Cannot seem to get clear shots in low light - HELP

[Kieran]

Superbee you have plenty of good advice sent your way. I have the 550D so this relates specifically to your camera dials and would be the setup I would use in the situation you have described
Use the eye-piece not live view to compose & focus your shot
Set ISO to AUTO
Set Camera to M
Scroll wheel to 1/50 seconds or faster (using a 50mm lens)
Push the AV button and scroll the f/stop to your desired (for a single head shot f/2.8 for a group f/5- f/8)
Press the right magnifying button (top right corner of camera back) This will show what focus sensors are on. Scroll until you have only the middle one highlighted
Press the Q button - select evaluative metering and AI SERVO for focussing
Now your camera is set to avoid camera shake (faster than the lens) & the depth of field is set to keep focus on only those people you want in focus. The focus point is exactly in the middle of the shot. The camera will automagically adjust the ISO for each shot. You can now concentrate on composing & focusing the shot.
You will find that sometimes there is just not enough light to get a good image and this is where you give up and put the flash on.
One other thing that I find useful is to not use the shutter release button to focus. Push the menu button. Go across to custom functions. Push the set button then arrow across to 9 "Shutter/AE lock button" Select option 3 "AE/AF, no AE lock" now when you want to focus push the de-magnify button (top right corner of camera back) and take the shot with the shutter release button as normal. This is useful because you keep the focus button pushed down constantly in AI servo to track someone on the move and push the shutter release button at exactly the right time hopefully, often just a little too early or late.
With low light shots you may find you get a more consistent result by taking shots with a little extra room around the edges with the intention of cropping latter. The reason for this is that it makes you step back a little more, allowing you to use a wider aperture without having the depth of field let you down. The distance to subject is really important as well as the f/stop in giving your depth of field, it is something that didn't dawn on me as soon as it should. Hope this is of some help. Good luck

[arthurking83]

Some good sounding tips from Kieran there.

The only thing that would concern me tho is the use of evaluative metering.
I remember years ago trying to help a couple out in the bush with setting the camera up so that it would expose 'correctly' .. and by that I mean that it didn't try it's best to protect highlights.
Not being a Canon person I had no idea on how to find the menu item to stop it doing this in Evaluative mode.
Anyhow, the person was trying to get a shot of the other against the distant landscape, but the camera tried harder to protect the highlights only .. hence the other person in the portrait was all in shadow(ie. black).
I told them to switch to spot metering, but that the background would look washed out. We tried flash(on board) but it didn't do a good job of it. Camera was one of the thousand series cameras .. maybe 1100D or something. Very hard to get to any controls coming from a D300.

Anyhow .. I'd prefer to recommend using spot or centre weighted metering, especially if the ISO value risks being at an elevated value.
As ISO level increases, the exposure of your subjects is far more important not to under expose(any subsequent processing to brighten/recover will only yield more noise levels, or blurry looking details).
In fact, as ISO increases up to a point .. I tend to over expose the subject by a small amount, and keep it that way in PP(a bit of over exposure hides or masks high ISO noise a bit).
On my D300 at above ISO 1600 I'd start to over expose, normally by about 1/3Ev .. even tho ISO 3200 and even 6400 may have been usable enough at low web sized images.
So far on the D800, I've yet to do it, and I don't mind shooting at ISO6400 if I have too.

Kieran's tip on setting a button on the back of the camera to continually press whilst leaving the AF function off the exposure release is one I'd highly recommend to do also(I didn't know the Canon had this feature).

Takes just a bit of getting used too it, but well worth it.

[superbee15]

One last question, does this model camera have live view mode?

Yes it does but I do not use as I figure that would be more prone to camera shake.

Use the eye-piece not live view to compose & focus your shot

One other thing that I find useful is to not use the shutter release button to focus.

Always use left eye on eyepiece - right eye not good enough!
Have tried back focus previously and it was a bit hard to use. I will retry.

