Michael C

There are some cameras, especially in the distant past (digitally speaking - say 2012 or before) that were so much slower in their transfer rate than theoretical limits for the specific SD card format they used that it would not surprise me if they were using SPI instead of SDIO. Maybe for royalty issues? If the patents on SPI had expired, but not for SDIO?

@misk94555 Well, if you're wanting to use the card in a camera, I think it would depend upon both if the card only uses one or the other for the Wi-Fi function. But, yes, I misunderstood the angle of your last question.

@misk94555 "Do the SD/MMC WiFi cards work for the SPI interface?" The answer to that might be camera specific on a model by model basis.

@misk94555 I've never looked for one, but I can't recall ever seeing one advertised or reviewed, either.

@EuriPinhollow Raw files, on the other hand, have no such single correct rendering. There can be an almost infinite number of possible renderings that are all equally legitimate expressions of some of the information contained in the raw file. That's the difference.

@EuriPinhollow Because other raster image formats have precisely defined ways that each bit should be displayed on the screen. A properly encoded JPEG has a single correct rendering. A properly encoded PNG has a single correct rendering. so do RGB TIFFs, HEIFs, etc.

@EuriPinhollow The point is, you never actually look at 'THE RAW IMAGE'. You're either looking at a preview of one specific interpretation among a near infinite number of possible interpretations of the data, or you're looking at an exported version of one specific interpretation among a near infinite...

@EuriPinhollow Adapting older Nikon F-mount lenses to even Nikon's Z-mount can be problematic. The Nikon adapters are only "guaranteed" to work fully with AF-P, AF-I and AF-S lenses. AF lenses will not autofocus when using the adapters because they lack a coupling for the screw drive receiver from the AF motor in the body required for AF lenses and none of the mirrorless bodies have screw drive motors in them.

@EuriPinhollow What it "suggest" to me was that the reason "!" was there was maybe you were saying a skylight filter should be used with daylight? That's why I asked what you meant.

@EuriPinhollow I'm not arguing against anything in the answer. You are.

@EuriPinhollow Not everyone who does photography codes. I don't. I have a strong background in advanced mathematics, and I've never seen that symbol used for ≠.

@xenoid 1080p TV is barely 2 megapixels.

@uhoh I'm guessing off the cuff that the technological advancement was in the area of servo motors precise and smooth enough to move both axes of the alt-az mounts in a way that made them functionally equivalent to equatorial mounts moving only one axis.

@OnBreak. Yes, it is just the same old questions over and over and over ad nauseam. No one wants to point to existing duplicates because then they couldn't answer it again... and again... and again...

@EuriPinhollow So a poor implementation of a raw convertor means the CFA design is flawed? Is this a camera generated JPEG? Or a raw file converted using Adobe products, which do an abysmal job with X-trans raw files?

@EuriPinhollow What makes you think that my suggestion that you could make a boatload of money if what you say is true means that I don't want to discuss it?

@EuriPinhollow If you can create a more efficient color sensor, you're free to do so. If it is so obviously superior as you seem to think, you'd put Canon, Sony, Nikon, etc., and, yes, Fuji, out of business in no time flat.

@EuriPinhollow Sure there is. If the distribution of wavelengths in "natural" light is strongest in the "green" section then making half the sensels most sensitive to green increases the overall sensitivity of the sensor compared to making only 1/3 of the CFA most transmissive to "green".

Cameras are designed to emulate human vision so that they can produce images that can fool human vision into thinking the colors are the same as the original thing being imaged.

@EuriPinhollow Huma nvision evolved to be more sensitive to "green" wavelengths because there is more "green" light in sunlight that has reached the Earth's surface, and plants are thus "green" to make efficient use of the energy from that light. Humans evolved to distinguish slight differences in "green" plant to tell good food from poisonous food.

@EuriPinhollow Having more green filtered photosites (which is not exactly the same thing as having "green pixels", since sensor don't have pixels, they have sensels a/k/a photosites) is due to the distribution of wavelengths in sunlight after having been filtered by the Earth's atmosphere.

