A Road to Sharpness

[TonyW]

This may be a bit off the track of the original post, but I normally don't do anything significant in ACR and rely on PS for all my editing. Is there anything I could do in ACR that I can't do in PS?

[xpatUSA]

This may be a bit off the track of the original post, but I normally don't do anything significant in ACR and rely on PS for all my editing. Is there anything I could do in ACR that I can't do in PS?

Tony, I've never owned PS but, as you know, ACR has four sliders: Amount,Radius, Detail and Mask. It has a trick where pushing the ALT key before adjusting gives a monochrome impression of the effect of the adjustment. It is said that the Detail slider applies de-convolution sharpening at it's higher settings.

If PS does all of that (Elements 6 does not), then the answer is "no, there is nothing you can do in ACR that you could not do in PS".

[TonyW]

Thanks, Ted. Following the information I have acquired in this a similar threads I have been using those sliders and found that they do seem to have good effects. So far they are the only things that I can't duplicate in PS.

[xpatUSA]

As a further illustration of 'before and after processing', I looked at yesterday's stunning shot of what was my drive-way before the undergrowth started moving in.



Here's a comparison screen shot of the wheel-barrow in FastStone Viewer:



It's a Bayer CFA sensor with a blur filter this time, to avoid Foveon obfuscation. At right, RawDigger's RGB export. At left, sharpened in ACR 5.4 followed by more sharpening in PSE 6.

Again, it's quite noticeable how blurry our raw originals are.

[NorthernFocus]

...it's quite noticeable how blurry our raw originals are.

Presumably by "our" you are not referring to the universal "we". Below is an example of an image with NO sharpening applied in LR as well as none applied during export. This image was taken with a D7100 which has no AA filter.

The eye is a full resolution crop. Surely is can be improved with a bit of sharpening but in its RAW state it's not what I'd call blurry.

[xpatUSA]

Presumably by "our" you are not referring to the universal "we". Below is an example of an image with NO sharpening applied in LR as well as none applied during export. This image was taken with a D7100 which has no AA filter.

The eye is a full resolution crop. Surely is can be improved with a bit of sharpening but in its RAW state it's not what I'd call blurry.

Yes, Dan, splendid images indeed. I should have made it clearer that, by "raw original" I was referring to an image made from raw data not involving a converter such as ACR or it's in-camera equivalent. Is there any chance you could post that .NEF anywhere so I could let RawDigger loose on it?

I'm glad you posted the 100% crop because the other, down-sized, image has more acutance just through the down-sizing.

[Manfred M]

This may be a bit off the track of the original post, but I normally don't do anything significant in ACR and rely on PS for all my editing. Is there anything I could do in ACR that I can't do in PS?

Tony - "best practices" work flow means you should be doing at least your WB, import sharpening (i.e. compensation for the AA filter) and global edits (exposure, white point, black point, shadow detail, luminance corrections) in the RAW converter before importing the file into Photoshop. Once you bring your file into Photoshop, the WB and gamma are "baked-in" and you get too many side effects trying to do these in PS later on. The RAW data gives you a lot more latitude to make these edits.

My workflow used to be like yours, and having gone to a more balanced workflow over the past couple of years has sped things up and given me better end results.

[NorthernFocus]

...I should have made it clearer that, by "raw original" I was referring to an image made from raw data not involving a converter such as ACR or it's in-camera equivalent...

I guess I'm confused. I didn't think ACR/LR read any of the proprietary information in the RAW files other than WB. I thought LR was displaying the unaltered RAW file. Is that not the case?

I'm not sure how I could get the NEF file to you. Might be able to through Zenfolio. I'll take a look when I get a chance.

[xpatUSA]

I guess I'm confused. I didn't think ACR/LR read any of the proprietary information in the RAW files other than WB. I thought LR was displaying the unaltered RAW file. Is that not the case?

It can indeed be confusing. I'm using RawDigger to show the raw image data. When a raw (Bayer) file is opened, the data is kept in R, G1, B, G2 format, not RGB. So a histogram looks like this:



As far as I know, when ACR/LR opens a raw file, it gets de-Bayered and converted into a temporary RGB file in Adobe's inner working space. At that point, the original, really raw data has already been transformed and messed with by Adobe. It is this file that you see on-screen and it is the file that gets all your edits.

[dje]

This may be a bit off the track of the original post, but I normally don't do anything significant in ACR and rely on PS for all my editing. Is there anything I could do in ACR that I can't do in PS?

Tony I try to do as much as I can in ACR. It is very good these days and I often find that I don't need to do anything much in PS. Limited selective adjustments can be done in ACR but in general, I prefer to do these in PS.