I tried a simple test with the 50 this afternoon and set camera on kitchen bench and focused on some food on bench (camera set to AI Servo). Sometimes the lens would lock and sometimes it just continued to hunt even though everything is static. I was using the center focus point from about 500mm and looking at same point on food. Light was not bad. Is this normal for AI Servo?

[Kieran]

No that isn't normal unless the food was very old and walking out of your kitchen. Try with AI SERVO again, then change to AI FOCUS, then to ONE SHOT to see if there is any difference between these 3 options on the amount of hunting going on. Low light is when it should have trouble so if you then put some decent light onto the subject the hunting should go away. Do you have AF-assist beam firing enabled? This is in Menu-Custom functions-7 "Autofocus/drive" select option 0 Enable. Then put on another lens and check to see that it is not a problem with this lens only.

[dtmateojr]

All these were taken at 1/20th second. Perhaps you might like to reconsider your conclusion?

The parts that are in focus are reasonably sharp which rules out camera or subject movement. The parts in focus are in front of their faces.

[arthurking83]

..... Sometimes the lens would lock and sometimes it just continued to hunt even though everything is static. I was using the center focus point from about 500mm and looking at same point on food. Light was not bad. Is this normal for AI Servo?

Actually! .. it can be.
It depends on your stability.
The is the rocking to a fro(or back and forth I wrote about previously.(apparently I do this .. so I think it's normal).
This is why you'd want to use this AI servo focus mode. It counteracts any fore aft movement you may naturally do when you're photographing.

Your movement may only be small, but it could be significant enough to make a difference.
Again, you can do this deliberately as a test to get a feel for it.
AI servo mode, find a static target and you deliberately rock yourself closer and further from the subject as you continually focus.
If you can manage the same target with the camera on a steady support(as before, even a solidly planted table will suffice, but tripod is better) .. and find a nice large target and using continuous focus then you will see that the focus system will not hunt(well it shouldn't!).

Also the type of target or the subject being focused on will have an impact on the focusing system and side issues like hunting and stuff like that.

[superbee15]

Well did some test shots this afternoon. Following method:

• Tripod for all;
• No shots with flash;
• Shots taken mid afternoon with light varied with blinds and fluors;
• Used a wired remote controller;
• Always focused over Snoopy's left eye - centered on the black;
• Used only the Center Focus Point;
• Tried the 50m lens then the 18-55 kit set (did not include any shots from this lens as seemed ok) at 50m then swapped back to the 50m all without moving camera to any large degree;
• Camera placed 1.63 meters from eye of Snoopy;
• Red figurine 65mm in front of plane of the left eye of Snoopy;
• The leading edge of the apple in fruit bowl is 160mm in behind of plane of the left eye of Snoopy.

First shot and all not bad was with the 50 and was set at

• ƒ/2.0
• 50.0 mm
• 1/20
• 100
• One Shot AF

IMG_4776 by Superbee15, on Flickr

Next shot review shows more DOF and seems good and was set at

• ƒ/9.0
• 50.0 mm
• 0.6
• 100
• AI Servo AF

IMG_4795 by Superbee15, on Flickr

3rd Shot
This was the 3rd shot with reduced light and increased ISO. The 1st 2 seemed ok but this seems to have not focused correctly with the camera/lens hunting a bit (inadvertently used 1.8) and settings were

• ƒ/1.8
• 50.0 mm
• 1/250
• 1600
• AI Servo AF

I do feel there seems to be much more focus of the red figurine in this shot when compared to the 1st shot!

IMG_4800 by Superbee15, on Flickr

4th shot
After the hunting with previous shot I opened up to light again and went back to better light settings. This is where things do seem to be wrong. It should have improved focus but did not, it even may be worse. The shot then ends up with more focus on the red figurine. Settings were

• ƒ/1.8
• 50.0 mm
• 1/60
• 200
• AI Servo AF

IMG_4801 by Superbee15, on Flickr

In between previous and next shots I did download shots and I moved the camera and then reset back in same place to retake some.

