Why buy L lens when you have PC

Digital manipulation is no real substitute for the best possible image quality to start with. Using software to overcome a lens' shortcomings is perfectly acceptable only to that group of people that can't or choose not to afford the best possible glass. No photographer worth his/her salt is going to want to suffer with MORE time at the computer to "perfect" their image when it could have been cleaner and sharper due to better glass in the first place.

But, if it floats your boat, more power to ya.

Mark
--
He who dies with the most red rings wins! (It won't be me!)
See my profile for the list of toys.

I think the "improved" photo looks as if the gal has had facial reconstructive surgery. Look close. The original is better and has character to it...IMO.

-- Greg

jeffx said:

Jeff

Julio Trujillo said:

Forrest said:

... Details you captured in your photo are
made to look better; details you didn't capture, still aren't
there.

Click to expand...

Of course not. But isn't this thread about "L" vs. "non-L" lenses?
In these cases detail is captured by both but the L lens does a
much better job of it. The edits required to compensate are not in
the realm of the miraclous - and a miracle is what is needed to
re-focus the extremely blurred background in your photo example.
This example below is a before and after 100% crop of a shot I took
while "abusing" my 135 f2.0 L - Just to see what I could do, I shot
the first half of a soccer game last Friday with both the 1.4X and
2X II teleconverters stacked (kids, don't try this at home!) As
you can guess the stacking of TC's produced very soft results.
Adding to the problem, the game was in the early evening so it was
shot at ISO 800.
Without further delay I present exhibit 1:



Retouching consisted of Neat Image, Focus magic and some minor tone
and color adjustments.
Disclaimer: I don't work for Focus Magic nor do I own any stock in
the company... hmmm... not a bad idea though - if they ever get big
enough to go public

Click to expand...

Click to expand...

Phil,

I'm truly not sure where the failure to comunicate lies but you're certanly not getting my point. One last try.

The reason that analog film, digitzed and processed in Photoshop is different from a purely digital file is that everything that happens to the analog film after it is digitized is a digital manipulation that, like any digital manipulation consists of pushing pixels around. But the film started out as an analog piece of plastic that has been conventionally processed using long-established analog (chemical) processes. Nowadays, more and more, film is digitized and then processed using Photoshop or similar programs. This has led to the belief (prejudice?) that processing in Photoshop is an inferiror process where you are limited by what was captured on film in the first place and the best you can do is approximate the old analog processes.

Fast forward to digital... digital cameras don't produce negatives - what they produce instead at it's most basic level is a mathematical interpretation of the meaning of the voltages on the photosensitive cells (CMOS on the 10D) it uses mathematical transformations based on weighted averages to guess at 2 out of the 3 colors for every pixel. So right from the very moment that a digital image is first created in the camera the mathematical "pixel pushing" begins. All that the raw converters and Photoshop do is continue the process started in the camera - it is not something 'new" being introduced just more of the same.

To say that there is some inherent "digital negative" similar to an analog negative beyond which "new data" can not be created is to ignore the fact that data is created using pixel-pushing approximations form the very moment the digital image is captured AND can continue beyond that.

Process a raw file with slightly blown highlights in Adobe Camera Raw version 2.2 and compare those highlights to the 10D JPG embedded in the raw file and then come back and tell me again that "new data" can not be created - that is exactly what camera raw does in the highlight end - it accesses information that is not normally accessible using the Canon converter. Then use the shadows/highlight tool in Photoshop CS to pull previously invisible detail from the dhadows and highlights... etc, etc...

You can argue that the data was in the raw file all along and all that ACR is doing is accessing what is already there and you would be right - up to a point. The point being that prior to some of the newer versions of the raw converters we are able to get for the first time, at data that was always there but we couldn't see before. Then, of course there is also the fact tha C1, ACR and FVU all interpret the RAW image in different ways and produce slightly different results. Which one of those 3 brings out what is "really there" and is therefore "correct" and which one "invents" new data? You wouldn't need to ask this question with a film negative would you?

DxO Optics Pro is hardly different in what it does to the recent advances in raw processing - except what it does is re-interpret the data into more accurate representation than you would get otherwise (or so they say and I believe them.) So... was the accurate data there all along or is it invented?

