No you should not do anything differently. ETTR is to get as much detail from the sensor into the RAW file. This is a flat scene with detail ranging from the dark green of the left hand tree to the sky in the histogram realistically 1/2 stop of detail in the dark areas would not give you anything that you would be able to see. If the scene was high contrast then the histogram would be clipped in either the left hand (dark tones) or right hand (highlights).
ETTR Raw files frequently look too light and awful, “correct exposure” gives a range of tones across the histogram and only if the dynamic range of the sensor is exceeded do you get clipping, RAW is the negative. ETTR is a principle to get the most detail in the file and only that, it doesn’t mean you are going to get a pleasing result. It’s up to you to choose to balance the contrast, darken the sky, lighten areas to lead the eyes around the frame. Frequently this means discarding detail, ETTR means you have as much digital detail possible that you can then throw away as required.
Next point. Go to your pharmacy with a memory card and make a hard copy, smallest size will do. Your camera is capable of a
16x11 inch print at 300 dpi, which is magazine cover quality.
30x20 inch print at 150 dpi, normal magazine pages
66x44 inches at front page of newspaper quality
get a print and look at it in your hand to align what this means in reality, what you are going to be able to get in your hands. Even with a magnifying glass you are not going to see that grain at iso 200 on anything under 16x11. Looking at a screen gives a false impression, like looking at tap water under a microscope, focus more on what you are going to get at the end. If your pictures are mainly going to be on the web, you’ll be scaling them down so grain is not an issue.
There isn’t an issue with sharpness at this aperture and any scale you are likely to reproduce the photos on the web or in the hand.
According to that page I did make a mistake in comparing how many times more light one aperture diameter captures compared to another [being (x/y)^2 instead of x/y], but not the general principal that aperture diameter (FL/FS) plays a large part determining how much light is captured.
If you spend seventy quintillion dollars on the latest and greatest you’ll still find noise when blowing the image up to 200%. Take more photos with what you have.
RT or any RAW processor gives a “better” image. That should be the end goal.
I don’t see anything in the water that really qualifies as noise. The image is well-exposed, no upper-end saturation, so it can be easily worked.
Your camera is fairly equivalent to the Nikon D7000, which is what I used until a couple of years ago. What got me going to upgrade is I think some of the same angst you’re experiencing, pushing the limits of a ~16MP APS-C sensor. They can only cram so many pixels in that sensor area, and they can only be so large, so IMHO the trades in dynamic range and resolution really start to show in subjects like landscapes. I still use my D7000 for scenes where the subjects are dominant and the destination renditions are for “small” things like web viewing. The denoising and sharpening alternatives available in the FOSS softwares are good, so you can still get clean images from that class of camera, you just need to become familiar with their dynamics and limitations.
Here’s my proof rendition of your raw from my hack software. The low dynamic range needed a solid S-curve to spread the data around, and I added just a touch of color saturation because I just felt like it…
That applies specifically to astrophotography where the subject isn’t filling the frame. If you increase the focal length, the subject fills the frame more, so more light is gathered from that subject.
It also applies to wildlife photography where you need to crop, for the same reasons.
It doesn’t apply to general purpose photography, though, because we do things on a per-full-image basis, not a per-degree basis. As you increase focal length, you gather more light per degree of subject but you have less subject in the frame.
Totally off-topic; the thread title reminded me of when my wife joined me overseas, literally, at Kwajalein Atoll in the Pacific. Laying in bed that night, she says, “What’s that horrid roaring sound?” I replied, “The ocean…”
Hm, well he has an abundance of articles, and uses this example of aperture diameter in many, not just astrophotography. However looking more closely, it does appear to be mentioned mostly in situations of low light. In those situations the subject does not have much light, therefore the light collecting ability of the lens (aperture diameter) becomes very important. But probably in scenes with a lot of light, its not so important, as it is much easier to saturate the pixel.
Seems like you could have focussed a tad further, as the background is slightly out of focus and you should have lots of DOF. You could have probably got away with slightly more wide open aperture and faster shutter speed hand held to get down to ISO 100, but too much seems like you’d have camera shake from too slow a handheld shutter or foreground or background slight blur that becomes distracting, as well as exposing lens weakness. Given ambient light levels and apparent lens performance, ideally, you’d want to shoot this on a tripod if you were after the best IQ results possible.
