Don't be afraid of digital zoom!

[Adam Kozlowski]

The quality difference between using a built-in-camera digital zoom and an external piece of software for it, pays so much more than the amount of time saved by using the camera digital zoom on the spot.

Trust me.

You're just using the wrong software for cropping/enlarging.

Trust me.

I know the maths behind it.

--
Adam Kozlowski Photography - http://www.spooky.pl
Creative and fun outlet - http://spooky79.deviantart.com

 

[Adam Kozlowski]

A different program may even do a better job.

Yes!!!!!! That's exactly my point. THERE IS great zooming software OUT THERE. But people need to google for it

I do recommend Shortcut PhotoZoom or the earlier version called S-Spline PRO. This is easy, with loads of great options. The S-Spline regular version was totally automated and also kicked it!

And if your image is noisy, first put it thru NeatImage.

I don't recommend Extensis pxlSmartScale, even though it does a better job than Photoshop or IrfanView (which, funnily, is better than Photoshop for resizing, i.e. more options) but it too clearly separates high-frequency areas in the image and processes them differently, and it is too visible in many images.

--
Adam Kozlowski Photography - http://www.spooky.pl
Creative and fun outlet - http://spooky79.deviantart.com

 

[Adam Kozlowski]

Yes...

The cameras upsampling methods are usually bilinear, which is the fastest and simplest algorithm to code. Bicubic used in Photoshop is way better as far as reproducing high frequencies (i.e. edges) in the image, however it is way slower and it leaves the unwanted artefacts, just as bilinear - i.e. the jagged lines and blurriness. IrfanView, being an extremely simple image viewer, does offer many more options as far as resampling of images is concerned. There you can choose other methods which can provide even better results than Bicubic Smoother or Sharper or Standard in Photoshop.

However please be advised that the only images that can be correctly and amazingly well enlarged by some standalone programs like PhotoZoom from Shortcut or pxlSmartScale by Extensis or Pictura or others, are the images that are inherently not noisy, devoid of those nasty JPEG artifacts plus they are sharp enough! This is the problem with most cameras - the image that goes to digital zooming is too soft. So no high frequencies (i.e. edges) can be detected and preserved for higher sampling rate photos.

Therefore it is very wise to shoot without digital zoom and postprocess the images at home/work using standalone software, but first by enhancing the quality of images, by, e.g. cleaning the noise with NeatImage, sharpening with PhotoShop (every camera has some nice USM tricks that work great on the photos from it) and only then blow it up in e.g. PhotoZoom.

--
Adam Kozlowski Photography - http://www.spooky.pl
Creative and fun outlet - http://spooky79.deviantart.com

 

[erichlund]

Ron:

Only Nick knows. Not me, not you. He says yes.

He hasn't even said if he was using the spot meter.

If you really had to, you could use the digital zoom to meter, but
still use a regular capture to get the image.

I've seen no spec on the angle of the S2 spot at the 12X zoom, nor
do I, nor do you know the angle of the moon subtended from where
location where Nick was taking the picture. I just KNOW that it is
a technique that works when the situation exists and that's all I
was pointing out. No need to get augmentative about it.

I agree with you that there can exist metering modes and situations
where it's to your advantage to reduce the area the camera is
monitoring in some way.

However, you won't get something for nothing. With a 4X of digital
zoom, you're taking a 640x480 crop from the center of the image and
interpolating it up to 5MP.

The only advantage to doing it in camera is that you might squeeze
more past the quantization threshold of the jpeg algorithm, so any
perceived benefit is an artifact of shooting with a camera that
doesn't support RAW.

The disadvantage is that for magnifications that are not integer
valued, your camera will use a worse interpolation algorithm than
what a good piece of software would do on your PC.

--
Ron Parr
Digital Photography FAQ:
http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

--

Of course, with the extreme high cost of taking two images, it's prohibitive to take one at full regular zoom and one at full digital zoom, then use what results in the best shot. ;-) Oh, and perhaps end this conversation once and for all, until the next time.

