1D X Mark II Owners Unite! Discuss & Post Photos

LJ3Jim wrote in post #18382652
I like sharp images, too.

Flickr creates a number of image sizes from the photo that you upload. It looks like the forum here is using the 1024 size. On Flickr, you have the capability of zooming in to the full 2547x1698 image that you uploaded. I compared the 1024 image on Flickr to the image shown here on the forum. A screenshot of the comparison shows that the Flickr image at that size is identical to the forum image. As would be expected, the Flickr image does become sharper when you look at larger image sizes.

Hope this helps.

LJ3Jim,

Thank you for presenting your theory.
Your theory may be the answer, and I will experiment with it later tonight.
Thank you again for taking the time to come up with your post and for the time & work that went into it.
Your comparison screenshots do help.
Your tips may further my education and add tools to my "toolkit".

In regards to the lion images, first, what a cool image!
I also agree with you that the first image does seem a bit softer, and the 2nd two are similar to each other.

I hope at some point to better understand these dynamics so I can better understand what the owners are posting, and make better decisions or interpretations where applicable.

Mark A. wrote in post #18382747
In regards to the lion images, first, what a cool image!

Thank you. That's Xerxes at the Woodland Park Zoo in Seattle. It's always fun to see him.

The first lion picture looks slightly better, I'd say. The other two a tad oversharpened.

LJ3Jim wrote in post #18382720
Here's a test. The first image below was uploaded to the forum as a full-size jpg (5026x3351). The forum software downsized it to fit in a 1280x1280 box. The second image was reduced to 1280 by Lightroom and then uploaded to the forum. I assume that the forum did not change it. The third image is the 1280 image as created by Lightroom, but it's hosted on GoDaddy. I think the first image is slightly softer than the other two. I don't see any difference between the last two. Your thoughts?

Hosted photo: posted by LJ3Jim in
./showthread.php?p=183​82720&i=i173329823
forum: Canon Digital Cameras

Hosted photo: posted by LJ3Jim in
./showthread.php?p=183​82720&i=i127349659
forum: Canon Digital Cameras

IMHO: Definitely a result of some extra compression/processing done by the forum.

I downloaded the three images, and put them into one layered image in Photoshop.

Flicking between them, there's clearly a sharpness difference between the first two images. The presized image (#2) is ~400KB and the forum sized image (#1) is 386KB. Not a huge difference, but I suspect the resize and compression algorithms used by the forum aren't doing such a good job.

Comparing image #2 with the one hosted by GoDaddy, the latter is ~659KB, and whilst the sharpness difference isn't huge, there's clearly more JPEG artefacts in image #2. That would indicate that the forum is recompressing even the 1280 sized image, and doing a more aggressive/worse job of it than your GoDaddy hosted image (which I assume is unmodified from the source file you uploaded). To be fair, there's probably an eye on bandwidth/storage on the forum; it all costs money.

I did try to prepare some small side by side crops, but they don't show the differences that well when enlarged; though the comparison of #3 (left) to #2 (right) clearly shows JPEG blocking artefacts in the background area to the right side of the image:

if (image needs to be resized [I.e. over 1280 pixels wide])
{
  resize and save new image
}
else
{
  recompress image and save
  if (new image is less than n% smaller than the original)
  {
    use the originally uploaded file
  }
  else
  {
    use the recompressed file
  }
}

sploo wrote in post #18383016
IMHO: Definitely a result of some extra compression/processing done by the forum.

I downloaded the three images, and put them into one layered image in Photoshop.

Flicking between them, there's clearly a sharpness difference between the first two images. The presized image (#2) is ~400KB and the forum sized image (#1) is 386KB. Not a huge difference, but I suspect the resize and compression algorithms used by the forum aren't doing such a good job.

Comparing image #2 with the one hosted by GoDaddy, the latter is ~659KB, and whilst the sharpness difference isn't huge, there's clearly more JPEG artefacts in image #2. That would indicate that the forum is recompressing even the 1280 sized image, and doing a more aggressive/worse job of it than your GoDaddy hosted image (which I assume is unmodified from the source file you uploaded). To be fair, there's probably an eye on bandwidth/storage on the forum; it all costs money.

