Optical vs. digital zoom tests.

[Marty Schultz]

I've been wondering about this for some time and did a few tests with full optical zoom and full digital zoom on the 2100. Several shots were taken of a dollar bill from about 5 feet away at full optical and full digital zoom, in TIFF, SHQ, and HQ. The 10x images where then resized 270% in PSP using bicubic resample. To test something else I've been wondering about I took several shots at full optical in TIFF (bumping the camera slightly between shots) and combined them after resizeing to 270%. After processing they were saved as bmp files and then I cut and pasted from those into a tiff file.

Here's the link to the image (2.3MB):

http://www.science.widener.edu/~schultz/temp/c2100/opticalvsdigizoom.tif

To me it looks like 10x tiff resized in PSP looks the best, even better when several are combined. In HQ the digital zoom image has a slight edge? In SHQ there doesn't seem to be much difference.--Later,
Marty

Olympus D490, C2100

 

[buzzy2]

I've been wondering about this for some time and did a few tests
with full optical zoom and full digital zoom on the 2100. Several
shots were taken of a dollar bill from about 5 feet away at full
optical and full digital zoom, in TIFF, SHQ, and HQ. The 10x
images where then resized 270% in PSP using bicubic resample. To
test something else I've been wondering about I took several shots
at full optical in TIFF (bumping the camera slightly between shots)
and combined them after resizeing to 270%. After processing they
were saved as bmp files and then I cut and pasted from those into a
tiff file.

Here's the link to the image (2.3MB):

http://www.science.widener.edu/~schultz/temp/c2100/opticalvsdigizoom.tif

To me it looks like 10x tiff resized in PSP looks the best, even
better when several are combined. In HQ the digital zoom image has
a slight edge? In SHQ there doesn't seem to be much difference.
--
Later,
Marty

Olympus D490, C2100

...hi, Marty...

...at 2.3 MB, with my 24K dial-up connection...I should be able to get back to you...say...January?...(JK...but not about the lack of broadband...)
newby

 

[Marty Schultz]

...hi, Marty...
...at 2.3 MB, with my 24K dial-up connection...I should be able to
get back to you...say...January?...(JK...but not about the lack of
broadband...)

I'm still using 28.8K at home, took about 10 minutes to upload the image to the server. For really big files I use the broadband at work and use zip disk to take them home.

Some of the subtle differences between the parts of the image would probably be lost with jpeg compression so I saved it as a compressed tiff.

 

[Maxven]

I'd say it's a very close call. For little (if any) noticable difference, I'd personally go for 27x SHQ. Saves a lot of time in postprocessing. I guess digital zoom is not as bad as it's rep.

Thanks for the test, Maxven

I've been wondering about this for some time and did a few tests
with full optical zoom and full digital zoom on the 2100. Several
shots were taken of a dollar bill from about 5 feet away at full
optical and full digital zoom, in TIFF, SHQ, and HQ. The 10x
images where then resized 270% in PSP using bicubic resample. To
test something else I've been wondering about I took several shots
at full optical in TIFF (bumping the camera slightly between shots)
and combined them after resizeing to 270%. After processing they
were saved as bmp files and then I cut and pasted from those into a
tiff file.

Here's the link to the image (2.3MB):

http://www.science.widener.edu/~schultz/temp/c2100/opticalvsdigizoom.tif

To me it looks like 10x tiff resized in PSP looks the best, even
better when several are combined. In HQ the digital zoom image has
a slight edge? In SHQ there doesn't seem to be much difference.
--
Later,
Marty

Olympus D490, C2100

...hi, Marty...
...at 2.3 MB, with my 24K dial-up connection...I should be able to
get back to you...say...January?...(JK...but not about the lack of
broadband...)
newby

 

[Roy F.]

Marty, I enjoyed examining your test results. Maxven,
my conclusion is similar to yours, but different:

I'd say it's a very close call. For little (if any) noticeable
difference, I'd personally go for 10x SHQ. Saves bits.
I guess digital zoom in camera is of little value. Let your
printing program (Qimage) do the interpolation on the
way to the printer.

Roy F.

I'd say it's a very close call. For little (if any) noticeable
difference, I'd personally go for 27x SHQ. Saves a lot of time in
postprocessing. I guess digital zoom is not as bad as it's rep

 

[Marty Schultz]

I'd say it's a very close call. For little (if any) noticeable
difference, I'd personally go for 10x SHQ. Saves bits.
I guess digital zoom in camera is of little value. Let your
printing program (Qimage) do the interpolation on the
way to the printer.