I don't know about you but for me the data not being there or me not being able to get at it ammounts to the same thing in the end. Likewise being able to extract it or re-interpret it in a new way that couldn't be done before is pretty well the same as creating it.

My only lens is the 28-135IS, so I have tried renting both L and non-L lens to shoot soccer games. Having tried both the 70-200L f2.8 and the 75-300 non-L, non-IS, I can tell you that, at least from a sports angle, there are pictures you can take with an L that you just can't take with a lesser lens.

When shooting with the L, I looked at the results and found very few shots to throw away. The 75-300 on the other hand, had a very difficult time keeping up with the action - it just wasn't up to the task of continuously adjusting its focal point as the players ran up and down the pitch. I was left with many missed shots as a result.

These are post-processed, web-sized samples of my favorite shot from each lens. To the point of the original post, a good shot from the non-L can be processed to look pretty much as good as from the L. Maybe not at full resolution, but once images are sized for the web or printed as a 4 x 6, there isn't much difference. (These were taken months apart, so the lighting differed)

From the L



From the non-L

Carl Hoang said:

Why buy L lens when you have PC and software to get you the best
sharpness of sharp.

I'm not sure if this subject has been discussed before. Since using
PC and photo software as part of 300D workflow, one could get by
with average lens and use the power of PC and software to get the
best picture.

I would love to see the comparison between L and non-L lens with
using software for correction. Please post your comparison if you
have any. TIA.

Click to expand...

--
http://imageevent.com/merritt

Frank Weston said:

Please don't be offended, but I find the "improved" image garish
and plastic, and to my eyes, definitely not an improvement on the
original. If anything, your presentation is an arguement against
the idea that all lenses can be made equal with software.

Click to expand...

LOL... I'm not offended Frank. I have to bow to the overwhelming consensus that barbie sucks

I thought it showed improved focus (especially the lettering on the shirt) perhaps not

which lens am I using.





More at:
http://www.softechie.com/Photos/2004MNStateCup/20040515U14BWingsVUSC/
http://www.softechie.com/Photos/2004MNStateCup/20040516U14BWingsBlackhawks/
http://www.softechie.com/Photos/2004MNStateCup/20040517U15BWingsMaplebrook/

bee1000 said:

When shooting with the L, I looked at the results and found very
few shots to throw away. The 75-300 on the other hand, had a very
difficult time keeping up with the action - it just wasn't up to
the task of continuously adjusting its focal point as the players
ran up and down the pitch. I was left with many missed shots as a
result.

These are post-processed, web-sized samples of my favorite shot
from each lens. To the point of the original post, a good shot from
the non-L can be processed to look pretty much as good as from the
L. Maybe not at full resolution, but once images are sized for the
web or printed as a 4 x 6, there isn't much difference. (These
were taken months apart, so the lighting differed)

From the L



From the non-L

Carl Hoang said:

Why buy L lens when you have PC and software to get you the best
sharpness of sharp.

I'm not sure if this subject has been discussed before. Since using
PC and photo software as part of 300D workflow, one could get by
with average lens and use the power of PC and software to get the
best picture.

I would love to see the comparison between L and non-L lens with
using software for correction. Please post your comparison if you
have any. TIA.

Click to expand...

--
http://imageevent.com/merritt

Click to expand...

I found only one canon camcorder:

Canon XL1S MiniDV Digital Camcorder with SuperRange Optical Image Stabilization & Interchangable Lenses

And withit you can use Canon EF (EOS) still-camera lenses--some of which also feature optical image stabilization. I suppose not all L-serie lenses.

PaulyOly said:

Ari Riutto said:

hockeynut said:

The high end Canon miniDV camcorders will take L lenses

Click to expand...

--

• That's my 0,02 € -

• Ari -

http://www.sci.fi/~ajr/ ; (16-35 L pictures)
http://koti.phnet.fi/arriutto/index.html ; (24-70 L pictures)
http://personal.fimnet.fi/koti/ari.riutto/ ; (100-400 L IS shots)

Click to expand...

--
http://www.pbase.com/paulyoly/root

Click to expand...

