Can an iPhone 11 Shoot Raw?

Time to replace your iPhone? Lusting after RAW photography? Too cheap to shell out $400 more for RAW capability (and fancier lenses and processing). Well, entrepreneurs have an alternative for you, a replacement camera app that better utilizes your current iPhone’s capabilities. This post will introduce the Lux Halide camera app and review manual exposure concepts.

Revisions

1. 2023-05-91 Original
2. 2023-05+23 Observations on low light photography

References

1. https://halide.cam/ Developer Lux product website.
2. https://en.wikipedia.org/wiki/Film_speed
3. https://en.wikipedia.org/wiki/Aperture
4. Marques Brownlee Reflects on the Color of My Skin

A non-Pro iPhone Can Shoot Raw?

The Apple Camera App is the Swiss Army knife of cameras. It has a nail cleaner, a wine opener, a can opener, a knife blade, dinky scissors, … you get the idea. Apple Camera App exposes many of the capabilities of the camera and its supporting software including each of the lenses, zoom, movies, portraits, panoramas, slow mo, and other still and video camera capabilities.

Lux Halide takes a different tack, it concentrates on giving you the ability to make a good still image as a starting point for post-shoot processing. Lux designers chose to leave out all the video camera features and in-camera processing and concentrate on the ability to use the camera in auto-exposure and manual modes and to produce camera raw or compressed H.264 images.

Raw or Rendered

Each image begins as light striking the photodetector at the camera’s focal plane. Red, green, and blue photo-receptors record the intensity of each of the three primary colors and colors near them in wavelength. These photo-receptors are continually responding to changes in the light incident on the detector array. When you trip the shutter, the camera electronics freeze the image and transfer it out as red, green, and blue raster images. These are saved to the camera’s memory. Once saved, they are post-processed to make a preview for you. This process is also happening continuously to make a preview image for the camera’s viewfinder as the viewfinder is unable to display the entire dataset that makes up the camera image.

After the image is captured, the camera can optionally compute a compressed image in any of the file formats that particular camera and its supporting software can render. These are the HEIC files you see. They are smaller in size at the expense of reducing exposure gamut in the rendered image and removing spatial detail from the image. Cable News video is an extreme example of lossy compression. Heavy handed compression to get those home shopping channels on the cable results in presenters appearances being wax fruit like lacking texture and shading. The presenter has a wax museum or post mortuary appearance.

By saving and exporting the complete pixel by pixel exposure data, a more detailed and more artfully rendered image is possible than that produced by the camera.

The camera has its built-in notion of skin tone, usually tailored to a particular tribe. YouTube creator Marques Brownlee [4] is always complaining about cameras that render his rich dark chocolate skin tone somewhere between milk chocolate and latte tan.

Each camera has a model of skin tones or perhaps several for different regions of the world. That model makes assumptions about the median skin tone of the people in that region and starts from there. Many cameras render African skin tones much lighter than reality because the internal rendering was designed to render European tribes. Google, for its Pixel phone cameras, has worked hard to enable them to correctly render skin tones for a larger subset of humanity.

RAW Pros and Cons

Raw giveth and raw taketh away, mostly storage and time.

Halide First Impressions

Halide has proven easy to use, actually, easier than Apple’s Camera App. I’m always doing something I don’t want to do with Camera: putting it in portrait, pano, slow mo, or time lapse, for example. Halide’s control design makes unintended control operations much less likely.

Halide exposes the following still camera capabilities.

• Display or hide the pixel intensity histogram
• Enable the self-timer and choose its delay
• Select the wide lens or zoom lens magnification
• Release the shutter
• Browse the camera roll
• Use auto exposure or manual with access to most of the exposure triangle.
• Use auto focus or manual
• Together, manual exposure and manual focus let you determine what is in focus and the depth of the volume that is in good focus.
• Halide makes it easy to coach auto-focus or switch to manual focus.
• Halide allows manual exposure with individual control of sensitivity, aperture, and duration (shutter speed).
• Halide supports portrait mode captures with an alternate user interface.
• Halide will do quick “instant” renderings of a raw image in the field.

Halide Quirks

Live Photos Disables Raw

So far, I’ve found only one, an interaction between the Apple Camera App and the Halide App. The Apple Camera Live Photo on/off setting, when on, disables Halide raw captures. This is because a live photo is actually a 3 second video clip. Selecting Apple Photos Live Photos on puts the camera in video mode. Halide is unable to turn off video mode to make raw captures.

Apple Low Light Not Supported

Apple has a computational photography, similar to HDR image combination, that gives the Apple Camera App an extended low light capability. Basically, the camera makes a sequence of noisy low light photographs, aligns them and combines them to create low light images not directly possible in the absence of the low light technique. Halide does not use the Apple low light technique.

The Exposure Triangle

The exposure triangle is a conceptual model that can be used to think about manual exposure of photographs. Camera sensitivity, lens aperture, and exposure duration form a set of values for specifying an image capture. Of the three, any two can be independent variables. Once chosen, they determine the value of the third that gives proper exposure. And, you can deliberately overexpose or under expose an image for artistic effect. I prefer to make a good capture and manipulate exposure in post.

