, he is talking about the standard photographic definition of COC.

I agree that when you use the definition that COC is the blur circle, then yes you could say that there is a diameter for the COC/blur circle of the sensor/film plane that is proportional to the one for the output format. Never heard it described that way but that just may be me.

No I read it, I just don't quite agree with it . If I shot it at f/2 they still would look the same. When you starting talking about 18MP vs 36MP (assuming the same format), you are talking about the sensors ability to "resolve" the circle of confusion (and detail in general) and not the circle of confusion itself.

e-k

E-K wrote in post #16324654
It's partially a terminology thing. In common parlance the COC criterion is the COC. In this case, it always refers to the diameter of the blur circle at the point of capture. In the video he states "the measurement where a point of light grows to a circle that you could actually see it in a final image is the circle of confusion". So yes, he is talking about the standard photographic definition of COC.

I agree that when you use the definition that COC is the blur circle, then yes you could say that there is a diameter for the COC/blur circle of the sensor/film plane that is proportional to the one for the output format. Never heard it described that way but that just may be me.

Hmm ok. I need to go to seep now lol.

No I read it, I just don't quite agree with it . If I shot it at f/2 they still would look the same. When you starting talking about 18MP vs 36MP (assuming the same format), you are talking about the sensors ability to "resolve" the circle of confusion (and detail in general) and not the circle of confusion itself.

Yes, for a small print size, you wouldn't be able to tell the difference either way, small aperture or big... and that's the point I was making. So I'm not sure what we disagree on since we seem to be saying the same thing but in different ways. Maybe I didn't phrase it clearly.

Kolor-Pikker wrote in post #16324720
Hmm ok. I need to go to seep now lol.

Sleep is good .

Yeah, maybe it is just a wording thing, I was reading this,

and was a bit confused. You have the same amount of diffraction in the 18MP shot as the 36MP one. Even for a larger print viewed closer, the 36MP shot should never look worse than the 18MP shot. So I wasn't clear on what exactly you were trying to get across.

e-k

E-K wrote in post #16324822
and was a bit confused. You have the same amount of diffraction in the 18MP shot as the 36MP one. Even for a larger print viewed closer, the 36MP shot should never look worse than the 18MP shot. So I wasn't clear on what exactly you were trying to get across.

Ah, I get it now, but whatever... I'll proof the post later.

Now the most important thing OP needs to take away from all of this, is that COC is a device photographers created for non-photographers, since as soon as another photographer is involved, there is no confusion as to whether or not he'll be looking at the print from sniffing distance.