cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

5D MK IV exposure issues - HIGH ISO necessary... tried two different lenses

Saraswati0
Apprentice

Hey. I've been shooting with my Canon for a couple of months now. It's already been back to the factory because of a shoddy touch screen. But recently i've noticed with TWO different lenses, some pretty serious issues with exposure and low light shooting.

I shot a yoga class in a studio with three large windows and in order to be at 1/250th f/5.6 I had to up the ISO to nearly 6000 which ended up being really grainy and almost useless. This was with my Sigma 35mm ART lens.

I JUST purchased the Sigma 50mm 1.4 lens and last night I had to shoot an indoor party event. I found myself getting all the was to f/1.4 1/30th and ISO 2500.

 

These numbers just DONT seem right.... is there something that could be wrong with the body itself? I just don't understand how two lenses could be having the same issues. I'm considering sending the 50mm back since its brand new and getting the Canon 50 1.2 instead... but thats only if the lens is the issue and not the body.

Anyone?

19 REPLIES 19

Thanks everyone. I wasn't necessarily trying to make those images quality in terms of HDR but rather just shot them to explore/show what seemed to me to be the need for high ISO in low light situations despite a wide aperture and slow shutter. I guess i'm confused how people can use a 50 1.4 at concerts with crazy inconsistent lighting and stop action without a flash if in the scenario I provided, though it isn't super bright in the room, I had to be at 1/10th...

Basically I bought this lens so that I would have the ability to take consistently sharp images, without flash at indoor events or for lifestyle shoots (like the indoor yoga session)

Here is a RAW crop of the in-studio yoga shoot: 1/100 f/4 ISO3200

I'm totally open to suggestions for software that works with noise, and also open to the idea that my expectations may be high? With the particular image below I addressed some of the noise in LR using the details panel and adjusting luminance and color noise reduction sliders...

 

To be fair most of my work has been done outside so moving into indoor spaces is bringing on a bit of a learning curve.

yogatestcrop.jpg

The picture of the Yoga class actually comes off looking a bit overexposed.

 

If you're new to indoor photography, don't overlook flash, particularly bounce flash. Bounce flash can often be a good answer in a moderate-sized room with relatively low ceilings. (Not so much in the City Council chamber I mentioned earlier.) I think I'd give it a try if I were you.

Bob
Philadelphia, Pennsylvania USA

"Here is a RAW crop of the in-studio yoga shoot: 1/100 f/4 ISO3200"

 

Of course it is difficult to impossible to suggest settings for you without actually seeing the room you are shooting in.

However, I think most folks know how the average school gym is lighted.  I shoot in that environment a lot.  A whole lot !

Do you think your yoga room is similar?

 

My setting for this shot was f4, 1/160 and ISO 1600.  It looks like your yoga room is at least as well lighted.

 

_DX_1399.jpg

 

As Robert offered, your shot does look slightly over exposed.  Maybe not a full stop but it might be.  That would even out our settings if accurate.  This leads me to believe you are doing things correctly.  Just some final tweaking.

 

Always shoot Raw format. Get the exposure as close as you can (forget the other settings) and let the post editing do the rest.  I use LR but you can use the free DPP4 that came with your camera for post.

 

EB
EOS 1DX and 1D Mk IV and less lenses then before!

TCampbell
Elite
Elite

The room is darker than you think.  I've owned a few different "incident" light meters and I've tested the meters and several cameras.  In all cases, camera metering and incident metering actually is _very_ close.  

 

An incident meter is theoretically perfect because it isn't subjective.  It's counting the photons landing on it's sensor.  A "reflected" meter is based on light first bouncing off a subject and then into the metering sensor.  But the reflection is skewed by the variable reflectivity of different objects.  "white" things reflect more light than "black" things ... even if those things are in identical light.  The camera is tuned to try to meter for "middle gray" (it's trying to correctly expose average things like skin tones, plant foliage, etc.   Things which are neither "white" nor "black".   

 

But knowing this... you want to be aware of what metering mode you use (evaluative, center-weighted, spot, etc.) and what you point the camera at to take the meter reading.   If necessary, it is possible to point the meter at something that should give a good "middle gray" value, meter it and use the AE Lock button to "lock" the reading... then point at the subject you want to photograph to take the shot (and the camera will use the locked meter value.  Just be aware that the "locked" meter reading is only locked until either (a) you take the next shot or (b) the camera time-out expires.  You'll notice that when you tap a button to wake up everything... the camera stays in that state for many seconds, and things go dark again.  Once things go dark, it releases the meter lock (it presumes you didn't really want to take that shot after all).   If you want to force it to release the metering lock, you can tape the AF point selection button and that will drop the locked meter reading immediately.