The triangle is a good model for ISO, aperture, and shutter speed. You can choose any two but, but to close the triangle (achieve a proper exposure), the third is determined.

For general shooting, you pick ISO based on the amount of light available. The ISO setting describes the sensor’s sensitivity to light. A larger ISO indicates a more sensitive image sensor.

ISO choice determines the working shutter speed range. Working with moving subjects, the photographer picks a high ISO speed. Working with bright subjects, a low ISO speed is chosen. Generally, dim lighting requires a large ISO value.

Choice of ISO influences the noise present in the captured image. At high ISO values, each pixel of the camera sensor captures less photons during the exposure making the thermal noise in the sensor a larger fraction of the signal captured. As ISO setting is increased, image becomes increasingly grainy with off-color and off-intensity pixels. For most shooting, pick a low to middling ISO in the 100-1600 range. (Or let Auto pick ISO).

From that working range established by the ISO sensitivity choice, then pick either a shutter speed to prevent motion blur, or an aperture to control depth of field or control shadow or highlight detail revealed. But you can only pick two of the three. Once ISO is picked, you can pick either shutter speed or aperture but not both. Shutter speed is expressed as a number, usually the reciprocal of the time the shutter is open. If the shutter is open 1/100 second, the shutter speed is shown as 100. Bigger numbers are less time open. Shutter speeds are related by a power of 2.

The lens aperture and shutter open time control the amount of light that falls on the sensor. The sensor’s sensitivity determines its response to that illumination. The f-stop number describes the lens opening in a standardized way. An aperture of F-8 gives the same exposure to the sensor independent of the lens focal length. F-numbers are determined from lens diameter and focal length in a way that each “stop” is related related to its neighbors by a factor of 2. F-11 passes twice the light of F-16. F-8 passes 4 times the light of F-16. F-22 passes half the light of F-16.

Adjacent shutter speeds are also related by a factor of 2 and are given in fractions of a second. The F number and shutter speed combinations that give an acceptable exposure are related to the sensitivity. This powers of 2 relationship makes it easy to adjust shutter speed and aperture as a one step change of either has the same effect on the light captured by the sensor. So moving 1 step faster requires opening the lens by one stop, say F-16 to F-11 to maintain exposure.

The sensitivity is given in ISO numbers defined in such a way that that a sunny day full sun exposure is correct at F-16 and 1/ISO seconds. This made it easy to set exposure outdoors without a meter. Using ISO 100 black and white film, a proper exposure would be obtained in daylight using F-16 and 1/100 second. Film speeds of 3o, 100, and 400 were common back in the days of halide photography.

An explanation of the choice of exposure for sunrise/sunset photography, etc is outside the scope of this note. Most of today’s cameras have exposure programs that handle these conditions correctly when the camera is told to use the proper one.

In practice, we let Auto find the proper exposure, switch to manual, and change the F-stop to lighten or darken the image in a way that we find “creative”. In practice, I almost always use auto exposure with iPhone cameras. The phone screen makes it difficult to judge shadow and highlight detail. This is much better done using an eye-to-viewfinder camera. Mirrorless cameras with their electronic finders are quite good for judging manual exposures in the finder (not the rear screen).

F-stop and depth of field are related. F-8 gives a good depth of field. Changing to F-16 or F22 gives more. Changing to F5.6 or F3.2 or F2 or F1.4 if you have a really “fast” lens gives less. F8 is a good general aperture for aperture-preferred shooting. Set shutter speed for proper exposure of the scene.

The depth of field is the “thickness” of the region that is in acceptable focus. Smaller F-stops give less, larger give more. Low depth of field images deliberately blur the foreground and background to emphasize the content in the zone in focus. Pick the aperture for the depth of field result and the shutter speed is determined.

Rendering Raw Images

I use Skylum’s Luminar Neo for most raw photo image rendering. Neo uses modern computational photography and statistical techniques (“machine learning”) to make an acceptable rendering of a captured image.

An Example

This image of my neighbor’s tree, the one Dismal Manor Gang calls Spooky Tree, appears commonly in Dismal Manor Gang’s mastodon.online musings. Halide took a iPhone 11 RAW image of Spooky on a gray Saturday morning. I’ve rendered the image in Luminar Neo and again in Apple Photos App.

To make the images below, I began with a Saturday morning snapshot. I made 2 copies of the Halide raw image. I used Luminar Neo to render one and Apple Photos the other. On the left, Luminar Neo. On the right, Apple Photos.

When I made the Neo version, I used one of the presets, likely Dynamic Result or Forest Stream. Dynamic Result is designed to render overcast images that have a textured sky. Forest Stream is designed to render images in open shade on sunny days but works well for general use.

I used Apple Photos magic wand to run the Auto feature of each rendering tool. Apple Photos Auto has gotten pretty good at rendering outdoor scenes.

As you can see, the two images have a similar overcast spring day feel but each has emphasized different aspects of the image. Grass color and detail differs. Photos makes the lawn look especially shaggy. Luminar Neo makes Spooky Tree look a bit healthier, greener, than Photos.