|
My
experience and tests with Canon 500D

In late april 2009, I have bought
the small Canon 500D. The main reasons for this choice is to
have a reflex camera with video (I want to take videos with
the Canon EF lenses!) and the 15 megapixel APS-C sensor, that
allows to make very precise lens tests and to take photos of
distant subjects (birds, wildlife) with good quality. I'm
not going to publish a traditional review of the 500D. Instead, I'll
publish here a "500D blog": what I learn every day
from tests and experiences in the field with the 500D!
Index
Camera
Specifications
|
Camera
|
Canon EOS 500D
|
|
Image sensor
|
15.10 megapixel APS-C CMOS sensor
|
|
File format
|
14 bit RAW (4752 x 3168 pixels, approx
20 MB .Cr2
files), JPEG
|
|
Video capabilities
|
1920 x 1080 (1080P, 16:9) @ 20 fps
1280 x 720 (720P, 16:9) @ 30 fps
640 x 480 (4:3) @ 30 fps
MOV format (H.264 video, linear PCM audio)
|
|
Viewfinder
|
95% coverage, magnification 0.87x
|
|
Autofocus
|
9 AF sensors (1 cross type)
|
|
Metering modes
|
Evaluative, Partial, Spot, Centerweighted
|
|
Exposure modes
|
Program, aperture priority, shutter
priority, manual, video + scene modes
|
|
ISO sensitivities
|
ISO 100–3200 (in 1/3 stop increments), ISO
100
- 12800 in expanded mode
|
|
Exposure compensation
|
+/-2 stops in 1/3 or 1/2 stop increments
|
|
Shutter speeds
|
30" - 1/4000 plus Bulb pose
|
|
X-sync
|
1/200
|
|
Continuous shooting speed
|
3.4 FPS for 9 RAW
|
|
LCD screen
|
3.0", 640x480 pixels TFT screen
|
|
Support
|
SD/SDHC Memory Card
|
|
Battery
|
One dedicated lithium-ion battery LP-E5
|
|
Weather sealing
|
No
|
|
Dimensions (W) x (H) x (D)
|
129 x 98 x 62 mm (5.1 x 3.9 x 2.4 in.)
|
|
Weight (Body only)
|
480g without battery
|
|
Price
|
$ 800
|
|
Announced
|
2009
|
|
Other features
|
live view, anti-dust
|
May
30, 2009: Some photo and video samples
During my trip to Finland, I have
used a lot the Canon 500D, both for videos and for still
photos. These are some samples that show the image quality in
the field and in studio!
The first photo is the fantastic
view from the Wild Brown Bear observation hide n. 4. This
photo is taken through the glass of the hide window,
nevertheless, the image quality is pretty good.
|

|
|
Canon
EOS 500D, Canon EF 24-105mm f/4 L IS USM, 1/60
f/5.6, iso 200.
|
|

|
 |
|
100%
crop from RAW
|
100%
crop from processed photo |
This is a photo taken from
another hide, this time with the 600 f/4 and with relatively
high ISO:
|

|
|
Canon
EOS 500D, Canon EF 600 f/4 L IS USM, 1/250 f/4, iso
800.
|
|

|
 |
|
100%
crop from RAW
|
100%
crop from processed photo |
I must say I am positively
surprised by the image quality at high ISO of the Canon 500D -
it is not as good as the image quality of 1DsIII, but it is
quite good, it even looks slightly better than the image
quality of its big brother Canon 50D. When I'll have the
occasion, I'll try to do a side by side comparison between
500D and 50D!
Do you want full size samples,
RAW and JPEG? here you are :-)
This is a good example of the
extraordinary detail you can get with the 500D, using good
lenses and perfect technique.
Even at 1600 ISO, the image
quality is still good. Of course there is noise into shadows,
but much less than I expected, and with a little of noise
reduction the images are 100% usable.
Here is an example of crazy high
ISO. The full size photo is pretty noisy, but for small prints
and web it is still usable. I'd try to avoid ISO 12800 as much
as possible on the 500D; nevertheless, if you absolutely need
it and you expose properly, you can still get acceptable image
quality for some usages.
And now, somef video
samples. These videos are straight out of the camera, I have
not applied any post processing; the camera was set on minimum
contrast and sharpening.
Even though the 500D videos are not super sharp, the image
quality is not bad; with nature subject it is difficult to
notice artefacts.
May
29, 2009: White Balance matters!
When I take still photos, I don't
care about white balance: I always keep the WB on Auto and I
select the correct white balance during RAW converstion. With
video, it is not possible: even though you can correct the
color balance during post processing, the result is not as
good as using the right white balance. In natural light Auto
WB works well, but with artificial light, it is easy to get a
result like that:
The image has a strong,
unpleasing yellow cast. In these situations, you have to
select manually the right WB setting; in this case, the
correct setting in "Incandescent"...and this is the
result:
Now, the image has natural,
pleasing colors. If you try to correct the WB of the first
image during post processing, you can get this result:

