Why use ACR vs Photoshop

walkerlangley

Ok, I've tried to find the answer to this but can't. So far, whenever I open an image in ACR, I just click on the "Open Image" tab to open the image in Photoshop. Once there, I do all my post processing. My question is, why do some people suggest doing things like white balance, exposure, and sharpening in ACR BEFORE importing into Photoshop? From what I have read, the only reason I can figure is because you're working on the actual RAW file itself instead of a RAW file that's been converted to a jpeg or tiff. Is this correct? Is it really that much better to do "the basics" in ACR and only do things like layer masks in Photoshop?

Any help will be greatly appreciated.

Walker

colourbox

walkerlangley wrote:

My question is, why do some people suggest doing things like white balance, exposure, and sharpening in ACR BEFORE importing into Photoshop?...Is it really that much better to do "the basics" in ACR and only do things like layer masks in Photoshop?

The important thing to understand is that when converted to Photoshop, the process of going from raw to RGB means certain decisions are being made permanent and certain limits are being defined, set in stone. Take white balance, for instance. Once the raw has been converted to Photoshop, white balance is locked down. If you need to make a large adjustment, you will hit the walls that the white balance was bricked into during the conversion and the range of possible adjustment (that actually looks good) will be limited. When still in raw, you can get at all the original data, like undeveloped film, before it was locked down to specific RGB values, and have your full range of adjustment. It's a similar story with highlights and shadows. Once you lock the tones into RGB values during the conversion, if you discover you need to recover highlights you can't go back and do as successful of a job as you could in raw because the full range of data was cut down during the conversion.

If you aren't picky, or you're a very precise shooter therefore you never need to make big moves, then you probably don't need to shoot raw at all: Just shoot perfectly exposed JPEGs and save a ton of disk space. If you choose to shoot raw, there can be benefits to getting the basics right before you cut down the data during the conversion.

Sharpening is an entirely different subject. It needs to be done different ways at each stage of production (capture, edit, output). Not all in raw, and not all in Photoshop.

Richard

In addition to all of the good information above, certain adjustments are simply easier in ACR, especially white balance and exposure tweaks. Noise is easier to control. You also can do two or more versions (one for highlights, one for shadows) and combine them in PS. Finally, keep in mind that ACR settings are non-destructive, that is, they don't change the contents of the RAW file. I usually process a group of pics in two stages--first getting the adjustments in ACR and second, tweaking in PS. I often find myself making changes to the ACR settings when I have a second look.

arodney

Photoshop is a pixel editor. You alter existing RGB (or other color models) values.
LR and ACR are metadata editors. And they are primarily designed for creating virgin pixels FROM Raw data using instructions (metadata) that you create when you alter the various tools.

With Photoshop, you work with the data as presented (in that color space, bit depth etc) and again, you alter (destructively) the pixel values. In LR and ACR, Raw data is taken along with the existing metadata instructions and the two are used to render, build, the actual pixels. When you take existing rendered images (lets say a TIFF in sRGB in 8-bit), that original data is never touched, its a source for new pixels like the Raw, but it is not non destructive! The original data is copied and converted to a unique colorspace, bit depth and linear encoded processing space (due to the fact that again, these are primarily Raw converters). The original sRGB file is left alone but a new file is created from this data, it can be in a differing color space and bit depth FROM this internal Raw processing color space depending on the export presets. That’s the differences under the hoods.

The other differences are the toolsets and how each product was designed for an accomplished task. Treating LR or ACR as pixel editors is not a good idea (although both have limited local adjustment tools). Treating Photoshop as if it worked like LR/ACR isn’t going to fly either and there are simply tasks you cannot easily cover in Photoshop that you can in LR (thanks again to the speed and ease of coping and pasting metadata). Think of taking a single image that needs white balance adjustment, moving a slider to do so, then pasting that “correction” on 200 other similar images. That in LR takes seconds. That in Photoshop takes a lot, lot longer (just imagine opening 200 21 megapixel files just to start the process).

ziggy53

What a lot of folks don't understand is that most digital cameras don't collect full color or full luminance information from each pixel site. When you capture an image with a Bayer mosaic chip, each photosite has a colored filter over it, Red, Green or Blue. When you process the image from RAW into RGB full-color (this process occurs in-camera if you shoot to JPG files, and it occurs in software if you shoot to RAW files), the color and luminance information is "built" using a complicated algorithm, primarily the "Bayer demosaicing" algorithm. Applying the WB and color tint corrections at this point is most appropriate and should yield the highest image quality.

