H9 - Its not the sensor losing detail

As far as the squirrel shots in this thread, the first one is a classic H-series AF snafu. The camera looks like it focused on the fence, which is the area of highest contrast. This is typical; this morning, my H1 AF had a love affair with some birch catkins behind a bird I was trying to shoot. The spot focus area is not small enough in general; plus the camera can actually hunt so far as to throw the foreground subject completely out of focus and lock onto contrasty detail behind it. It doesn't happen all the time, but it does happen.

In addition, yes, 1/60s is not enough to stop motion on an active animal. The head moved more than the tail here, which is why it appears blurrier even though it is closer to the focus point. In my tests on birds, I found progressively more feather detail at least through 1/400s on a small perched (but active) bird. Worse than that when they're really moving. IS reduces camera shake, but doesn't do a thing for subject motion. In fact, it should probably be turned off when trying to pan, which is the only way to keep up with a moving subject like this. Canon has a "panning mode" IS; not sure how effective it is. The second shot looks like it focused on a background element, too.

As far as the images themselves, there's a lot of smearing and some noise visible. The colors look very weird and harsh. Exif isn't included with these images, so I don't know what the settings were. To be frank, they look like ISO 400 from my H1, which I would only use in an emergency situation. Out of 20k+ total shots from my H1, I've only kept two that I shot at ISO 400, and one of those was a mistake that turned out to be artistic. The other was in a bad situation anyway, and I needed to get a higher shutter speed to try to stop motion so I fired a burst at ISO 400. Otherwise, ISO 200 always won when I shot at both settings. The H2 and H5 seemed to have improved high-ISO performance by about a stop. That's not to say one couldn't get a noisy-looking image at any ISO with inappropriate settings and why it's really hard to judge a camera from pictures one didn't take oneself.

The H9 appears to have entered the Panasonic zone, where the camera's processor is doing things to the images that aren't always for the better. It looks like some of this can be minimized by choosing correct settings (I wouldn't trust my H1 on full Auto in 99% of circumstances; especially when you go to macro or full tele, the disposable-camera model doesn't apply anymore). There's probably some sample-to-sample variation as well.

--
Wes Stone
Chiloquin, OR
http://wesweb.homestead.com/birdpix.html
 
This picture is just awesome for a non-DSLR camera!
No EXIF, so I am guessing Canon S80... ?

That is what I call good detail, good focus, good depth of field and good exposure! You can see every single detail, and it appears natural!

I've taken many excellent pictures with good detail with a 4 year old 4MP Canon A80 myself. The only concern I had was a slow focus. The camera got lost, and I miss it.
 
Back in the good 'ol days before manufacturers started cost cutting
like crazy and shrinking sensors, 8Mp with a 1/1.8" sensor in a
non-DSLR achieved much better quality images. Today, we see
impressive specmanship with little to show for when it comes to
pure image quality.

A sure sign that a camera is less than satisfactory is when owners
show a reluctance to post large images or crops. Afterall,
practically any camera can be made to look good with liberal
downsampling.

As an example, 8Mp from a non-dSLR, circa 2004. Makes a lot of
today's 8Mp's look like "Fisher Price" quality:

Here is the challenge:

Who can post an H9 picture with the same or better amount of detail, downsized to the same size as Jared's picture (1200x1800)?
 
As an example, 8Mp from a non-dSLR, circa 2004. Makes a lot of
today's 8Mp's look like "Fisher Price" quality:
Olympus 8080? What was the range of the 8080? 5x?

Let's not blame it all on the sensor size. Don't forget the lens. A shorter range lens is always going to give you better quality than a longer range lens.

Next to MPs, the magnification is the other "number" guide the public pays attention to, not realizing that the bigger number may compromise quality. So the makers of these cameras try to satisfy that pursuit by the consumer for the bigger numbers.

There are consumers like me who are aware of what these numbers mean but still decide that we like the convenience and fun of 31-465 in a small and light camera, despite the loss of tack sharpness.

Otherwise we would all still be shooting with our Canon G3s and Oly 8080s. :)

Olga
 
Yes, by 'Processor', I meant the software that processes the images. The chip used is Bionz.

