Will new "blur" LR Classic feature make fast f/4 wildlife lenses obsolete?

If you would like to post, you'll need to register. Note that if you have a BCG store account, you'll need a new, separate account here (we keep the two sites separate for security purposes).

I didn't realize that. I'm largely a purist when it comes to any of this sort of stuff but I think banning denoising is outright dumb.
I remember looking at the NANPA Showcase entry rules a few years back and as I recall denoising in general wasn't prohibited but AI based tools including AI based denoising wasn't allowed. I expect some of this is just the world trying to figure out how to navigate the new ML and AI technologies and not knowing where to draw lines.

IOW, I hear non-photographers ask whether an image was 'photoshopped' without any understanding of the difference between basic raw processing and heavy handed image manipulation like adding subjects or replacing backgrounds which wouldn't be allowed under a lot of nature photo contest rules. I suspect the term AI has folks equally suspicious as to how much manipulation has occurred.

It probably doesn't help that the term AI is very hot right now so lots of software companies are trying to frame their products as AI products. Some no doubt are and use AI/ML methodology, trained supervised or unsupervised models and the like but some are just adaptive algorithms using more traditional heuristic decision making that have jumped on the AI bandwagon for marketing reasons which muddies the water even further.

But yeah, I'm with you and don't think modern noise reduction is really a problem.
 
remember that you can get shallow dof with slower lenses if you arrange the camera/subject/background ratio. fast glass also gives you the ability to shoot in darker conditions. fast glass doesn’t have a monopoly on shallow dof
Actually fast glass is the only way to get shallow DOF. But I think what you're saying is that shallow DOF isn't the only way to get subject isolation with blurred FG/BG. And actually if we really pay attention when we look at photos with nice blurred FG/BG even those shot with fast glass are usually composed well to create that effect.
 
Remember that the new future in Lightroom is betta, or "early accesss" as Adobe likes to call it. I'm sure it will get better, and eventually will probably be quite good. That said, like a lot of editing tools, I'd rather just get it right in the first place. Like correcting motion blur with Sharpen AI, it never quite looks as good as getting it right in the first place. :)
 
The technology isn't there yet - but I have no doubt it will be at some point. The biggest issue I see with it is that it's not great at transitions from sharp to blurry like a real lens - it's applied in too general a way without enough subtle granularity in the transitions.

Still, let's play a game using some bokeh examples from one of my recent videos. In this case, we're going to compare the 600 6.3 to the 600 F/4. Also keep in mind this is an INCREDIBLY easy and unrealistic scene - real wildlife shots would be much tougher.

First, just what the two lenses look like side by side, the backgrounds unmolested by technology :)

View attachment 73732

Now, how it looks with 100% blur applied to the F/6.3 lens.


View attachment 73733

At first glance, the 600PF looks really good here, but look closer. Notice how the blur "spreads out" in the 600 F/4 shot, as opposed to the 600 6.3 where it simply takes the background and applies uniform blur to it. You can especially see this in the greens in the lower left and also the yellows right next to the clamp (note the size the the dark areas). Also, note the transitions between the blurry areas - they are still contrasty and hard compared to the 600 F/4. However, the tool does soften the textures in the background which is a plus. Will a casual facbook user notice things like this? Probably not. However, they may notice that the effect looks somehow "fake" without being able to put a finger on it.

Here's a quick side by side with just the 600 6.3 so you can see the difference the filter made.

View attachment 73734

Also, keep in mind that you can also apply this to the F/4 shots too. In this case, both at 100% blur.


View attachment 73735

Personally, I think the best use for this tool isn't using it at 100% :)

I think the best way to use it is for just knocking off the hard edges in the background and not trying to completely simulate a faster lens - I think when you push it that far, it tends to look a little unrealistic (although, this can also depend on the shot - texture of the background, that sort of thing).
Good comparison Steve.
I think your final advice about the best way to use it really applies to all the brushes and sliders. Too much sharpening, too much saturation, too much clarity, too much gaussian blur, too much shadow boost, too much of anything is, well, too much.

