A quick note on RAW, EXIF and native converters.
In short: Only original RAW converters understand all camera settings and sensor calibration for 'correct' debayering. The color problem is partly due to the structure of the sensor. IN EXIF a huge amount of data is stored for the correct conversion of RAW to full color image.
Materials on the topic:
- Pixels and Subpixels
- Five things I start processing in Adobe Photoshop Lightroom
- About RAW converters
- Very easy way to sharpen with some Nikon cameras
Comments on this post do not require registration. Anyone can leave a comment. Many different photographic equipment can be found on AliExpress.
Material prepared Arkady Shapoval. Training/Consultations | Youtube | Facebook | Instagram | Twitter | Telegram
I have been shooting for over a year with D300s. The native converter reads the source better, and the Picture Controls are displayed immediately. However, lightroom, although it transmits the color of the source differently and does not see the PC, gives much more opportunities for photo processing. It is a pity that the manufacturer has scored on the software - to equalize its capabilities with the LR, and it will not have a price.
That's it, this looks like your Picture Control. Roughly speaking, these are the same filter settings that the camera applied to your frame to get the desired image. All of these adjustments to brightness, saturation, etc. And the native developer then allows them to play with them on a computer using the same algorithms.
Only adobe camera raw is the fastest and most convenient.
I see no reason to twist my own.
ACR is not optimized for batch operation at all
Well, essentially LR = ACR + batch processing
Yes, there is the same processing module
True, although it is there, but uncomfortable.
However, I still do not process a batch photo, I twist three or five frames at a time, no more. So unnecessarily.
To be honest, you need to pick a question - it is very difficult. But this “extra data” in EXIF is very similar to the settings that the camera used when creating a JPEG file (or TIFF file) for previewing from RAW data. And as far as I came across information (I do not know nikons) - when "developing" with a native converter, it writes metadata with settings directly to the NEF file, while third-party converters store them in some file nearby.
Not that much data is actually collected from the matrix and from the camera. The sensor is covered with a set of filters, the most common of them is this RG1G2B circuit, Fuji plays with his own sets (starting from s5), Canon with 5D4 uses DualPixel - two LEDs instead of one. Each LED, covered with its own filter, gives a number - roughly speaking, what voltage is fixed on it from the incoming color. Our eyes do the same, in a very simplified way.
Each LED under this filter has a certain spectral sensitivity - how strongly it will react to incident light of a certain wavelength, given the filtered light. Therefore, yes, in fact, the resulting image is monochrome - there is no light yet, there is only the response of the LED, which can be associated with some kind of brightness - from 0 (black) to the maximum value (how many there are - 14 bits, it seems, - 16384). In reality, there is not even 0-16384 - but some larger values - with an offset, the middle of the range is 16 bits.
Well, the white balance is collected and calculated in the form of four coefficients for each type of filters. By what values should all the RAW data be multiplied to make white white. In this case, the RAW data itself does not change, by the way, the camera always captures the same way, it does not know how to adapt. Therefore, for a camera, a white sheet of paper illuminated by a yellow lamp will give yellow color (conditionally, the camera does not know how to color, but it will be as yellow as a yellow sheet of paper illuminated by a white lamp). And four factors allow you to adjust the white balance. By and large, these data can also be neglected and a custom white balance can be built, and in Exif, along the way, in addition to the calculated balance, they also pushed the coefficients for all camera presets.
Everything else in the exe is a set of data about the lens, camera, GPS, which focus points worked, which values were collected from the metering sensors. Well, except that the matrices of translation into the XYZ space after demosaicing. The characteristic curve is also, perhaps, because the camera has a linear sensitivity, and we are non-linear. Well, maybe (POSSIBLY) this nagging skintone is Hue Shift. This is the “creep” of the hue with decreasing-increasing brightness, which occurs when superimposing a gamma curve. Artists are familiar with this phenomenon when watercolors are diluted - the color in the mass and the diluted color can have different shades, here it already depends on the converter. But also, as it were, everything plus or minus is considered the same for everyone - all tables are known, nothing new has been invented.