 

Looking at your shots and your exposure... f/4 at 1/10th & ISO 800...   and you were saying you had to use ISO 6000 to use f/5.6 and 1/250th.  So lets do some math...

 

ISO 6000 is 2 2/3rds stops up from ISO 800  (800 -> 1600 -> 3200 -> 6400)

f/5.6 is down 3 stops from f/1.4  (f/1.4 -> f/2 -> f/2.8 -> f/4)

1/250th is about  4 2/3rds stops down from 1/10th  (1/10 -> 1/20 -> 1/40 -> 1/80 -> 1/160 -> 1/320)

 

This tells me the exposure was different by about 1 stop (the shots you took at ISO 6000 f/5.6 & 1/250th) vs these shots you posted here at ISO 800, f/4 & 1/10th.

 

A slight difference in lighting, time-of-day (e.g. how much sun was shining in the window) and the metered subject in the room could all account for that.

 

The camera is pretty good at ISO 6000.  You said the images were "unusable" (I shoot my 5D IV at ISO 6400 and don't even think twice about it.   But you would want to make sure you understand how to use a workflow to process the images.

 

Here's an ISO 6400 shot from my 5D IV that I posted a few days ago.

 

2W0A3663.JPG

 

A human eye is unlikely to be able to detect any noise in this image (all images technically have "noise" ... it's not a matter of "if", it's a matter of "how much" (specifically defined by a signal to noise ratio (SNR)).  

 

You can find the original RAW files here:  https://drive.google.com/open?id=1ZekVErNHUn7NU-RXSZWu5lnuQn--UkYv

 

(that's a Google Drive folder that contains both original RAW and JPEG versions of two images... one at ISO 100, the other at ISO 6400)

 

The trouble with the web posted version is that it's (a) resized and (b) converted to JPEG.  Resampling an image to a smaller size means many pixels get averaged into fewer pixels and this tends to "average out" the noise.  The JPEG conversion results in a bit more de-noising.   I've noticed that when I compare the two JPEG images side by side, I cannot tell the difference.  But when I compare the original RAW files both opened at zoomed up to 100% size (one screen pixel = one image pixel) then I *can* notice a tiny amount of noise in the shadows.  THAT noise is EASILY handled in software (nothing I would ever think twice about).

 

ISO is only applied to an image AFTER the shutter is closed and the exposure is complete.  ISO is not technically part of "exposure" ... even though you'll find lots of books & articles that talk about the "exposure triangle".

 

Anytime you have an insufficient exposure, you get noticeable noise.  You can eliminate noise by getting a sufficient exposure.  But remember that since ISO isn't really part of exposure... boosting ISO doesn't count toward getting "sufficient exposure".  

 

There are ways to deal with noise in images and most photo-processing apps have tools for it (albiet fairly basic tools that tend to try to reduce noise globally).  There are better tools that don't "globally" de-noise... they selectively de-noise based on tonality, use of image masks, or even let you "paint on" the de-noising (only de-noise the spots you select) and they produce a much nicer result.

 

If you post an example of an "unusuable" image as a RAW file somewhere (you wont be able to post it here ... you'd have to put it in a cloud drive somewhere and share the URL) I can show you what can be done with it.    Also, what you *can* do is export the image as a JPEG with no compression and no re-sizing... but then "crop in" on some section you feel has too much noise (so that the cropped area isn't too big) and then post that.  

 

Sometimes what people think of as "noise" is really something else. 

 

 

 

 

Tim Campbell
5D III, 5D IV, 60Da

TCampbell
Elite
Elite

 

The yoga shot is overexposed.  Here's the histogram (I have an EXIF viewer plug-in installed in my Chrome browser that let's me right-click and view this for any image):

 

Screen Shot 2018-04-18 at 11.09.49 AM.png

 

Notice the histogram graph at the bottom shows lots of data all jammed up against the right-side of the graph?  This indicates heavy over-exposure.  But it helps to understand how to read a histogram.

 

The left-right direction of the graph indicates the brightness or tonality where the left side is complete "black" (zero light registering) and the right edge indicates complete "white" (the maximum value that the camera can record.  You can think of it as "100% bright" but the camera is either recording images in 8-bit JPEG or as 14-bit RAW (so the actual values are 0-255 for JPEG and 0-65535 for RAW).