Technology marches on and, generally, improves over time. I remember, not that long ago, when everyone in the photo magazines and reviews said digital will never replace film, not enough resolution, too grainy, unrealistic colors. Yet, here we are today.

Interesting discussion. Honestly, I think it will be a few years before technology makes the fast glass more or less obsolete but it will be a few years. As with most technology, early adapters and early versions have lots of glitches but rapidly, the technology becomes better and perfected.

It's a good time to be a photographer.

Jeff
 
I experimented with the Adobe Lens blur with multiple shots taking less than 5min on each photo. In my view it can improve a photograph but it is not a replacement for a F/4 lens or for getting down to the animal's level to avoid a close background. I have experimented with more or less of the effect according to taste. Some shots that I have used the effect on.
395659515_10231875237328409_4480593523698415917_n.jpg
You can only see EXIF info for this image if you are logged in.
394309917_10231843381412031_2006890649044234969_n (1).jpg
You can only see EXIF info for this image if you are logged in.
 
Since we can now considerable blur backgrounds with the new "blur" feature in LR Classic and Adobe Camera Raw, will this feature become so refined that expensive f/4 telephoto lenses will become obsolete? I doubt it, but what are your thoughts?

I think the whole subject against a creamy background thing has gotten over-used at this point and makes for mostly dull and boring shots.
BUT ... 1/2000s @f4 is enough to freeze plenty of motion. That becomes 1/500s @f8 which is not enough to freeze most motion. For me, that is the main reason for having a faster telephoto :).

Then again, I often found that even f8 was not enough depth of field to properly get a bird in focus.

And I often found that thinking of background as not something that needs to be dealt with and more like a part of the story of your photo helps a lot.
You might actually want to make it stand out if it provides context to your shot or maybe take advantage of some patterns or colors in it to make your subject stand out.

One a more technical note, I find it much more natural to de-emphasize backgrounds using AI subject selection in LR, inverting the mask, making quick refinements to it and then using the contrast/sharpness/texture/clarity sliders. I find the output subtler with much less jarring transitions between in focus and out of focus areas.
 
I think the whole subject against a creamy background thing has gotten over-used at this point and makes for mostly dull and boring shots.
BUT ... 1/2000s @f4 is enough to freeze plenty of motion. That becomes 1/500s @f8 which is not enough to freeze most motion. For me, that is the main reason for having a faster telephoto :).

Then again, I often found that even f8 was not enough depth of field to properly get a bird in focus.

And I often found that thinking of background as not something that needs to be dealt with and more like a part of the story of your photo helps a lot.
You might actually want to make it stand out if it provides context to your shot or maybe take advantage of some patterns or colors in it to make your subject stand out.

One a more technical note, I find it much more natural to de-emphasize backgrounds using AI subject selection in LR, inverting the mask, making quick refinements to it and then using the contrast/sharpness/texture/clarity sliders. I find the output subtler with much less jarring transitions between in focus and out of focus areas.
Oy vey - Sorry for your troubles but your post and statements are a tad much for this nuclear blond - and I am challenging you to prove your exact statement:

Creamy backgrounds makes for mostly dull and boring shots.

I think one needs to be careful to put everything in a box. However you or anyone shoots, has it’s place for that person. Another person’s images are like having visitors - I am always happy to see them. Some when they come and some when they go
 
I experimented with the Adobe Lens blur with multiple shots taking less than 5min on each photo. In my view it can improve a photograph but it is not a replacement for a F/4 lens or for getting down to the animal's level to avoid a close background. I have experimented with more or less of the effect according to taste. Some shots that I have used the effect on.
One of the issues with doing blurring in post - Halos/ghosting, and / or hard jaggety edges - the solution is time consuming.
 
Oy vey - Sorry for your troubles but your post and statements are a tad much for this nuclear blond - and I am challenging you to prove your exact statement:

Creamy backgrounds makes for mostly dull and boring shots.