One other option in situations like this would be to shoot a burst, and use something like Kandao Raw+ to get an AI stacked raw, for less noise. Free, but not open source.
If you read the article that you linked, under the very heading you suggested, here’s what it says:
Technical note. The lens aperture area times the subject angular area is called the Etendue . Etendue is is a key parameter used in designing optical systems, including cameras for spacecraft, aircraft, lab or field use. While the term is not known in the general photography community, Etendue describes the basic physics of light collection by an optical system and is key to distinguishing what is true from what are myths in the photography community. Etendue is also called the A Omega product, where A is the lens aperture area, and Omega is the subject angular area. For example, the Moon is one-half degree in diameter, so if the moon where the subject, the subject would be about 0.2 square degree, then Omega = 0.2 square degree. Etendue, combined with lens transmission, sensor quantum efficiency and exposure time can be used to measure absolute light levels with an imaging system, including digital cameras. Or it can be used to predict signal levels to set exposure times, e.g. by an orbiting spacecraft, or to compute the integration time to achieve a specific signal-to-noise ratio on a galaxy with your telephoto lens and consumer digital camera.
You cannot just use the aperture size itself, you need to also take into account the subject angular area.
If you’re cropping, then the subject angular area is constant, and a longer focal length at the same f-number improves light gathering.
If you’re shooting a landscape and your subject is “whatever your lens is facing” then the angular area goes up with wider lenses, such that the etendue remains constant for a given f-number.
This may be a basic question, but how can you tell that it is only a half-stop shy of proper ETTR? To me it looks like about 1/3 is remaining to the right of the green histogram?
I believe I was shooting at ISO 250 because I didn’t have a tripod and was trying to keep the shutter speed at at least 1/40 to avoid any blurring. For this type of landscape I like to try and keep as much in focus as possible so I usually try to shoot around f/8 - f/10 for DOF. Maybe this is not a good technique?
I think this makes sense if you think about “zooming in” on the area so you don’t have to crop, therefore a larger part of the sensor would be capturing light for the desired area; is that a correct way to understand this?
Wouldn’t a newer APS-C sensor have the same problem in terms of the sensor area (or even worse), since it’s the same size sensor still but now has even more pixels (e.g. 24MP)? Or do you mean that you can increase the photo to a higher resolution because of the additional pixels before noticing this type of noise?
Well, it sure is mighty confusing to read him bang on about the importance of focal length and aperture diameter in collecting light every article (see the bookcase example: https://clarkvision.com/articles/does.pixel.size.matter/index.html#etendue) only to find out its not that important. But what you’re saying does match my experience punching values into spreadsheets. I would read that focal length was supposed to have a dramatic effect, change the value, and see no change, while things like f-stop and pixel pitch would cause change. So I thought my spreadsheet must be wrong, but as it turns out, maybe it isn’t. You might have helped me understand something.
The histogram is linear on the horizontal axis, so that is just over 1/2 stop shy of clipping the green channel, 1 stop shy of clipping the blue channel, and 1 2/3 stops shy of clipping the red channel.
For better noise performance, most older-sensor Canons have better dynamic range at 100 and then multiples of 160. Unfortunately photonstophotos only has full stops listed for the 50D, though, so you’ll have to try it yourself by measuring the raw dark noise and comparing against white clipping points at various ISO.
But if it’s true that multiples of 160 perform better, then you need to make sure to bump ISO up to 320 in that situation. The noise performance vs 250, which is on the worse end of the DR kinks, may be noticeable.
I THINK it now has support for outputting a DNG that is after align-and-merge but prior to demosaic/tonemap/etc. I’m not sure as I have a hackish fork and I haven’t used it in a while. If there’s interest and there is still no DNG output capability, I’ll post my hackish patch.
Alternatively, for this scene when tripod-mounted, traditional bracketing combined with HDRMerge would provide far more DR improvement with fewer shots. Unfortunately, with most cameras, there’s enough deadtime between shots that the lack of rotational correction in Tim’s HDR+ implementation is problematic. Although since there’s no in-scene subject movement, one might be able to do:
Burst shots
Develop all to linear TIFFs
Feed those to hugin’s align_image_stack and hugin_stacker