It is not too bad, but takes some effort
(using the correct WB directly in-camera is much easier!) and
overall it is still not as good as the video taken with the
right WB: there are still slight color casts and the fine
detail looks a bit smeared.
May
08, 2009: Canon 500D vs Canon 350D
After so many tests about video,
today I have decided to try the still capabilities of the
500D, comparing it with the old 350D. This is the test target:
I have taken the same photo with
both cameras, using mirror lock up, 10 seconds self timer,
tripod and Canon 24-105mm at 50mm f/8, its sharpest setting.
The photos had been taken in RAW and coverted with Canon
Digital Photo Professional (currently it is the only program
that converts 500D RAW files). This is a 100% crop from the
500D photo:
And this is the crop from the
350D photo, upsized to 15 megapixel:
The 500D shows clearly more
detail. The 350D is not bad, but the 500D is better! (not a
big surprise here)
Other than detail, I was curios
to compare high iso performance. These are two crops from a
photo taken at 1600 ISO with the 500D and 350D.
The 500D is the first crop on the
left, while the 350D is on the right. Considering that the
500D has much smaller photosites than 350D, I expected a lot
more noise: instead, the 500D shows nearly the same noise
performance of 350D, and more detail.
May
06, 2009: Picture Styles and "video RAW"
While for photos you can use RAW
so you don't have to worry about the setting of sharpness,
contrast, etc, during video these settings become very
important. If your setting give oversharpened video, for
example, you can't remove the excessive sharpening during post
processing, so it is important to set the picture styles on
the right settings: the video quality can be quite good!
This is the standard picture
style. There is too much contrast and an exaggerate
sharpening.
|

|
|
Click
on the image to view it at full resolution, 1280x720
px!
|
To get the best video quality, I
have created a picture style with the minum value of contrast
(-4), minimum sharpening (-4) and very low saturation (-2); I
leave tone on the average value (0). Even tough it is not
possible to get RAW video, with these settings you get a video
that is "as RAW as possible": it has the maximum
possible dynamic range (look at the detail into the white
shells and in the shadows!) and it has no sharpening
artefacts. Other than that, these settings reduce the
"jaggies" problem!
|

|
|
Click
on the image to view it at full resolution, 1280x720
px!
|
This video requires some
post-processing to give the best results; with some curves
adjustment and sharpening you get this:
|

|
|
Click
on the image to view it at full resolution, 1280x720
px!
|
The resulting video is better
than video taken in-camera with the "Standard"
Picture Style; the details looks more natural and it has less
sharpening halos and "jaggies".
May
02, 2009: Low light and bright lenses As
previously said, the 500D is not a great performer
in low light. How can you improve the results? In this
situation, there is only
one way: bright lenses. Since the video is taken at 30 frames
per second, the shutter speed can not be longer than 1/30; the
only way to have more light is to use a bright aperture. Until
now, I didn't like my Sigma 20 f/1.8, but with the 500D I
have changed my mind: for video this lens is an awesome
choice. Here is a quick test! This is a still frame from a video taken with a darker
lens, the Sigma 12-24mm at 20mm f/5.6:
|

|
|
Click
on the image to view it at full resolution, 1280x720
px!
|
Image
quality is almost as bad as the video quality of my crappy
Nokia 5800 mobile phone. And now, let's try the same video
with the Sigma 20mm at f/1.8:
|

|
|
Click
on the image to view it at full resolution, 1280x720
px!
|
There
is a night and day difference! This time, the video quality is
quite good, with much, much better colors and way less noise. The
Sigma 20 f/1.8 was not a good lens on my fullframe 1Ds3, but
on the 500D you don't have problems of vignetting and soft
corners (because the APS-C sensor uses only the center part of
the image, that is the sharpest), moreover, the image is quite small (1280x720 is about
1 megapixel) so it is easy to get a good sharpness. With the 1.6x crop factor, you have about the
same angle of view of a 32mm: it is the widest f/1.8 lens you
can get on APS-C. If you want to take videos in low light, I
highly recommend the Sigma 20mm f/1.8! Other
lenses that would give great results in low light are 50mm
f/1.8, 1.4 and, even better, f/1.2; 85mm f/1.2, and the
majority of lenses with apeture between f/1.2 and f/2...the
brighter, the better.
May
01, 2009: Video Quality...big sensor advantage? Since
video is one of the main reasons that led me to buy the 500D,
it is also one of the first things I have tested. Overall,
video quality is good in day light, while it is so-so in low
light - actually, it is slightly worse than the quality of my previous
HF10. How is it possible? The 500D has a much bigger sensor
than consumer camcorders as the HF10, so one could expect way
better low light performance. Whatching the 500D video
artifacts revealed me a possible explaination: I think that the 500D does not use
the entire surface of the sensor to generate video, instead,
it scans a line of pixels every 4 or 5. In other words,
instead of taking 15 megapixels images and resizing them on
the fly to 1280x720px, it is likely that the 500D during video take directly
1280x720px images, using only a small portion of the sensor
surface. The resulting video is quite noisy in low light and,
even in bright light, sometimes it shows "jaggies"
(aliasing artefacts). To verify
this theory, I have done a simple test. This is a still frame from
video:
|

|
|
Click
on the image to view it at full resolution, 1280x720
px!
|
And
this is a 15 megapixel photo, cropped to 16:9 and resized to
1280x720px:
|
|
|
Click
on the image to view it at full resolution, 1280x720
px!
|
The difference, in terms of crispness and
detail is huge. If the 500D actually used the
full surface of its sensor the video quality would be awesome,
but by using only a row of pixel every 4 or 5, the video
quality is ok, but not as good as one may expect.
Here is an example of the jaggies artefacts:
|

|
|
Click
on the image to view it at full resolution, 1280x720
px!
|
This artefact is likely a result of using
only one row of pixel every 4 or 5; during the video it is not
as apparent as in these still frames grabbed from the video,
but sometimes it is still annoying.
Even though I have not done in-depth tests
with the 5D MarkII, I believe that it uses a much larger
portion of the sensor for video; its videos are much more
crisp and it has better performance in low light. Of
course, the 5D2 costs $ 2700 while the 500D costs $800: you
get what you pay for.
Do you have
comments or questions?
If you have comments or questions about this
article, feel free to ask in the Juza
Nature Photography Discussion Forum!
|