I think one needs to be careful to put everything in a box. However you or anyone shoots, has it’s place for that person. Another person’s images are like having visitors - I am always happy to see them. Some when they come and some when they go
His statement began with "I think". Yet you cherry pick the last part and present it as if he was stating it as fact, you even went so far as to capitalize the C in creamy as though that was the beginning of his statement. How would someone go about proving their preference to you and why do you feel they need to after you twist their words around?
 
I would agree there is plenty of room for each person's point of view, and even see some merit to it, depending on the shot, as the environment might tell more of a story than a blur in some cases.
 
Since we can now considerable blur backgrounds with the new "blur" feature in LR Classic and Adobe Camera Raw, will this feature become so refined that expensive f/4 telephoto lenses will become obsolete? I doubt it, but what are your thoughts?
There’s a time and place for almost every software tool, as long as it’s used in moderation. I view the new blur feature no differently than, say, using DeNoise or sharpening an image, again as long as applied judiciously. Ideally of course, we’d all get it “right” in the camera and wouldn’t need software to clean things up but the reality is that just doesn’t happen often enough. Of course, the best chances of capturing perfect images are with quality glass. I know I’m stating the obvious, so apologies for being so preachy.
 
Last edited:
The technology isn't there yet - but I have no doubt it will be at some point. The biggest issue I see with it is that it's not great at transitions from sharp to blurry like a real lens - it's applied in too general a way without enough subtle granularity in the transitions.

Still, let's play a game using some bokeh examples from one of my recent videos. In this case, we're going to compare the 600 6.3 to the 600 F/4. Also keep in mind this is an INCREDIBLY easy and unrealistic scene - real wildlife shots would be much tougher.

First, just what the two lenses look like side by side, the backgrounds unmolested by technology :)

View attachment 73732

Now, how it looks with 100% blur applied to the F/6.3 lens.


View attachment 73733

At first glance, the 600PF looks really good here, but look closer. Notice how the blur "spreads out" in the 600 F/4 shot, as opposed to the 600 6.3 where it simply takes the background and applies uniform blur to it. You can especially see this in the greens in the lower left and also the yellows right next to the clamp (note the size the the dark areas). Also, note the transitions between the blurry areas - they are still contrasty and hard compared to the 600 F/4. However, the tool does soften the textures in the background which is a plus. Will a casual facbook user notice things like this? Probably not. However, they may notice that the effect looks somehow "fake" without being able to put a finger on it.

Here's a quick side by side with just the 600 6.3 so you can see the difference the filter made.

View attachment 73734

Also, keep in mind that you can also apply this to the F/4 shots too. In this case, both at 100% blur.


View attachment 73735

Personally, I think the best use for this tool isn't using it at 100% :)

I think the best way to use it is for just knocking off the hard edges in the background and not trying to completely simulate a faster lens - I think when you push it that far, it tends to look a little unrealistic (although, this can also depend on the shot - texture of the background, that sort of thing).
Steve,
As usual, a great analysis. Thanks for the explanation. Kinda what at expected at this stage.
 
It is an interesting feature that no doubt began its life for phone cameras that don't do very good with background blur, in fact, the iPhone and Samsung (et al) software tools to "enhance" images have bled into the more professional post processing environment. That being said, I tend to agree with others here, it is interesting and has potential but not quite "there yet"; however, with what we've seen in technology it will likely "get there" pretty quick. Right now the tool doesn't do well with foreground blur and the line of delineation between blur and no blur is sometimes too obvious. I am excited about where these AI tools are going but also see a danger in them as well - to be mishandled and/or misappropriated. We have been altering images to meet our liking since the advent of film itself, even the film companies altered formulas which gave us different results, and even our DSLR and mirrorless mfr's are giving us their own interpretations of the rendered digital capture (not all RAW engines are created equal) - I definitely saw this going from Nikon D series to Z series, even using the same lenses the images had more "life" with the Z RAW files vs. the D RAW files; could I modify the D images to get close, sure, but Nikon did it right out of the box so to speak. I've been using Presets for years with LR and PS because of a certain look I am going for and I see AI as another tool on my PP belt with which to be creative. How those tools effect the final image and are perceived by the viewer is another story.
 
