Canon S90 - Impressions

[Michael Slade]

Doug,

Wow, that is a lot of work. I'm still going through the images. Thanks for putting that up. I'm thinking that my PNS review (woah Nellie!), if it happens, will require me to be very systematic. I want to attain the standard you have set.

 

[dhackney]

Doug,

Wow, that is a lot of work. I'm still going through the images. Thanks for putting that up. I'm thinking that my PNS review (woah Nellie!), if it happens, will require me to be very systematic. I want to attain the standard you have set.

That was about one day's work, doing everything manually.

If you set up an automated workflow you could probably shave hours off the time investment.

Things that would help:
- documented processes (enables training and process improvement/optimization)
- Logical and trackable tagging/ID system for cameras under test. This may sound redundant considering that theoretically each camera has a unique mfgr & model, but when you start looking at the insane product line nomenclature of the Canon SD line, things can get really crazy really quickly. From the data standpoint, you definitely want what is called a unique ID for each camera, but I think it would help to have an additional logical ID, somewhat more informal, just to aid the testing process, e.g. mfg/mod/rez/class (DSLR, EV, compact, etc.)
- Dedicated handling bins / boxes / containers for each camera under test. Helps when checking them in, inventory control, moving from test station to test station to table-top sweep (for glamour shots), post-test return shipping, etc.
- Documented, standardized processes for testing (to aid the production of the "how we tested" editorial section)
- Documented standardized camera settings for test shots (to aid apples to apples comparison)
- Documented standardized testing criteria (resolution, frame rate, IQ, ergonomics, etc.) all within the overall framework of an overlanding / expedition perspective, meaning "the exterior surface was rated a 1 out of 5 because it is too slippery when your hands are covered in fish guts" or "the construction and durability was rated a 5 out of 5 because it survived the drop test from the rooftop tent."
- Automated workflow for download, image file naming, tagging, meta data population and cropping.
- Database for test result data entry, compilation and ranking
- Dedicated testing setups that could remain consistent across a long series of tests (months/years). This could be accomplished with a setup subject package that could be put into a case and stored away and retrieved when necessary. The lighting could be documented and recreated, although you would need to account for falloff in the output of the heads over time, etc.

 

[Michael Slade]

That was about one day's work, doing everything manually.

If you set up an automated workflow you could probably shave hours off the time investment.

Things that would help:
- documented processes (enables training and process improvement/optimization)
- Logical and trackable tagging/ID system for cameras under test. This may sound redundant considering that theoretically each camera has a unique mfgr & model, but when you start looking at the insane product line nomenclature of the Canon SD line, things can get really crazy really quickly. From the data standpoint, you definitely want what is called a unique ID for each camera, but I think it would help to have an additional logical ID, somewhat more informal, just to aid the testing process, e.g. mfg/mod/rez/class (DSLR, EV, compact, etc.)
- Dedicated handling bins / boxes / containers for each camera under test. Helps when checking them in, inventory control, moving from test station to test station to table-top sweep (for glamour shots), post-test return shipping, etc.
- Documented, standardized processes for testing (to aid the production of the "how we tested" editorial section)
- Documented standardized camera settings for test shots (to aid apples to apples comparison)
- Documented standardized testing criteria (resolution, frame rate, IQ, ergonomics, etc.) all within the overall framework of an overlanding / expedition perspective, meaning "the exterior surface was rated a 1 out of 5 because it is too slippery when your hands are covered in fish guts" or "the construction and durability was rated a 5 out of 5 because it survived the drop test from the rooftop tent."
- Automated workflow for download, image file naming, tagging, meta data population and cropping.
- Database for test result data entry, compilation and ranking
- Dedicated testing setups that could remain consistent across a long series of tests (months/years). This could be accomplished with a setup subject package that could be put into a case and stored away and retrieved when necessary. The lighting could be documented and recreated, although you would need to account for falloff in the output of the heads over time, etc.

You just described the person that I am NOT. :Wow1:

 

[haven]

Michael Reichmann has a review, in his fashion, of the S90.
He likes it:

"For now though, if you want the world's smallest
and lightest shirt-pocket-sized (SPS) camera that
shoots raw and comes in at under $500, the
Canon Powershot S90 is the one to get."

http://www.luminous-landscape.com/reviews/cameras/s90.shtml

I'm still drawn to the Panasonic Lumix DMC-LX3, but
the S90 is a strong candidate to replace my Fujifilm
Finepix F30.

 

[dhackney]

New York Times has a review out on my two finalists: the Canon S90 and the Lumix GF1.


http://www.nytimes.com/2009/11/12/technology/personaltech/12pogue.html?_r=1&scp=1&sq=S90&st=cse


I went with the S90 due to its lower cost. I still think the Lumix GF1 is a beautiful proposition to fill a long empty hole in the market. Selfishly, I'd love Canon to come out with something similar just because I already know their menu systems, etc.

 

[Lost Canadian]

My wife was looking for a smaller camera, one she could easily pocket and take with her to social events, so we picked up an S90 the other day. I must say, what a great little camera this guy is. I really love the front dial, such a great feature. Anyway what really surprises me is the cameras IQ when compared to my not so old G9. Canon has certainly made strides in just a few short years. I'm not going to go in depth like Doug did, but here are a couple sample shots I just took to quickly illustrate the improvements Canon has made with respect to higher ISO's. The two cameras used here are the S90 and the G9. Both camera's were set at F4 at ISO 800. These are straight out of camera jpgs. Sorry for the framing, they are not exact as I was hand holding.

First the S90

...and here the G9.

S90 at 100%

...and G9 at 100%

And if you're wondering, this was what I was able to get with a RAW file, same settings as above, and running it through LR3. You can squeeze a bit more out of the RAW file but Canon has done a pretty good job of balancing noise reduction and sharpening. If you're looking to save time processing and space on a memory card, the jpg output is nothing to sneeze at.