--

• That's my 0,02 € -

• Ari -

http://www.sci.fi/~ajr/ ; (16-35 L pictures)
http://koti.phnet.fi/arriutto/index.html ; (24-70 L pictures)
http://personal.fimnet.fi/koti/ari.riutto/ ; (100-400 L IS shots)

I'm still looking for the plane of focus in the original picture. I think it was either shot with a soft-focus lens, or blurred with software. but you're right, it is more real.

GPB said:

I think the "improved" photo looks as if the gal has had facial
reconstructive surgery. Look close. The original is better and
has character to it...IMO.

-- Greg

Click to expand...

--
Tom

No idea. On the web, without knowing if they're retouched, etc, I don't think I could tell - that was part of the point of my post, actually.

It looks like a cloudy day. If it were sunny, I would guess not an L lens, because they look a little hazy/uncontrasty, but that could be a result of the weather and not the lens.

As it is, it looks like it could be an L. The creases and patches on the jersey look like they have more detail in them than a non-L would catch. But I am no expert! I only have 2 days' experience with any lenses other than my 28-135!

Carl Hoang said:

which lens am I using.

http://www.softechie.com/Photos/2004MNStateCup/20040515U14BWingsVUSC/Lucas06.jpg
http://www.softechie.com/Photos/2004MNStateCup/20040515U14BWingsVUSC/Ben02.jpg

Click to expand...

snip

Carl Hoang said:

bee1000 wrote:

Click to expand...

snip

Carl Hoang said:

Carl Hoang said:

To the point of the original post, a good shot from
the non-L can be processed to look pretty much as good as from the
L. Maybe not at full resolution, but once images are sized for the
web or printed as a 4 x 6, there isn't much difference. (These
were taken months apart, so the lighting differed)

Click to expand...

Click to expand...

--
http://imageevent.com/merritt

Carl

so the reason it is ok for you is because you are happy with completely useless mediocre photos.

Well you could have said that in the beginning.

--
Michael Salzlechner
http://www.PalmsWestPhoto.com

The images have been scaled down for the web, so there is no way to tell, because both types of lenses have resolution higher than this scaled down image.

Post 100% crops of a picture taken with a soft, consumer grade lens, processed with whichever software is producing good sharp results, and a picture taken with an L-series lens, then we'll have something to debate...

Carl Hoang said:

which lens am I using.

Click to expand...

Julio Trujillo said:

Phil,
I'm truly not sure where the failure to comunicate lies but you're
certanly not getting my point. One last try.

Click to expand...

Dude! I sure hope this is just miscommunication...!

Julio said:

The reason that analog film, digitzed and processed in Photoshop is
different from a purely digital file is that everything that
happens to the analog film after it is digitized is a digital

Click to expand...

Julio said:

3 colors for every pixel. So right from the very moment that a
digital image is first created in the camera the mathematical
"pixel pushing" begins. All that the raw converters and Photoshop

Click to expand...

So why is that relevant? No one is claiming that the camera is adding real detail when it interpolates 2 colors out of 3 for every pixel. I'm certainly not. The "real data" is what was captured by the CCD, not what comes out of the camera. No "real data" got added during the Bayer interpolation.

(I mean hey, you could use a Foveon sensor then, which has no interpolation pattern, then your argument becomes moot).

Julio said:

do is continue the process started in the camera - it is not
something 'new" being introduced just more of the same.

Click to expand...

Well, I agree with you there... it is more of the same. More of the same which does not add any detail or extra information to the image.

Julio said:

To say that there is some inherent "digital negative" similar to an
analog negative beyond which "new data" can not be created is to
ignore the fact that data is created using pixel-pushing
approximations form the very moment the digital image is captured
AND can continue beyond that.

Click to expand...

Well, you could say the digital negative is the data that was captured by the CCD, with only one colour per pixel. Unfortunately that would be very inconvenient to edit in photoshop Kind of like it would be hard to make prints from film negatives if the dye layers weren't conveniently sandwhiched together in one piece of material

Julio said:

Process a raw file with slightly blown highlights in Adobe Camera
Raw version 2.2 and compare those highlights to the 10D JPG
embedded in the raw file and then come back and tell me again that
"new data" can not be created - that is exactly what camera raw

Click to expand...

New data can not be created.