Cheers,
Eric

 

[Ron Parr]

Of course, with the extreme high cost of taking two images, it's
prohibitive to take one at full regular zoom and one at full
digital zoom, then use what results in the best shot. ;-) Oh, and
perhaps end this conversation once and for all, until the next time.

Taking two images wasn't on the table.

At some point it was suggested that one could use the digital zoom to meter, and then take the shot at regular zoom, but this did not imply actually clicking the shutter - just doing AE lock.

--
Ron Parr
Digital Photography FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

 

[tkz4]

nada

 

[tkz4]

nada

 

[denomolos]

Bull... it's all bull. The golden rule is NEVER use digital zoom IN
THE CAMERA! The built in zooms suck! They blow all the way! I've
never seen a good one yet.

Try making all shots at 12x optical and then use post processing
with such programs like Shortcut PhotoZoom (formerly known as
S-Spline Pro). Try it. Trust me. Just try it. You will never regret
it. This is the software that will kick ass if you produce nice and
sharp photos with your camera. If you have a crappy photo, don't
expect the digital zooming software to do the job for you. However
a shot with a 300D/DRebel, with a decent lens and with the Fine
setting, will enlarge so beautifully you will cry.

Trust me. I'm not here to boast but to prove that i know what i'm
talking about i can say that my MSc thesis was about linear and
adaptive enlargement of digital images.

--
Adam Kozlowski Photography - http://www.spooky.pl
Creative and fun outlet - http://spooky79.deviantart.com

Then you must have some example images to prove your point, so prove it.

 

[alexandr6868]

Try making all shots at 12x optical and then use post processing
with such programs like Shortcut PhotoZoom (formerly known as
S-Spline Pro).
--
Adam Kozlowski Photography - http://www.spooky.pl
Creative and fun outlet - http://spooky79.deviantart.com

I haven't try that program. I surely will give it a try. From what you say it must do much better job than resizing with PhotoShop or PaintShopPro.

George

 

[William Saw]

I want to thank NicksGarage for starting this thread.

Until now I have always thought that the digital zoom was a stupid marketing thing and I could always crop the picture later. But as many people have pointed out it will save me time in cropping the photo later. More importantly it will help me compose the shot in a way that I might not have otherwise been able to. Since I have never used the digital zoom I will have to look at how my camera does it. If it works like the S2 was described then it seems harmless to leave it enabled since I can avoid inadvertently using it.

This discussion seems to have gone away from the original intent. Some people are arguing that in camera digital zoom will always produce poorer results than post processing from a clean image with good software. Fine. You are right. But I think the differences are probably insignificant. Once you choose to save the file in jpg format in the camera you have already decided to loose data. Just as optical zoom is better than digital zoom, raw format is better than jpg. And if I had 60megs of flash memory for my camera I would take everything in raw. It’s a trade off. I choose jpg, and now I think I’m willing to try digital zoom out.

Thanks
--
The Dude

 

[Greg Henry]

Play around with the levels and contrast a bit and you can bring out a bit more detail as well...

 

[NicksGarage]

I appreciate everyone's comments in this thread. Even the experts. We all learned something and everyone can make their own decision about the usefulness of digital zoom versus doing post processing.

I still stick with my original statement. Don't be afraid of it.

Now go out and take some pictures!

Nick.

 

[tom graham]

Ron replys-

This is a flawed comparison because you aren't upsampling the image at all when you do this. You're comparing your camera's upsampling with NO upsampling.

Ok, but is not this the technique people are using ? That is, simply cropping the image to make it appear larger on a given size print? Maybe I've misunderstood the whole thing. If I do it your way, the right way, do I have to get that bicubic-hyperfocal-nyquist-quantum (BHNQ) upsampling program and run it with Linux?
regards - tom

 

[Ron Parr]

Ron replys-
This is a flawed comparison because you aren't upsampling the image
at all when you do this. You're comparing your camera's upsampling
with NO upsampling.

Ok, but is not this the technique people are using ? That is,
simply cropping the image to make it appear larger on a given size
print? Maybe I've misunderstood the whole thing. If I do it your
way, the right way, do I have to get that
bicubic-hyperfocal-nyquist-quantum (BHNQ) upsampling program and
run it with Linux?