 

The right side of this particular histogram shows lots of data jammed to the right.  The left side shows a thin red band up against the left side of the histogram ... indicating the "red" channel had some under-exposure.  

 

 

 

An easy way to help the image is to increase the light in the room so that the room brightness is nearly as bright as the outdoor brightness.   You can use a flash to do this... bounced off the ceiling (if the ceiling is white).  

 

Ideally you'd take a meter reading through the window to check the exposure.  It's ok that the light "outside" be on the bright side... you just don't want it blown out.  Then use the flash in E-TTL mode and angle the head up to the ceiling.  The flash should determine how much power it needs to correctly expose the shot based on the exposure YOU set on the camera (which is already set so that the outside wont be too bright.)

 

This should help create a more balanced exposure (at least balance w.r.t. light levels.)

 

HDR (High Dynamic Range) is another way to solve the problem... but works best if nothing in the image is moving.  This is because HDR is based on taking multiple images (at different exposures) and then combining the data.  All the exposures should have identical composition with nothing changing from frame to frame ... other than the exposure level.  If anything changed positions between frames it will result in "ghosting" of the HDR output.  (there are techniques to deal with the ghosting problem.)

 

In general, HDR usually works best for capturing a very wide dynamic range when nothing is changing between frames.

 

 

 

Tim Campbell
5D III, 5D IV, 60Da

"Notice the histogram graph at the bottom shows lots of data all jammed up against the right-side of the graph?  This indicates heavy over-exposure."

 

We both know that's not correct. Don't we?  Again an issue where the graph reader isn't coming to right conclusion. A graph that ends before it gets to either side is under or over exposure. Gaps on either end indicate you are missing information and your exposure needs to be shifted to not lose any detail.  His histogram extends completely to each side. A slight over exposure, possibly, or is it just a light colored subject? 

 

HDR for moving subjects may not be the answer but bracketing certainly is.  In some scenes it may be impossible to get a perfect histogram.

EB
EOS 1DX and 1D Mk IV and less lenses then before!


@ebiggs1wrote:

"Notice the histogram graph at the bottom shows lots of data all jammed up against the right-side of the graph?  This indicates heavy over-exposure."

 

We both know that's not correct. Don't we?  Again an issue where the graph reader isn't coming to right conclusion. A graph that ends before it gets to either side is under or over exposure. Gaps on either end indicate you are missing information and your exposure needs to be shifted to not lose any detail.  His histogram extends completely to each side. A slight over exposure, possibly, or is it just a light colored subject? 

 


If the graph is bunched on the right side, there are a *lot* of pixels at 255,255,255 (Histograms are always based on the JPEG) and so have no information and so are overexposed. (It could be you want that overexposure because it is a window or something, but that is why we are photographers, not snapshooters, we make choices.)

"...we are photographers ... we make choices."

 

Amen!

EB
EOS 1DX and 1D Mk IV and less lenses then before!

Ernie, I'm thinking you don't understand how to read and interpret a histogram.

 

There are a few times when I'm ok with a histogram clipping.  

 

1)  If you take a photo that includes the Sun and you aren't using a sufficient filter, there are many situations where (barring the use of a filter or a lot of atmospheric haze, etc.) there really isn't any exposure that wont result in the Sun being blown out.  That's a situation where you just have to accept it.

 

2)  If you take a photo of a predominantly dark (e.g. night) scene, you'll get a lot of shadows that will clip.  Shooting a high enough exposure to prevent the shadows from clipping would result in losing the highlights.

 

But this shot isn't either of those situations ... nor anything similar.

 

In this shot, the histogram consistently gets higher and higher values... until it hits the right edge of the graph and runs out of room. 

 

This indicates a high amount of not just very bright pixels... but clipped data that is unrecoverable.   In this image, the window, the shoulders and a bit of the faces at the leading edges are clipped.  The image appears to be over-exposed by possibly 2/3rds of a stop... possibly a full stop needed to avoid the clipping on the subjects. 

 

Use of a flash is probably the easiest way to fix the problem.  This brings the "indoor" exposure closer to the "outdoor" exposure and reduces the dynamic range needed for the shot.

 

Bracketing just shifts the exposure but doesn't help the dynamic range needed if you want to capture everything without clipping.  HDR can address the dynamic range without clipping but can have ghosting issues if anything moves from frame to frame.