Julio said:

does in the highlight end - it accesses information that is not
normally accessible using the Canon converter. Then use the

Click to expand...

That's not creating new data. The data is already there. That the Canon converter has to turn it into an 8bpp jpeg (since your monitor only displays 8bits).

Julio said:

shadows/highlight tool in Photoshop CS to pull previously invisible
detail from the dhadows and highlights... etc, etc...
You can argue that the data was in the raw file all along and all
that ACR is doing is accessing what is already there and you would
be right - up to a point. The point being that prior to some of the
newer versions of the raw converters we are able to get for the
first time, at data that was always there but we couldn't see
before. Then, of course there is also the fact tha C1, ACR and FVU

Click to expand...

Here's an analogy:

You have an audio recording. It has recorded frequencies up to 20KHz, beyond the range of the human ear. So you can't hear it. Then you run a pitch-shift effect on it, bringing those 20KHz frequencies down to 10KHz. All of a sudden you can hear those frequencies (i.e. you can see the detail in the formerly blown out highlights).
Ok, so that's obvious.

Now you're saying, shift the frequencies down another octave... so those 20KHz frequencies are now 5KHz, and 40KHz frequencies are now 10KHz, and you can hear them.... oh but wait!.. the original recording wasn't able to recording anything above 20KHz, so there's no data there. But you are saying there is. That's what you seem to be claiming these post-processing steps are doing - yet you are equating it to RAW converters being able to extract "invisible data". It's different. The invisible data is only because of the limitations of the file format/computer monitor/human ear.

Julio said:

I don't know about you but for me the data not being there or me
not being able to get at it ammounts to the same thing in the end.
Likewise being able to extract it or re-interpret it in a new way
that couldn't be done before is pretty well the same as creating it.

Click to expand...

No. Re-interpreting it is one thing. I can load the 16bit RAW file on photoshop and extract detail from the shadows. It was always there, it wasn't created out of thin air... you just couldn't see it because your monitor can only display 8bpp. Just like the human ear couldn't hear the 20KHz frequencies that were recorded.

This is fun, let's keep it up!

I own and use Focus Magic, and it has its place, but its place isn'tr generic sharpening, noise reduction, or for improving the results delivered by a low quality lens. What Focus Magic is very good for, is correcting motion blur, particularly the small amount of blur associated with long focal lengths shot at a slightly too slow shutter speed.

I've experimented with just about all the photo software out there, and the bottom line is that there's just no substitute for starting with a good, clean, clear, properly exposed image right out of the camera. Software can sometimes reclaim a mediocre image and make it good, but software can never make a mediocre image great.

Julio Trujillo said:

Frank Weston said:

Please don't be offended, but I find the "improved" image garish
and plastic, and to my eyes, definitely not an improvement on the
original. If anything, your presentation is an arguement against
the idea that all lenses can be made equal with software.

Click to expand...

LOL... I'm not offended Frank. I have to bow to the overwhelming
consensus that barbie sucks
I thought it showed improved focus (especially the lettering on the
shirt) perhaps not

Click to expand...

--
Frank Weston - http://www.weston.smugmug.com

If i had any software of hardware that produced a image like the processed one you posted, i would be asking for my money back. Your shot is the best example of garbage in = garbage out. Yes, the processed shot may be considered as "sharper" but the subject no longer looks human, all the features have been blurred together with the end result being a very plastic look, the girl looks like a manequin now - and the colours look like someone has drawn make up on with crayons after the fact. If you were posting that example to show why you need a good shot to start with i could understand, but to post it thinking that the processed results are good is not understandable.

WRT the original topic, regardless of any advances in software, you are always going to be better off with good glass. If software is capable of making an image shot with a poor quality lens look good, just think how much better the same shot taken with a good lens will look after some processing. Software cannot automagically put in details with pixels if the detail isnt there to start with, it can only manipulate what is there. Software can only make some assumptions about what might be considered to look better to the eye, and when analysing the pictures, can only make some educated guesses abou the manipulation that should be made in the hope that the end result will be considered as a better looking image to the viewer.

Cheers,

Ben
Its not the speed that kills, its the sudden stop at the end.