I did try to prepare some small side by side crops, but they don't show the differences that well when enlarged; though the comparison of #3 (left) to #2 (right) clearly shows JPEG blocking artefacts in the background area to the right side of the image:

Hosted photo: posted by sploo in
./showthread.php?p=183​83016&i=i183305698
forum: Canon Digital Cameras

Take home message: resize to 1280 pixels before uploading (it's better than letting the forum resize it); otherwise, for the best quality; upload an image to another host and link.

NB: The crop above was deliberately resized using nearest neighbor; so each pixel is a single colour. You can see it's definitely been recompressed as each single colour pixel (now a square of pixels) themselves now exhibit JPEG artefacts. There are some on the source file I uploaded, but nowhere near as much. Downloading the file you can see in my post: it's ~36.5KB vs the 38KB original. That would point to the forum's recompression of small (<=1280pixel wide) images not really gaining much in terms of size, and costing a lot in terms of quality. My advice would be a check (in the forum code) that would keep the originally uploaded file if the recompressed version wasn't much smaller. E.g.:

if (image needs to be resized [I.e. over 1280 pixels wide])
{
  resize and save new image
}
else
{
  recompress image and save
  if (new image is less than n% smaller than the original)
  {
    use the originally uploaded file
  }
  else
  {
    use the recompressed file
  }
}

Thank you for the analysis. For quite some time, I've assumed that the forum software modified files in a not-so-good way. Most people wouldn't notice; but this forum is full of, well, photographers.

As you surmised, the image stored on GoDaddy is unmodified from the one on my computer.

LJ3Jim wrote in post #18383047
Thank you for the analysis. For quite some time, I've assumed that the forum software modified files in a not-so-good way. Most people wouldn't notice; but this forum is full of, well, photographers.

As you surmised, the image stored on GoDaddy is unmodified from the one on my computer.

To be fair, a ~660KB to ~400KB reduction would be a worthwhile reduction in file size (in my suggested pseudo code above); though the difference on screen for the whole image (between #2 and #3) isn't that much to my eyes (you need to zoom right in to see the problems).

One way might be for the forum software to not attempt to recompress an image that was <=1280 pixels AND below a certain file size. I.e. a sort of "contract"; whereby if a user gets a jpeg file under a certain pixel and file size (edited and compressed to their personal taste) the forum software will keep it unmodified. Clearly for the case where it's badly damaged an image to reduce a 400KB file to 386KB it wasn't really the right choice.

Easy for me to say of course, probably harder to work out a suitable set of software rules.

sploo wrote in post #18383076
Easy for me to say of course, probably harder to work out a suitable set of software rules.

I wrote software for 35 years. Tradeoffs were made pretty much every day. I have no complaints with the forum software; I'm just trying to figure out why some folks say that their Flickr images look different than the forum image. Today's test is just one clue...

LJ3Jim wrote in post #18383217
I wrote software for 35 years. Tradeoffs were made pretty much every day. I have no complaints with the forum software; I'm just trying to figure out why some folks say that their Flickr images look different than the forum image. Today's test is just one clue...

Ah. A fellow code monkey. Was every day of your career also like this https://www.stilldrink​ing.org/programming-sucks ?

Great tit with 1000mm

13.5 Megapix crop

Is it possible to link spot metering to the selected af point?

sploo wrote in post #18383290
Ah. A fellow code monkey. Was every day of your career also like this https://www.stilldrink​ing.org/programming-sucks ?

The author forgot to mention punch cards and paper tape.
I was fortunate to work (mostly) on good projects with good people. I did have a lot of fun along the way. Retired in 2011.

fordmondeo wrote in post #18383374
Is it possible to link spot metering to the selected af point?

Yes, as long as the AF point is manually selected. The option is at the bottom of the first menu in the camera section (the orange menus).

umphotography wrote in post #18381409
Snoqualmie Falls Washington

Hosted photo: posted by umphotography in
./showthread.php?p=183​81409&i=i160118919
forum: Canon Digital Cameras

Very nice composition.

daughter and her 7 amazon girlfriends at prom

IMAGE LINK: https://flic.kr/p/VPQg​X1