 

 

 

The 5D IV does have a highlight tone priority feature will will try to protect against over-exposure (but can only do so much).  Basically it changes how ISO gain is applied to resist boosting highlights to the point of clipping... but mapping the tonality to preserve relative brightness.

 

Tim Campbell
5D III, 5D IV, 60Da

coachboz68
Enthusiast

Per usual, the answers from the more experienced photographers in this forum are packed with fantastic information! 

 

I am not at their level, so will share from a less informed perspective.  If nothing else, perhaps just to commiserate. 

 

Basically, you are saying that based on reviews, you thought your camera would be able to shoot lower ISO in low-light situations, thereby resulting in a "cleaner" (i.e. less noisy) image.  I've been there.  Oh, have I been there. 

 

I predominately shoot gymnastics for my daugthers at the "club" level, which is to say advanced levels before college.  This results in, to my knowledge, some of the most challenging sports, stop-action, shooting out there. Horrific, I mean horrific, lighting.  Fast moving girls twisting and turning in all directions.  and unifoms that blend in oh-so-perfectly with the both the background and with other gymnasts who are constantly walking in front of, and behind, the subject -- all of which pushes any continuous AF (Servo mode in Canons) to their limit.  

 

FWIW, here are some things I did/do technically and emotionally, in no particular order of importance... 

 

1) Comparison.  At some point, if you're super ruminating like I am, and just can't get it out of your head that your camera isn't performing well, hunt down a friend or fellow photographer and have them shoot one day with you to compare results.  Or, if you want to spend money, rent a different body brand.  Whatever.  Just find a way to get real-world comparisons in front of you instead of manufacturing a comparator in your mind's eye.  The latter just drove me crazy.  When I shot next to a comparable Nikon user, I felt better as I understood the concrete differences.  One of your examples was "f/1.4 1/30th and ISO 2500" ... if you can bring along a friend (or borrow/rent a camera/lens) and shoot those same parameters from the same shoot point, you'll at least know if you're WAAAY out of the ballpark or just minor differences.  

 

2) If point (1) yields the "OK, I know it's at least not BROKEN" level of confidence, just stop comparing altogether and work with what you have.  I just love watching an experienced photographer work with her/his gear.  It's like a true artisan who knows the tool inside and out.  Allows them to focus on the picture, not the mechanics, once they get really good at knowing their gear's characteristics. 

 

3) If you want lower ISO in the yoga shots, work on slowing down your shutter speed and catching the movement at it's natural "stop point" rather than using a faster shutter speed and trying to stop the motion at any point during the movement.  In gymnastics, the example would be a beam routine where the gymnast jupms and splits her legs in midair.  At a point in the jump, the legs are actually motionless as they reach their apex.  And, they slow down considerably as they're nearing the apex.  With some practice, I found I can slow down my shutter speed and nail a good shot more often than not.  The way you describe your yoga shooting, it sounds like you have plenty of opportunities, which means you get to experiment a lot throughout the shoot.  In gymnastics, when it's my daughter I'm shooting, you don't have that luxury!  So I practice a lot with the gymnasts before her.  

 

4) Learn to accept some noise in the picture.  Unless it's for a professional magazine or something, I find that I am WAAAAY more obsessed with noise than the people actually using, or enjoying, the pictures.  Personally, I find Canon's noise more appealing aesthetically than some other brands, but that's pure preference.  In the end, so many athletes just want the great shot and don't even mention or notice the noise in a picture that drives me crazy.  Also, it seems the vast majority of viewing happens on devices with small to moderate screens, or with the image reduced in size to fit on a webpage, and for those, the impact of noise is reduced even further (especially the levels you're talking about, which for the better cameras, really aren't that high).  

 

5) The noise-reduction I use in Lightroom is quite good and while it will soften the image, goes to my point above.  You will care a lot more than most of your viewers, unless you are a professional and they are paying you top dollar for pin-sharp clarity.  And if that's the case, you need to go all-in with proper lighting, off-camera flashes, etc., etc.  

 

6) Given how much I spend on the body, I feel more comfortable with Canon lenses.  I've heard from enough pros that it does matter, and I don't want to leave that kind of thing to chance.  In for a penny, in for a pound, as they say. 

 

Again, I am nowhere near the level of many participants on this thread so remember my opinions are really novice in comparison.  I'd listen to the pros! 🙂 . Just wanted to share my experiences as I have personally dealt with the kind of nagging questions you reported in the initial post.  

 

Good luck! 

 

 

Announcements