@EuriPinhollow Probably because most raw convertors aren't compatible with non-Bayer masked color sensors. X-Trans wasn't a gimmick. There are use cases where it is a benefit. There are other use cases where it is a detriment, just like practically every other informed design decision when it comes to photographic equipment.

But even then all is not equal. Because the human retina has fewer S-cones (short wavelength "blue" cones than L-cones (long wavelength "red" cones), the R and B filtered sensels are given different, not equal, weighting.

There are so many different demosaicing algorithms as to be nearly countless. But at the basic level, a Bayer sensor with 4 million photosites will have 2 million filtered with "green", 1 million filtered with "blue" (that is actually more violet than blue), and 1 million filtered with "red" (which is actually yellow-orange, though the "red" filters tend to vary more from one brand to the next that the other two). The output is 4 million pixels (one for each photosite) that all have an R, G, and B value. The green filtered sensels are represented by an RGB pixel. So are the "red" and "blue".

Photography started out with natural light sources, all of which burn very closely along the CT axis. More modern light sources do not. CT was already embedded in the entire system before light sources significantly off the CT axis began to appear.

@AntonShepelev You're totally incorrect about interpolation. There is both color interpolation and spacial interpolation, but they are not the same thing. An RGB value is interpolated for every photosite. The photosites with a "red" filter in front of them, which is actually most transmissive somewhere between yellow and orange (580-600nm) for the vast majority of digital cameras, must be interpolated and transformed to "Red" (640nm) that is used in RGB color spaces.

@AntonShepelev RE:#3 Anyone who thinks Kodak's Velvia daylight film gives the same color as Fuji's Provia daylight film needs to have their eyes examined. Ditto for Ektrachrome vs. Kodachrome. Both come in daylight emulsions, but their color response is far from the same. Just because a film is labeled "daylight" or "tungsten", or anything else does not mean it has a scientifically standardized color response curve.

@AntonShepelev Re; #2: two light sources can have the same color temperature and have vastly different spectral composition. That's one of the main reasons WB correction is needed. A 3700K old style fluorescent light and a 3700K LED light at a club are nowhere near the same color. One is biased heavily towards green, the other heavily towards magenta. The green←→magenta axis is more or less orthogonal to the blue←→amber CT axis.

@AntonShepelev "r 1.0 1.0 1.0 1.0" is not doing what you seem to think it does. It still must interpolate to correct for the difference between the colors of the filters in the CFA and the colors used in color reproduction systems, such as RGB or CMYK. Again, the colors of CFAs are nowhere near a match for the target colors of RGB emitting devices, much less CMYK printing systems.

Start by looking at the two existing questions and their answers linked above. Then look at What are the pros and cons of different Bayer demosaicing algorithms?. Much of it is all already here and does not need to be repeated, especially in comments.

You keep saying "I know..." followed by incorrect information. I do not think "I know..." means what you think it means.

Again, color separation, if that's what you want to inaccurately call the difference between how sensels behind different color filters record the same light, is NOT the same thing as White Balance.

Interpolation is necessary to produce anything that remotely produces a perception of color by the viewer that is in any way similar to the perception of color the viewer sees when they look at the actual scene that was imaged. Please see What does an unprocessed RAW file look like?

CFAs do NOT separate colors (or even wavelengths). Not at all. That's not how they work. Not at all. They attenuate some wavelengths more than others. But if the light entering the camera is equally distributed full spectrum light sufficiently bright to take a normal image, or even natural sunlight (which is stronger in some parts of the visible spectrum than other parts of the visible spectrum), every sensel behind all three of the CFA filters will receive some light from all of the visible spectrum. CFAs do NOT separate colors. Digital image processing does.

Nowhere above have I said raw data does not contain color information. You can stop arguing against that. But color information that needs to be interpolated from monochromatic luminance values is not the same thing as White Balance.

In other words, dye couplers determine the color space of the developed image. Just as the color space we choose after the fact (when the analog raw energy values collected by the sensor are converted to digital information and processed in various ways) determines what colors we can display in a image after collecting far more information than any current color space can reproduce.