Oy vey - Sorry for your troubles but your post and statements are a tad much for this nuclear blond - and I am challenging you to prove your exact statement:

Creamy backgrounds makes for mostly dull and boring shots.

I think one needs to be careful to put everything in a box. However you or anyone shoots, has it’s place for that person. Another person’s images are like having visitors - I am always happy to see them. Some when they come and some when they go
I don't want to put words in Stefan's mouth and do not wish to claim any special insight into what he was thinking. I can say the way I read it and what I tend to think about the topic.
1) There is a risk for photos to start having the same look. A few years back, everyone was posting and shooting the "animal portrait" "Hero" shot. Those are still great but it got kind of old. I have always gravitated toward more environmental shots. Hero shots still have their place and do showcase magnificent creatures but variety is the spice of ife.

2) completely blown out backgrounds can, at times, detract from an image, especially when it strips the sense of time and place from the photo. Sometimes it is difficult to for the viewer to discern if the image was takin in the backcountry wilderness, a city park, a zoo or an animal farm. Sometimes context is important.

Kind of like the old joke, "how can you tell the difference between one Disco song and another?" "they have different names."

Again, I don't presume to have any special insight into what the other fellow was thinking, just musing my personal reaction to his post.

The cool thing about photography is it is a form of art, thus there really are no correct or incorrect opinions.

Jeff
 
I have some time using masking and gaussian blur and it seems that the Adobe "lens blur" is better. What the current iteration of the software lacks is blending between in focus and blur.

Tom
 
I have some time using masking and gaussian blur and it seems that the Adobe "lens blur" is better. What the current iteration of the software lacks is blending between in focus and blur.

Tom
That's what seems tough to me, since the automatic subject or background mask was introduced in LR a while back it doesn't seem too difficult to use some form of blur on just the background. The tricky part is actual lens blur and DoF is a function of distance and not all background blurs equally. To emulate actual DoF of a wider aperture lens you'd want the blur to increase as you get further from the subject. Not saying this is impossible, but it sounds a lot trickier than just selecting the background and blurring it.
 
That's what seems tough to me, since the automatic subject or background mask was introduced in LR a while back it doesn't seem too difficult to use some form of blur on just the background. The tricky part is actual lens blur and DoF is a function of distance and not all background blurs equally. To emulate actual DoF of a wider aperture lens you'd want the blur to increase as you get further from the subject. Not saying this is impossible, but it sounds a lot trickier than just selecting the background and blurring it.
In PS you accomplish that by using the brush at different opacity on the mask.
 
In PS you accomplish that by using the brush at different opacity on the mask.
Well not really. You can clone a layer, apply a given radius blur and then paint on a mask to vary the opacity of that blur for different parts of the image. But you still have a fixed radius blur being applied with different opacities throughout the image which isn't what happens optically with an actual wider aperture lens.

Optically the far background has a much larger radius blur than the close background nearer the subject it's not the same as one radius of blur applied at varying opacities. Sure you could conceivably create several layers each with a different blur radius and then use masks to selectively apply them but still the transitions wouldn't be continuous and smooth between them the way it is with actual optics.

Stated differently the size of the Circles of Confusion increase as you move deeper into the frame and further behind (also further in front) of the plane of focus so out of focus objects bloom into larger and larger areas the deeper you get. A single blur radius applied with varying opacity to different parts of the image doesn't behave that way.

BTW, this is what I took from @Steve's examples further up in this thread. The far background doesn't seem to transition as smoothly and have as large a blur radius in the software blur solution as it does with an actual wider aperture lens. Ideally there would be some sort of distance or depth into the image transfer function that would emulate increasing blur radius as you get further from the plane of best focus. To do that accurately somehow the camera would have to know not just the distance to the subject but the distances to the background and varying distance across the background for things like landscape shots that show depth from a high vantage point which seems quite difficult.

Lot's of things can be accomplished with enough clever software so maybe this could all be done in time but I don't think the technology is there just yet especially in a fully automated form though someday it might be.
 
Last edited:
Back
Top