Thanks, Haider.
 
-----As a former Canon-Nikon user I am extremely disappointed with my brand new H9's image processing! It seems like smaller file size was more important when they were designning this camera. A shame considering the fine Zeiss lens!-----

I think there might have been something behind it other than smaller file size.

-----I expect sony to deliver a new firmware version additionally allowing us to choose either RAW or Super Fine for JPEG.-----

It's not up in any of the sports books yet, but I'd make the odds at 8 1 against.

This is my first superzoom, prior digi's being Nikon 990 and Canon S60 then S70. I thought the S60 was sharper with better dynamic range than the 'newer' S70. The 990 in many ways is still my favorite (perfectly throttled flash for macros).

It's just that the H9 is sooo fast handling, it makes the other digi's operation seem glacial in focusing and menu navigation. And, the IQ is acceptable for my non-pro needs.

Sorry for the ramble...
 
Yes, this was shot with the Olympus 8080. But I would disagree that the lens is largely responsible for the poor quality of 2007 8Mp cameras. It's clearly the sensor because all recent 3X cameras with 1/2.5" or 1/2.7" sensors with 8mp+ pixels demonstrate the same poor image quality with respect to detail preservation from noise reduction. This is why with my recent Sony purchase I deliberately stayed away from the 8Mp models and stuck with 7.1Mp. Early sample images from 2007 cameras clearly showed that this was not going to be a good generation for high MP point and shoots.

The joke is that manufacturers are going to start selling us on the idea of 'large' 1/1.8" sensors as a feature when in fact this was standard size several years ago even on the smallest ultra-compact cameras.
As an example, 8Mp from a non-dSLR, circa 2004. Makes a lot of
today's 8Mp's look like "Fisher Price" quality:
Olympus 8080? What was the range of the 8080? 5x?

Let's not blame it all on the sensor size. Don't forget the lens. A
shorter range lens is always going to give you better quality than
a longer range lens.

Next to MPs, the magnification is the other "number" guide the
public pays attention to, not realizing that the bigger number may
compromise quality. So the makers of these cameras try to satisfy
that pursuit by the consumer for the bigger numbers.

There are consumers like me who are aware of what these numbers
mean but still decide that we like the convenience and fun of
31-465 in a small and light camera, despite the loss of tack
sharpness.

Otherwise we would all still be shooting with our Canon G3s and Oly
8080s. :)

Olga
 
Yes, this was shot with the Olympus 8080. But I would disagree that
the lens is largely responsible for the poor quality of 2007 8Mp
cameras.
I didn't say "mostly." But it's another contributing factor.

Olga
Yes, but it's always been a contributing factor. But given the type of image defects we've seen from over-packing sensors, even the best opticals wouldn't save them. The 'watercolor' grass and foliage is not an optical attribute.
 
Yes, but it's always been a contributing factor. But given the type
of image defects we've seen from over-packing sensors, even the
best opticals wouldn't save them. The 'watercolor' grass and
foliage is not an optical attribute.
But you cannot attribute it to sensor alone. Did you see the S5's images? Same sensor, different results. Personally I think there is some internal processing that is causing that effect.

Olga
 
Yes, but it's always been a contributing factor. But given the type
of image defects we've seen from over-packing sensors, even the
best opticals wouldn't save them. The 'watercolor' grass and
foliage is not an optical attribute.
But you cannot attribute it to sensor alone. Did you see the S5's
images? Same sensor, different results. Personally I think there is
some internal processing that is causing that effect.

Olga
The S5 is better, but the images are just less bad rather than impressive. For a better perspective of how things have declined, compare it to the Canon 8Mp Pro1 from 2004 which uses a 2/3" sensor. Many cameras of that generation lack the gritty veil of noise you see on even the best 2007 8Mp low-ISO samples.

While camera prices have dropped over the years, so has quality. By shrinking sensors, makers can get away with using smaller lenses and less glass thus reducing costs. Consumers have demanded smaller cameras and lower prices and are paying the price through lower quality images.
 

Keyboard shortcuts

Back
Top