Color films have three monochromatic layers that have sensitivities tuned to either the lower, medium, or higher wavelengths of the "visible spectrum". Each emulsion layer has a dye coupler that is not visible until activated by the chemicals used in development. The color of each of the dye couplers determines a color film's "white balance". Dye couplers are the equivalent of the specific colors emitted by monitors for each of the three color channels. The CFA in front of the sensor is not the equivalent of a set of three dye couplers.

Where, exactly, in nature? Color does not exist in nature. Color is a product of the perception of electromagnetic radiation by a vision system (such as the human eye/brain system). Even if one is talking about spectral distribution and the relative values of various wavelengths the source can vary significantly, even from one black body radiator to another. Even when the source is Sol, the location of the observer and the angle of the sun in the sky will affect the spectral distribution of the light falling on the scene and the colors the observer perceives under that light.

Look at it this way. When we shot with B&W film, putting a red filter in front of the lens did not create a red photo. It created a photo where red objects in the scene were brighter in the photo than similarly bright green or blue objects.

Just because there is a difference in the sensitivity to various wavelengths of light depending on whether there is a violet-blue filter, a slightly yellow green, or a yellow-orange filter in front of a sensel does not mean the sensor has a WB. It has a sensitivity curve. But that's not the same thing as WB. That's not even close to the same thing.

Only a spectrometer can discriminate between specific wavelengths of light. The CFAs in front of sensors do not prevent any visible wavelength from passing through them, they only reduce some wavelengths more than others. And our cameras and RGB color reproduction systems are not capable of reproducing many, if any spectral colors. EVERYTHING within sRGB, AdobeRGB, CMYK, etc. are not spectral colors. They're simply not capable of reproducing them. ProPhotoRGB is the only common color space that can reproduce some spectral colors.

Please see RAW files store 3 colors per pixel, or only one?. Yes, the single luminance values recorded by each photosite/sensel (only a digital image file actually has pixels), but none of them discriminate between the light that makes it past the filters and into the photosite's well. Even if there is a green filter in front of a sensel, the "blue" and "red" light that makes it past the filter is recorded as energy, just the same as the "green" light that makes it through. It's just that more green is allowed to pass than red or blue.

Each sensor design (and even individual copy, due to manufacturing variances) will have varying different sensitivities to different wavelengths even before the CFA is put in front of it. The color and strength (density) of the CFA will also have an effect. But that does not mean the sensor has a WB. It doesn't. The light sources it records have WB (spectral emission/color rendering index). Sensors do not.

The reason WB settings have no influence on raw files is because raw information is monochrome. There are no red, green, or blue channels in raw data. The filters in front of the sensor do not block all of the light from other parts of the visible spectrum, they only attenuate it to a greater or lesser degree depending on the wavelength of light and the color of the filter. The "red", "green", and "blue" filters are not even the same "Red", "Green", and "Blue" colors emitted by RGB color reproduction systems (monitors).

To be more precise, 135mm format film uses 36mm + 2mm gap for a total of 38mm/frame moving horizontally through the camera with exactly 8 perforations per frame. 35mm movie film used 16mm + 3mm gap for a total of 19mm/frame moving vertically through the camera with exactly 4 perforations per frame.

@MarkMorganLloyd NASA doesn't run America. NASA does NASA, the rest of America does whatever they want. You're taking a very localized specific example and insisting on stretching it to a Universal Truth for a culture in which you've never lived. Surely even you can see the problem with that? I mean surely the only way anyone in all of Europe travels anywhere is via a "Tube", right? I mean they have the Underground in London and all, so ALL Europeans ONLY travel anywhere via underground railroads?

Cine film has always been metric, or at least the perforations have always been exactly 4.75mm apart/8 holes per 38mm (and not 8 holes per 38.1mm/1.5 inches).

What does NASA have to do with Kodak, other than they bought film from them (for 70mm Hasselblad cameras)?

Large format view cameras that had been around since the 1800s used Imperial measurements for paper sizes and even lens focal lengths well into the 20th century, but even at the same time the same photographers using Imperial units for LF used mm for focal length of their 135 format ("35mm film") lenses.