The other thing is that an imager records the scene in a linear fashion, but the output format needs to be logarithmic. If you didn't perform this operation the image would appear very dark. Again this is a form of interpolation and the power of a full computer can usually beat the dedicated processor in a camera in terms of data accuracy. Better data accuracy results in a more accurate image.

Likewise, since the demosaicing operation is a type of interpolation, upres and downres operations are potentially better performed at this stage than performed later. This assumes that interpolation is performed at the same time as the demosaic operation.

Default or improper settings of the RAW processor will cause some image data to be discarded before you transfer the image to the image processor, making further corrections problematic or impossible.

By taking charge of all the controls of a RAW processing software you retain potentially more control over the ultimate quality of the resulting image, and you provide better input data to the image processing software, like Photoshop.

walkerlangley

First off thanks for all the replies. You guys obviously know a LOT more about the inner workings of these programs than I ever will. The thing I like about using PS to set white balance is that I can use the threshold adjustment to set black and white points, which gives a more accurate white balance than just sliding a slider or picking a neutral grey color with the eyedropper. Is there a way to get an accurate WB in ACR or do you just have to use your own judgement?

I understand the idea of ACR "building" new pixels from the metadata while PS "alters" the pixels, but in PS, I do everything in layers, so I can always go back and re-do or get rid of any adjustments I've made. I guess my main questions is

"is it worth doing WB, hue, saturation, etc in ACR or should I just stay with PS? Can you really tell the difference in the final print between a photo processed in ACR and one in PS?"

Again, thanks for all the help.

Walker

arodney

walkerlangley wrote:

I understand the idea of ACR "building" new pixels from the metadata while PS "alters" the pixels, but in PS, I do everything in layers, so I can always go back and re-do or get rid of any adjustments I've made.

Yes but at the point you flatten (to print) or save off a flattened version, then the edits are applied and the damage results.

Plus in LR/ACR you have a history of every edit that exists no matter how often you quit (unlike Photoshop).

If you have Raw data, use ACR or LR of course.
IF you have camera generated JPEGs, you might want to do as much as you can in LR or ACR but its not the same as #1. After you pretty much exhaust the global and near global tone and color work, when you need precise pixel editing, move into Photoshop. Stay there and end metadata editing from that integration forward.

LiquidAir

The sensor in your camera is a linear device. What linear means is that 20 means twice as much light and 10, 200 means twice as much light as 100 and 2000 means twice is much light as 1000. However our eyes are logrithmic which is why in photorgaphy we usally work in power of 2 stops, (2, 4, 8, 16, 32...) rather than in linear light levels.

When you convert a RAW file to a color space to work in Photoshop, one of the things that happens is that a gamma curve is applied to it which approximates our logarithmic perception. When you open a RAW file, ACR applies this curve for you automatically when rendering it on the screen beucase displaying the linear data looks very flat and nothing like what you saw.

White balance and exposure compensation are things that can only be computed correctly on linear color channels. Once the gamma curve is applied to the file any adjustment you do will only get some of the luminance range right. For instance, setting the white point in the highlights will not properly correct the mid tones and the shadows. Because of this, basic white balance and exposure compensation should definitely be done during the RAW conversion (in ACR or your choice of RAW converters).

As for other corrections, major adjustments done on the RAW file will be somewhat less noisy if done in ACR. Whether that impacts you depends a lot on the kind of image and the details of the adjustment. In particular, shadow and highlight recovery can be done better on the RAW file than on the converted file. Personally I get the image as close as where I want it to be before I go to Photoshop. However, for many kinds of shooting just correcting the white balance and exposure before going to Photoshop works just fine.

Richard

walkerlangley wrote:

The thing I like about using PS to set white balance is that I can use the threshold adjustment to set black and white points, which gives a more accurate white balance than just sliding a slider or picking a neutral grey color with the eyedropper. Is there a way to get an accurate WB in ACR or do you just have to use your own judgement?

Just to clarify: white point and white balance are different animals. What you describe is setting the white point. While balance really means making sure that neutral objects (anything between pure white and pure black) do not have any color cast to them, that they are true grays. ACR has temperature and tint sliders to adjust colors along blue/yellow and green/magenta dimensions. It also has a dropper tool that lets you click on a spot you know should be neutral and it will adjust temp and tint automatically. While possible, it requires a great deal more skill and knowledge to accomplish this in PS.