If you need to save storage space and are shooting in HQ the digital zoom might be worth using. For SHQ or TIFF there is no need for digital zoom unless you don't have a good graphics program (or the time to process them).

Guess I should also point out that the camera sharpness was set to soft, the results might be different with normal or hard.

 

[Niko55958]

Your test pictures don't show the jpeg-loss really. I did another similar test with uzi (SHQ).

http://www.cc.jyu.fi/~tanipe/comparison.html
(warning, three 900KB bitmaps on a page)

You take a picture without digital zoom and crop and resize afterwards using software. In this case the jpeg-compression is done to the original picture before cropping and resizing. The resizing will bring the jpeg-artifacts up very clearly.

But with digital zoom, you do resizing to the original picture in the memory of the camera and the jpeg-compression is done after the resize. So the effect of the jpeg-artifacts is much less visible in this case.

So digital zoom really is useful when:
-you shoot jpegs

-you cannot get any closer and your optical zoom is not enough, so you need to frame the picture anyway
-your camera has a digital zoom that also resizes the picture, not just crop

-Niko

 

[misha]

Your test pictures don't show the jpeg-loss really. I did another
similar test with uzi (SHQ).

http://www.cc.jyu.fi/~tanipe/comparison.html
(warning, three 900KB bitmaps on a page)

You take a picture without digital zoom and crop and resize
afterwards using software. In this case the jpeg-compression is
done to the original picture before cropping and resizing. The
resizing will bring the jpeg-artifacts up very clearly.

But with digital zoom, you do resizing to the original picture in
the memory of the camera and the jpeg-compression is done after
the resize. So the effect of the jpeg-artifacts is much less
visible in this case.

I would not though that in your samples the PSP interpolations are sharper than the in-camera digital zoom; if your apply some blur to these pictures they'll probably be close to the DZ.

Misha

 

[Niko55958]

I was thinking about the noise, which is quite bad in those PSP interpolated pictures. Maybe the softness of the digitally zoomed pictures comes from the worse interpolation algoritm of uzi, comparing to PSP.

Well you can always sharpen those digitally zoomed pictures, on the other hand you can always add some edge preserving smoothing to the noisy PSP pictures. I tried both and I cannot say which was finally better. I think I'd stick to digital zoom (versus cropping) since there's less work at least.

Niko

I would not though that in your samples the PSP interpolations are
sharper than the in-camera digital zoom; if your apply some blur to
these pictures they'll probably be close to the DZ.

Misha

 

[L-5]

I've been wondering about this for some time and did a few tests
with full optical zoom and full digital zoom on the 2100. Several
shots were taken of a dollar bill from about 5 feet away at full
optical and full digital zoom, in TIFF, SHQ, and HQ. The 10x
images where then resized 270% in PSP using bicubic resample. To
test something else I've been wondering about I took several shots
at full optical in TIFF (bumping the camera slightly between shots)
and combined them after resizeing to 270%. After processing they
were saved as bmp files and then I cut and pasted from those into a
tiff file.

Here's the link to the image (2.3MB):

http://www.science.widener.edu/~schultz/temp/c2100/opticalvsdigizoom.tif

To me it looks like 10x tiff resized in PSP looks the best, even
better when several are combined. In HQ the digital zoom image has
a slight edge? In SHQ there doesn't seem to be much difference.
--
Later,
Marty

Olympus D490, C2100

I'm on my way to work now but I cant wait to get home this
should prove to be very imformative as well as something to do
when I get home. I can't thank you enough for the test.

 

[L-5]

Your test pictures don't show the jpeg-loss really. I did another
similar test with uzi (SHQ).

http://www.cc.jyu.fi/~tanipe/comparison.html
(warning, three 900KB bitmaps on a page)

You take a picture without digital zoom and crop and resize
afterwards using software. In this case the jpeg-compression is
done to the original picture before cropping and resizing. The
resizing will bring the jpeg-artifacts up very clearly.

But with digital zoom, you do resizing to the original picture in
the memory of the camera and the jpeg-compression is done after
the resize. So the effect of the jpeg-artifacts is much less
visible in this case.

So digital zoom really is useful when:
-you shoot jpegs
-you cannot get any closer and your optical zoom is not enough, so
you need to frame the picture anyway
-your camera has a digital zoom that also resizes the picture, not
just crop

-Niko

Thats food for knowlege. Thanx