I'm not sure what most people are doing.

If you want to display something larger on screen, then you have no choice but to upsample. If you have a 640x480 image and you want to display it at 800x600, you have to upsample or include some text with your file asking viewers to zoom in some other way.

It's generally accepted procedure to upsample before printing if you would have low PPI at the chosen output size. This is covered in the Luminous Landscape article I mentioned above.

If you do this in Photoshop, then you would want to select bicubic resampling when you resize. Now, someobody in this thread has already claimed that bicubic isn't as good as what the camera does (surprising since the camera should only be able to run a less powerful algorithm than bicubic). If we take that comment on faith, then you might consider something else. Even irfanview (great, free, image viewer BTW, if you don't already have it), offers some more sophisticated resizing options such as Lanczos.

Of course, you could just use digital zoom and not worry about it if you're happy with the results.

Perhaps you could also humor me by keeping in the back of your mind that you could probably find something better if you were willing to put it some time and experimentation.

Here's a link for irfanview if you haven't heard of it before:

http://www.irfanview.com/

--
Ron Parr
Digital Photography FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

 

[wazza]

Perhaps you could also humor me by keeping in the back of your mind
that you could probably find something better if you were willing
to put it some time and experimentation.

Agree with all you say but for many (if not the majority) of the masses, digital zoom is fine and just works without delving into post processing.

Time and experimentation is something that many just don't want to use.

I know my daughters are happy with 320x240 images for the net. They never print, they certainly never PP and digital zoom works for them.

I'm a little more discerning but at the end of the day they probably get more use and viewing pleasure out of their shots than I do

Regards

 

[tkz4]

That is one of the best straight out of the camera (no converter lens) moon shots I have seen. Nice joc.
TK

 

[tkz4]

I guess I used moon shot and joc in the same post... not good

 

[Greg Henry]

I did not take this shot, guy. I was showing the original poster how tweaked a bit more, his shot looks even better.

 

[erichlund]

Of course, with the extreme high cost of taking two images, it's
prohibitive to take one at full regular zoom and one at full
digital zoom, then use what results in the best shot. ;-) Oh, and
perhaps end this conversation once and for all, until the next time.

Taking two images wasn't on the table.

Table??? I apologize if your reaction was to my "flip" answer.

However, I believe the OP was simply saying to give Digital Zoom a try, which led to the "usual" debate as to whether or not to use digital zoom. Therefore, it is not invalid to offer the alternative of doing both and use the picture that works best. Certainly, this is just another form of bracketing, which I don't think you will deny is a valid photographic technique.

I recognize that there will be times when you have only one chance to get the photo. In those cases, it's likely you don't have time to get into digital zoom, so you just spin to max optical and take the shot. But if you're taking a shot of the full moon on a clear night, you likely have the time to try several things out. Doesn't it seem reasonable to give yourself the best possible chance of success? And, if a reasonable number of tries results in a common best behavior, then experience allows you to eliminate the lesser option.

Meaning no disrespect, I'd rather rely on results, than take the word of someone I've never met. That does not invalidate the concept of teaching, but teaching is reinforced by practice: both by negative results and positive results. The result is experience, which then qualifies you to teach, but not demand acceptance of you're experience. You impart knowledge and encourage the "students" to gain their own experience.
--

Cheers,
Eric

 

[Ron Parr]

Table??? I apologize if your reaction was to my "flip" answer.

I thought you misunderstood the suggestion about using the zoom for metering purposes.

Meaning no disrespect, I'd rather rely on results, than take the
word of someone I've never met. That does not invalidate the
concept of teaching, but teaching is reinforced by practice: both
by negative results and positive results. The result is
experience, which then qualifies you to teach, but not demand
acceptance of you're experience. You impart knowledge and
encourage the "students" to gain their own experience.

I don't want anybody to do things just because I say so; I want people to understand.

Experimentation is fine too, but it's not a substitute for understanding since experimentation without insight just leads to bad experiments and deeper confusion.

--
Ron Parr
Digital Photography FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/