5th Shot
This shot seems to be reasonable although the red figurine still seems to have very slightly more focus than the 1st shot. Settings were

• ƒ/2.0
• 50.0 mm
• 1/30
• 100
• AI Servo AF

IMG_4803 by Superbee15, on Flickr

Thoughts ?

[fess67]

My personal thought is you are looking for a camera focus issue when I think it is much more likely to be a user error. Please do not take that personally, I have done exactly the same with my cameras.

I think a better way for you to check your focus is to use the set up you already have but use a ruler on the bench as your subject - or one of a number of focus cards easily found online. Have the 0 of the ruler closest to you and the end of it furthest, use single point focus on One Shot. Throw lots of light on it and make sure the shutter speed is 1/60 or higher (preferably higher). Focus on the 10 cm mark and see if that is the point in focus.

You may also consider using mirror lockup to ensure no anomalies are thrown in due to mirror slap.

AI Servo is designed for moving subjects, using that mode may induce errors as well when mounted on a tripod.

Of course you may well have a focus issue as well, just worth while doing some more scientific tests and give the camera a chance by increasing your shutter speed above that which you have been using so far.

[superbee15]

My personal thought is you are looking for a camera focus issue when I think it is much more likely to be a user error. Please do not take that personally, I have done exactly the same with my cameras.

No offense taken as the post was commenced under assumption I was the one doing something wrong and I am trying to determine what. I had also never heard anything about front/back focus.

I took your advice and found a card and tried some more shots. Again like yesterday following method:

• Tripod for all;
• Used Mirror Lockup - new;
• No shots with flash;
• Shots taken mid afternoon with 1 set done outside to ensure good light and the reduced light with blinds and fluors inside to take higher ISO;
• Used a wired remote controller;
• Always focused on center line;
• Used only the Center Focus Point;
• Used the 50m lens only;
• Camera placed 0.63 meters from center focus line.

I took 3 sets (covering ISO 100, 6400 and 1600) of 8 shots covering following f-stops - 1.8, 2.0, 2.8, 4.0, 4.5, 5.0, 9.0 and 13.

Analysis - to my limited experience - seems to display slightly more focus consistency in good light around f4 and f4.5. An example is the 1st photo below.

I only seemed to achieve 1 shot at both ISO6400 and 1600 that I could call balanced focus. All other shots varied from slightly front focus to definite front focus. Mind you I am NOT saying there is a lot in it. I also could NOT get any shot to come out as bad as the worst shot from yesterday. After taking the 3 sets I set mirror back to normal and varied between AI Servo and one shot and also tried one with flash and all seemed consistent with other shots taken.

The 2nd and 3rd photos below are to me front focused and were shot at ISO 6400 and 1600.

It also seems to me that in low light and large aperture the lens tends to have more front focus while in good light the smaller aperture seems to have more front focus.

Shot 1

• ƒ/4.0
• 50.0 mm
• 1/60
• 100

IMG_4812 by Superbee15, on Flickr

Shot 2

• ƒ/2.0
• 50.0 mm
• 1/500
• 6400

IMG_4824 by Superbee15, on Flickr

Shot 3

• ƒ/1.8
• 50.0 mm
• 1/125
• 1600

IMG_4833 by Superbee15, on Flickr

[arthurking83]

One thing .. be careful with your terminology.
Smaller aperture = larger number. If you mean smaller aperture number .. call it f/stop. f/stops are the actual ratio values .. such as f/5.6 or f/11(smaller aperture)
(the aperture is the physical opening of the lens! .. technically it's called the diaphragm .. but that's not important).

Larger aperture means bigger opening .. smaller f/stop number(eg. f/1.8 f/2 .. etc).