There is one more thing: "the camera has linear sensitivity" - this is the ideal. In real hardware, there are deviations that only the manufacturer of the matrix knows about.
I don't think it's a big problem to measure and plot your own gamma curve for each type of matrix. It's just that in most cases the profile includes only nine-factor conversion matrices from the camera's color space to XYZ and judging by the fact that few manufacturers are soaring about this, the problem is especially acute only among esoteric citizens. Most of those who need “beautiful” (and beautiful is always not reliable, it is always richer, brighter and with a distorted color rendition for a number of reasons - including the fact that in order to convey the sensation from a real picture, its reduced image must act more strongly ) ... so, those who need more beauty will do everything with their hands based on tasks and feelings, adjusting each color separately.
No, just modern Nikon software does not touch RAW (NEF), and writes metadata on the corrections made to a separate file (which is inconvenient).
Interestingly, you are using the Exif analyzer from Jeffrey Friedl. He went headlong into image processing and plugins for such processing. And I remember him as the author of the best (and almost the only) book on regular expressions, this is one of the few technical books that I bought in paper.
Arkady, what do you think about using Phase One?
Not Arkady, but I will say that I like the initial color on C1 with files from D300 much more than in any other software. Less fiddling to get what you want. And fine-tuning the color in any range is also pleasant.
I mainly use native raw converters:
For Nikon: CAPTURE NX-D
For Canon: DPP
For Fuji Proof / Proof: Hyper-Utility HS-V3 Version 1.1b, sometimes Silkypix.
X100T sold while looking at X-T3 / X-T4
For Sigma SD-1 Merrill: Sigma Photo Pro 5.5.3, they say it is better than the later versions, which were sharpened on the Quatra.
I'm currently using Lightroom classic 9.4. When you select the “camera standard” profile in it, the colors in the photo from the nikon d7500 correspond to what I see in the capture nx-d with standard picture control.
Does your monitor allow you to see the difference?
For Nikon with matrices from a Sony camera, in C1, the Sony profiles are good, the curve is standard, and the curve is edited according to the Lla template. All of this can be set to be automatically applied upon import. Sony is clearly better cooperating with C1 than Nikon, Nikon bastards and their software do not saw and do not help others.
only original RAW converters understand all the camera settings = !! = RAW writes the conditions for setting the matrix ... but does not apply them in the construction of the file .. TIF is more convenient for work and also 48 bits ... RAW IS GOOD THAT IS A LEGAL DOCUMENT IN COURT
In fact, most RAWs are 36 or 42 bits, and in TIF, 48 bits are often redundant, because this is already a debayerized image and there is nothing to “pull a picture” from. TIFs 24 bit and 48 bit stretch + - the same, of course, there are situations when this is not the case.
"For every iso ..."
-Well well. In general, tell me from which class of the gymnasium you were kicked out for failure? Sixth? (from)
Arkady, the girl filmed at 70-200 / 2.8 is very pretty :-)
Good afternoon. I accidentally cut off the parolonchik on the DSLR, but gave it up and took pictures further. But now I remembered again for this crap, is it even needed? If so, what to replace it with? And is it okay that I was shooting without a pad?
this is a damper, the mirror did not beat))
Can anyone tell you more? There is no information on the Internet at all (
and what exactly is not clear? when it rises quickly, the mirror hits the limiter with great force and therefore it is necessary to somehow damp it.
I'm wondering what can replace it and how.
Obviously
Another piece of foam rubber
In general, nothing has changed since the time of Prokudin-Gorsky. Three rgb filters give us color
it is in monitors three RGB subpixels give one full color pixel, and on sensors one monochrome subpixel from the RBGB set (two green, one blue, one red) gives one color pixel. The difference is enormous.
Well, just one green one was added. The camera is also software
everything is much more interesting there. The debayering process is quite complex, and involves interpolating adjacent subpixels. without this, the final frame resolution would be much lower. you can see what RAW looks like in reality, before debyer. it's unusual))
No, not just one green was added here. Once again, read carefully: in monitors three RGB subpixels give one pixel, and on the sensors one monochrome RBGB sub-pixel gives one pixel... 3-> 1, and conditionally 1-> 1. It was not added, it is a completely different structure.