Carl,

You miss the point. A cheap lens can take good pictures. What a professional quality lens can do is take good pictures more of the time. A pro quality lens will deliver better results in low light, better results with a more distant or faster moving subject, better results when it's raining, hot, cold, or dusty....just plain better results more of the time.

On the other hand, no matter what lens or camera or software you use, the photographer is responsible for 99% of the quality of any shot. Good gear just makes it easier.

--
Frank Weston - http://www.weston.smugmug.com

Just go to the bloody DxO Pro website and argue with them.

Frank Weston said:

Software can sometimes reclaim a mediocre image and make
it good, but software can never make a mediocre image great.

Click to expand...

Agreed.

But that's not what this thread is about. The original poster wanted to know if Software could make a non-L lens L-like. Everyone quickly jumped on him, insulted him, called his point useless, etc. I'm not saying that you jumped in and joined the lynch mob - you didn't and I have never known you to do anything remotely like that.

I reacted to the feeding frenzy and the typical knee-jerk ignorance that is exhibited in this forum much too often by pointing out that DxO Optics Pro claims to do just that and some people who have tested it seem to be very impressed. (Obviously you can also improve the L lens image so that argument only holds if you compare a non-L lens + Optics pro vs. an L lens without it - everything else being equal.)

My original post in this thread and point simply was: this question is not as silly as you guys are making it out to be. My mistake, if I made one beyond jumping in to play devil's advocate in the first place, was to mention Focus Magic as another example of software that goes beyond what is conventional to enhance a photo. And then when some smart-ass said "let's see Focus Magic fix this" and posted an extremely blurred background... I unfortunately couldn't resist biting... mea culpa!

As you have, I perhaps should have emphasized Focus Magic's ability to correct motion blur. Mind you I don't think FM's ability to do this is any more worthy of respect than its OOF filter. But motion blur fixing is certainly unique, relatively new and unusual. So let's talk about that.

I can almost guarantee you that somewhere in these archives dating back from a time before Focus Magic's motion blur algorithm was introduced, someone, somewhere, suggested that it might be possible to do just that. I also guarantee you that unless that speculative post came from an acknowledged guru, the troll mob would have jumped all over him and called him an idiot because "the data-without-motion-blur just ain't there and if it ain't there you can't create it." Well guess what... someone applied math to the problem and now Focus Magic can find a sharp "silk purse" in a formerly blurred "sow's ear."

By the way, imho this whole "create" thing is just a red-hering. Why don't we call it instead "making-visible-what-formerly-could-not-be-made-visible." I know, it's one helluva mouthfull but it may prevent semantics-based arguments.

Phil Fortier said:

Dude! I sure hope this is just miscommunication...!

Click to expand...

OK. I finally understand the probllem... d00d... you're deliberately missunderstanding to try to continue the argument. Hmmm, isn't there a word for that that begins with "Tr" and ends with "oll"?
Believe whatever and sell someone else on it...

Just one rhetorical question (that's the kind that requires no answer, by the way) What would you like to call the process of getting information out of a digital image that no one else could ever get out of it before... obviously you object to "creation", so go ahead pick any word you like and use it to describe what focus magic does when it corrects (as in makes it go bye-bye) small to moderate amounts of motion blurring... any word at all...

The real truth is a Sharp photo is a sharp photo. You can't take a nearly sharp photo and make it a sharp photo using any software. The image posted at the very beginning of this thread shows some amazing results but this is only based upon where it was when you started with it. That photo compared to a truly sharp photo would seriously pale in comparison end of story.

You've missed my point. Software can enhance photos from average lens to look "almost like" L lens. If photo is off 10% and software can help it to get to 95-99% of L lens then why spend the big bucks when you can do with $100 software.

For photo quality only: Sigma Lens + Software = Canon L Lens ??? Just a question.

Frank Weston said:

Carl,

You miss the point. A cheap lens can take good pictures. What a
professional quality lens can do is take good pictures more of the
time. A pro quality lens will deliver better results in low light,
better results with a more distant or faster moving subject, better
results when it's raining, hot, cold, or dusty....just plain better
results more of the time.

On the other hand, no matter what lens or camera or software you
use, the photographer is responsible for 99% of the quality of any
shot. Good gear just makes it easier.

--
Frank Weston - http://www.weston.smugmug.com

Click to expand...