Your preliminary tests seems to indicate that the lens may be front focusing .. but it's still as yet inconclusive.
If it was, it would do so all the time(unless something went awry in the first test shot).

In the lower light test shot, the conclusion would be that the combination of the 50/1.8 and the 500D aren't perfectly suited for perfect and consistent low light AF!

So in Ai mode, the camera(+ this lens) seems to struggle to maintain focus.
This shouldn't be an uncommon situation .. although not one you want to have to put up with.
Now that you know this, you can 'teach yourself' to work around it.

If you can ... just one more quick test which shouldn't take up too much time for 'ya.
Switch the subject matter around.
Put the red figurine up on a pedestal(she seems to be the star of the moment anyhow! ) .. and use her as the focus point.
Put Snoopy on the sidelines still in the frame, as a gauge on focus depth.
The red figurine has more contrasting patterns and lines for focus not get confused, whereas snoopy can cause focus locking errors.
Keep it low light. F/2 and f/4 or f/5.6 should be sufficient for determining if you do have a focusing issue with this lens on this camera or not.

Something else that you can do as well. You don't need to actually take the shot .. this is more for confirmation purposes and posting questions to others.
As you have liveview, what you can do is to focus, using Ai mode, briefly .. say a second or so, stop and then switch to Lv mode.
Zoom in to 100% view and determine if the image looks sharp on the review screen. If it's sharp, great. If it looks slightly blurry, you can 'jimmy' the lens' focus collar to get a sharper looking image.
The trick is to determine which way you need to adjust focus on the lens, and by (roughly) how much.
two or three runs should be enough .. but be sure to defocus the lens before each run .. and do so both ways(that is, throw focus out towards infinity and then the next attempt towards minimum focus distance too).

Repeat the above test in Single shot AF mode too .. just to see what happens, and if it works differently to Ai mode.

FWIW: the reason for the above test again is to discount that Snoopy wasn't the issue here!
The way focus systems work is that they see lines. Basically they see a series of pixels, which make up lines and try to render these pixel lines with as much contrast as possible.
The less fine the pixel/line subjects are .. the harder the AF system has to work to achieve consistently accurate focus.
The red figurine has a lot more detail for the AF system to 'see'.

If you want to understand another technical aspect that can affect focus systems .. read my reply to THIS thread too.
My reply with the large red squares, apparently, is a common problem that probably affects all cameras in some way .. some less so, others more so. The tolerances are so tiny in AF systems, that this can't be helped.

FWIW Pt II: the best outcome with your issue will be that it is a user error issue, and not a physical issue with your gear! User errors is easily fixable .. gear issues cost you $$$!(hope for the former! )

[superbee15]

Larger aperture means bigger opening .. smaller f/stop number(eg. f/1.8 f/2 .. etc).

Yep was across that even though it has taken a bit for it to sink in. Will refer to as f/stop in future.

Thanks, will look at further tests.
Cheers

[superbee15]

One thing .. be careful with your terminology.
...Put the red figurine up on a pedestal(she seems to be the star of the moment anyhow! ) .. and use her as the focus point.
Put Snoopy on the sidelines still in the frame, as a gauge on focus depth.
The red figurine has more contrasting patterns and lines for focus not get confused, whereas snoopy can cause focus locking errors.
Keep it low light. F/2 and f/4 or f/5.6 should be sufficient for determining if you do have a focusing issue with this lens on this camera or not.

Something else that you can do as well. You don't need to actually take the shot .. this is more for confirmation purposes and posting questions to others.
As you have liveview, what you can do is to focus, using Ai mode, briefly .. say a second or so, stop and then switch to Lv mode.
Zoom in to 100% view and determine if the image looks sharp on the review screen. If it's sharp, great. If it looks slightly blurry, you can 'jimmy' the lens' focus collar to get a sharper looking image.
The trick is to determine which way you need to adjust focus on the lens, and by (roughly) how much.
two or three runs should be enough .. but be sure to defocus the lens before each run .. and do so both ways(that is, throw focus out towards infinity and then the next attempt towards minimum focus distance too).