I feel that some adjustments in ACR are better and or easier to use than the equivalents in PS. Eg contrast, sharpening and noise reduction. Also it is said by some that better results can be achieved in ACR because you are working on linear 16 bit data (ie without gamma applied).

One reason I like to use the sharpening and noise reduction tab in ACR is that sharpening and noise reduction can be done in conjunction with each other. If you are working on a high ISO shot that needs noise reduction, often a bit more sharpening needs to be applied to account for the softening caused by the noise reduction.

On the other hand, if I want to use the Nik filter plug-ins, I need to take the image into PS. Also I prefer to do perspective adjustment in PS because has a lot more options.

Cropping can be done in either and I usually do this in ACR.

One good thing about ACR is that the adjustments are made parametrically - the pixel values are not permanently changed but rather edit settings are saved in an XML siidecar file and are applied every time you open the raw file in ACR. Using the Snapshot facility, you can save several different versions of your edit very easily.

You can do similar things in PS with layers and Smart Objects but file sizes end up much larger and Smart Objects with several filters can be slow to re-render every time you make a change.

Dave

[TonyW]

Thanks, Dave. It does look as though my strategy is not optimal and I must experiment with different practices. I'm finding it difficult because the effect of the different adjustments are not explained very well at a fundamental level. It is not clear to me what the Detail slider does precisely although it does seem to be effective and there does not seem to be any replacement in photoshop.

Another thing I just read somewhere is that it is better to do colour balancing before demosaicing. Does ACR do that or am I misinterpreting something here? Usually I accept the As Shot option for colour balance and don't make further adjustment. When I do change the balance in ACR or PS it is very slight and for a special reason or for a special purpose.

[dje]

Another thing I just read somewhere is that it is better to do colour balancing before demosaicing. Does ACR do that or am I misinterpreting something here? Usually I accept the As Shot option for colour balance and don't make further adjustment. When I do change the balance in ACR or PS it is very slight and for a special reason or for a special purpose.

Tony I have read something on that somewhere too but I don't know whether WB adjustment is done before or after de-mosaicing in ACR. If it is done before, then presumably de-mosaicing would have to be re-done every time you change the WB setting and in my experience, the WB adjustments are virtually instantaneous (which means that de-mosaicing would have to be fast). It is not something you have any control over. Yes As Shot usually works well for me too but the first thing I do in ACR is fiddle with WB if the colours look a bit off.

Dave

[NorthernFocus]

It can indeed be confusing. I'm using RawDigger to show the raw image data. When a raw (Bayer) file is opened, the data is kept in R, G1, B, G2 format, not RGB. So a histogram looks like this:
As far as I know, when ACR/LR opens a raw file, it gets de-Bayered and converted into a temporary RGB file in Adobe's inner working space. At that point, the original, really raw data has already been transformed and messed with by Adobe. It is this file that you see on-screen and it is the file that gets all your edits.

Well now I'm even further confused. What does the data set prior to creating the RGB image have to do with sharpness/sharpening?

[xpatUSA]

Well now I'm even further confused. What does the data set prior to creating the RGB image have to do with sharpness/sharpening?

Nothing. I really do wish I hadn't started this thread which was never intended to be about how to sharpen. Neither was it intended to be about the difference between various sharpening methods; nor was it about ACR/LR, PS (CC or no), SPP, RT, HF, ID, or any other converter.

As a matter of fact, now I come to think about it, the OP is actually quite trivial and says nothing that we don't already know: that what the sensor captures, what you call "the data set" for some reason, is usually blurred and needs post-processing to look sharp.

That is all that I sought to illustrate. Trivial indeed.

I do apologize for wasting your time.

[TonyW]

Ted, you have not wasted my time. I might be partly responsible for taking the discussion off course but I have learned something on the way.

[NorthernFocus]

Nothing. I really do wish I hadn't started this thread which was never intended to be about how to sharpen....I do apologize for wasting your time.

Actually I thought the thread started rather well but it did drift, didn't it? Not a waste of time. Particularly if the thread was followed by anyone out there who is still clinging to that nonsense that only SOOC images are true photography. In that context this thread has been rather informative in demonstrating how much processing is indeed necessary from the original sensor capture to a usable image. It's really simply a question of where the processing occurs and with what.

[Texas Dave]

I like to sharpen in photoshop as well and often dupe the layer, sharpen and then mask out area that are to sharp with a layer mask. I can vary the opacity of the layer and brush to do so.

One other "trick" I sometimes do is convert to lab and sharpen only the lightness channel, then convert back to my rgb profile. This method can bring in some sharp highlights and fool the eye nicely without allot of unsharp mask halos.