For a long time I asked you, boyar, wanted - how do you change the font? Do BB codes work here?
Admin rights
All people are equal, but some are much more equal)) Clearly.
Is it so difficult to introduce formatting and other elements inherent in any forum on the network into the local response form?
And Radozhiva is not a forum, but a blog.
It’s probably not entirely correct to compare the Adobe Standard profile with Nikonovsky. ACR has camera profiles that simulate Picture Controls and these profiles give results close to the original. ADL, of course, goes sideways in third party converters. As an example, the same RAW development in a native carpet:
Photoshop Adobe Standard
Photoshop Camera Standard
In all converters, corrections are "in zeros"
Recently came across a nice article on raw decoding. May be useful to someone https://m.habr.com/ru/post/516228/
I lived peacefully with my canons, there was only one Lightroom for all occasions, sometimes I also shot on Nikon 3200 - and I didn't know grief. Until I suddenly began to feel nostalgic for the color of old CCD matrices and got a D80, like in the old days. And since since that time I have been shooting exclusively in Raw for a long time, then everything surfaced immediately, at the very first development. I carried out a number of very detailed tests with all available converters and was forced to come to a disappointing conclusion: Arkady is absolutely right. Probably, I am already the million-first amateur photographer who made this “discovery” for myself (it would be better if I hadn't, by God). Third-party converters lie shamelessly. Or you can say “interpret”. In some photos it is not so obvious, in some it is very noticeable. With Nikon it is much stronger than with Canon, but they also misinterpret him, although not so noticeably.
I don’t know, I don’t exclude that the same C1 works well with modern cameras, all sorts of fashionable sleepyheads, etc. At least I've heard such reviews. People work, photographers, and don't even bother. But I use old cameras: one ds MII Canon, 40d, Nikon d80. Everything is pretty sad with them. Lightroom is slightly better than C1, although C1 is, of course, a cut higher in processing and quality. Still, the Camera Standard profile is somehow still trying to imitate the original, but still not that. Moreover, the most unpleasant thing about Lightroom is that in general, everything is somehow dirty and pale. C1 is just serit (from the word gray)), and this is the most unpleasant thing. It would seem a trifle, just a small shade. And the meaning of the photo sometimes changes beyond recognition. And a light joyful sunny day can turn into a dull and cold day. And I did not shoot at all.
The best result, as I determined for myself, is given by development in native converters plus Photoshop. Moreover, native converters are also that little thing. If CNX-D is still all right, then DPP is generally a complete horror at the interface and everything else. Well at least there is a main button "transfer photo to Photoshop" You can, in principle, straight away and press.
Lightroom, C1 with the resulting tiffs, of course, also work, but worse. Options, lens base and so on disappear. And the quality is not the same. But Photoshop is great, it's like a fish in water. What is sad, I do not really like this program, it is already too heavy, complex, too much interference in the photo, and I am not a supporter of this approach, life needs to be shot well, and not well drawn. Well, no batch processing, of course. But in general, adding a couple of adjustment layers, sharpness in color contrast is also possible, and the result is very good. But of course, now you can forget about the convenience of simultaneous cataloging, development, complete processing in one universal converter ((
In general, it was a cry from the heart. How did life become more complicated) Interestingly, everyone is still bothering, developing, typhoid, Photoshop? .. Or have they been hamstrung for all this pain long ago?
And a special thanks to Arkady for perhaps the best Russian-language photoblog on the Internet! And youtube channel. I always read and watch with interest.
I use my own developers on my cameras.
For Fuji S3 / S5 Pro - Hyper-Utility Software HS-V3 Version 1.1b
For Sigma SD1 Merrill - Sigma Photo Pro
For 1Ds_mk2 / 1Ds_mk3 - Canon DPP
For cameras - Nikon Capture NX-D
Well, I see, now I do the same. Only two of them, since I don't have fuji and sigma. Two and that is not very convenient, in each everything is different, you always get confused, and four - even more so.
So, the color from the well-known large developers also did not suit?
And where do you bring tiffs to mind? FSH or right there?