Repeat the above test in Single shot AF mode too .. just to see what happens, and if it works differently to Ai mode.

Well did some more shots this afternoon and again have come across a miss on focus. Had same setup as yesterday in terms of tripod and mirror and remote.

However substituted Snoopy for bottles on basis that they may show focus better than him. Included ruler as a backup - which I think was just as well. The focus was over Reds eyes and nose. The vertical plane for that focus ends up on the 23 of the ruler markings.

I guess I did 22 shots and just near the end when changing from 5.6 to 1.8 it did not focus as it should (I think it was short) the next at 2.0 was same (missed focus) and the next shot was ok. All four shots were comp[leted inside of a 2 minute period with only changing some settings. Have included 1st 2 shots of this 4 shot group.

I also tried the 2nd test and I did adjust minutely the focus- it was very fine and I know my eyes are not the best. However when pushed to infinity to the right (is that correct) I adjusted slightly back to infinity on live view and when pulled to minimum I have to adjust slightly back to minimum - this was in 'one shot'. In AI Servo it seemed to perform ok.

I also think it did very well in live view - I tried a few re-focuses while zoomed so not sure if this a different circumstances.

Overall I do not think I did anything wrong on these shots but again seem to have come across random misses.

Shot 1

• ƒ/5.6
• 50.0 mm
• 1/15
• 1600

IMG_4861 by Superbee15, on Flickr

Shot 2

• ƒ/1.8
• 50.0 mm
• 1/160
• 1600

IMG_4862 by Superbee15, on Flickr

[jev]

Seems to me you definitely have some front-focus issues in this setup. Unfortunately, the 550D seems to be missing microfocus-adjustment features... *sigh*

[arthurking83]

Yep! As jev said .. 99.9% front focus issue with this lens.
If you have the time an patience to check any other lens you have .. it'd be recommended. This step will just confirm if it's the camera or the lens that's causing the misfocus.
(in almost all instances, it'll be the lens .. but this is not set in stone).

Liveview focusing will be immune to any AF issues, as the AF system in Lv mode is the actual sensor.
When using the viewfinder, there is a separate AF module that does the focusing. This tells the lens where to set the focus distance.

The two possibilities are that the AF sensor is slightly out of whack and tells the lens to move to an apparently correct focus distance .. which in reality is not.(camera issue)
In this instance you can confirm this by testing another lens.
The other possibility is that the lens is slightly out of whack, and is simply not setting itself to the correct focus distance that the camera tells it too. It can be adjusted by the Canon service department.
The question is, the price of this service relative to the cost of the lens.

Other things you can do to work around the issue for now are to stop the lens down a bit more(say f/2.5-f/4) .. or use Lv mode for more accurate wide open settings.
More DOF will slightly mask the effects of misfocusing.
This is one thing to watch for if you use a kit type lens as the other lens you test with. The smaller inherent aperture values of these lens types mask the effects of misfocusing.

[superbee15]

Ok will check both kit lens as they are all that I have.

The 50mm lens is just 9 months old so if there is an issue with just this one would it likely be covered? If not at just over $100 bills it most likely not worth worrying about.

And thanks for your advice and help.

[arthurking83]

.....

The 50mm lens is just 9 months old so if there is an issue with just this one would it likely be covered?

....

I think it should(but I don't know the conditions).
Because it's still under warranty, for sure take it to get it checked out, or replaced or whatever.

if you check your other lenses and they work OK .. if you take it back be sure to explain to them that this 'back focus' issue is limited to this lens only!

What they can sometimes do is to ask for the camera and lens to calibrate them.
There is a way to quickly adjust the camera so that it makes the lens focus correctly(or at the correct point) by adjusting the mirror.(this is easy enough for most folks to do it themselves!)
problem is tho, if the other lenses work OK now, if you adjust the camera to compensate for the one lens, the other lenses will almost certainly be out of whack. (as you'd expect . not ideal)
So it's important that they know that this(or any other) lenses focus OK, or not.
If you have this written down in a service agreement form(or something like that), and they stuff it up(to try to cheap out on costs) .. you have something to come back on.

The lens can be adjusted to focus correctly .. it's just more time consuming.

good luck with it all.

[superbee15]

Thanks Arthur

Have finally checked the other kit lenses and they both perform better overall than the 50mm lens. I took quite a few shots with both and the majority of the time they seemed very close to even. There were a few that had slight front focus and a odd one that had slight back focus. But I did not get any misses like what has happened with the 50mm lens.

Going away soon for a short trip and would have liked the 50 to be true - but not enough time to put in to a service agent. So I guess I will do that when I return.

Thanks to all for input.

Cheers

[arthurking83]

.....

Have finally checked the other kit lenses and they both perform better overall than the 50mm lens. I took quite a few shots with both and the majority of the time they seemed very close to even. There were a few that had slight front focus and a odd one that had slight back focus. But I did not get any misses like what has happened with the 50mm lens.

.....

Just remember that if you compare your other lenses to the 50/1.8, they should be at the same focal length(or very close to it) and aperture settings.
eg. your 18-55 should be at f/5.6 and 50-ish mm and your 50/1.8 should be set to f/5.6

if your other lens is a 55-200(for example) then 55mm is OK, but if it's f/3.5 then compare it to the 50/1.8 set to f/5.6(just to be sure that it's not hazy when set wide open).

if you don't compare them at the same apertures, then the naturally deeper DOF of the other lenses will fool you into thinking that they focus more accurately.

Faster lenses are harder to test wide open because of this lack of DOF when set wide open .. and stopping them down isn't much help as they can also be affected by focus shift as aperture values change.

If you had access to another camera to assess that the lens positively misfocuses .. then it would be cut and dried! But faster lenses can be harder to focus perfectly on less capable camera bodies.

[swifty]

There are already a lot of good diagnostic info in this thread but I'll add just one observation I've noted in your tests.
Not only does it appear you have some front focussing with your 50mm lens, it appears to have a small degree of focus shift.
Your wide opened test shots (@f1.8) appear to be clearly front-focussing.
In your stopped down test shots, you do get an overall sharper image due to the increased DOF but if you look closely at the plane of focus, it appears to have shifted, compensating for some of the initial front focus issue.
In ruler test 1(2nd test), max sharpness at f1.8 appears around the ruler 1.4 whilst max sharpness at f4 appears around 0.6 to my eye.
In ruler test 2(3rd test), the f1.8 shot appears to have max sharpness around the 15cm but shifts to around the 17/18cm mark in the f5.6 shot.
Both are a bit hard to see at these images sizes but you get the point there's a shift. But these are not definitive since this isn't the correct way to test for focus shift. Nor do I think this is your main issue but just making the observation that if focus shift is present on your lens, even if you can get focus spot on at max aperture, you may notice some slight inconsistencies stopped down (although this may be largely hidden with the increased DOF) except maybe around minimum focusing distance when DOF is pretty thin.

[arthurking83]

I think you're right swifty .. an element of focus shift needs to be taken into account too. In fact it's quite common for fast lenses to have a degree of focus shift inherent(my three all have some).
But also! .. what's not so obvious and a bit hard to determine is the straightness of the sheet of paper.
You can clearly see the curvature of the sheet at the lines, but obviously not on the plain section. This could have an impact.

But it's being pedantic, and I think there may be a small amount of front focusing on this combo.

What is going to be hard to figure out (without some cost) is if it's the lens alone(ie. just out of adjustment!) or is the combo of this lens on this camera

ie. there's no point in paying money if it's the combination of both, just taking the lens in for service and then finding the camera itself is causing some of the issue.