Balconies by the river, seen better days.

Apparently the photos are private on Flickr and they look all horribly over-compressed here.
 
What exactly does that mean?

Lack of fine detail, jaggies on curves, tonal compression so that there are steps between small changes of light & dark instead of smooth gradation - the obvious place for this is in bright areas like the surfaces of leaves that all look equally & excessively bright.
 
. . . bright areas like the surfaces of leaves that all look equally & excessively bright.
I understand that (photo 2 I believe?)
Lack of fine detail, jaggies on curves, tonal compression so that there are steps between small changes of light & dark instead of smooth gradation
Struggling to understand your jargon, sorry!

However, what has caused this and what might have been done (or not done) to avoid this? I'd learn from that information rather than from saying only what might be wrong

Thanks, a.m.
 
Have a look at image 1 - the green section of ropes against the black window - for an example of jaggies. Look at the wall below the LH balcony in the same image, following the line of rendering downward - the flat ends of the stones in the wall are a similar density and the rendered area to the right has an almost evenly bright patch with an evenly darker patch above where the fine changes of light and shade are flattened into single blocks of tone.

There's several possible causes, probably all of which came together here.

First is exporting the image using too much jpg compression. The smaller the number selected for output quality, the smaller the image file will be as more and more data is lost in compression. Lose too much data and effects like this become increasingly obvious - this often shows up as banding across skies where the colour changes in steps instead of a smooth gradient. Cure - don't over-compress images on jpg output: I normally use 91 for Flickr where images get exported or 75 for images that don't get resized.

Second exporting with too much sharpening. That can make jaggies more obvious as the software tries to increase the fine contrast between areas that change sharply in tone and shade. Cure - if using Lightroom then output at either low or standard sharpening for screen depending on the nature of the image.

Third allowing Flickr to re-size an image for embedding. Sometimes the resizing algorithms don't favour already slightly over-stressed images, and what looks fine at 1200 X 800 gets really messed up when resized to 900X600 on the fly in a browser. Cure - for uploading to flickr keep image sizes a bit larger (to give more data for smoother resizing) and avoid deliberately embedding the images at smaller resolutions on sites like TP where we'll analyse each others pictures critically. If you want a smaller image to look its best on the web then output it at required resolution and don't allow any further resizing.

HTH.
 
Thanks Toni - but I'm quite confused and rather depressed :(

I CAN see a problem with the rope, yes - but struggle to understand about the wall (but let's not dwell on that)

First is exporting the image using too much jpg compression. The smaller the number selected for output quality, the smaller the image file will be as more and more data is lost in compression. Lose too much data and effects like this become increasingly obvious - this often shows up as banding across skies where the colour changes in steps instead of a smooth gradient. Cure - don't over-compress images on jpg output: I normally use 91 for Flickr where images get exported or 75 for images that don't get resized.

I'm totally lost on this - I haven't come across "jpg compression", or numbers like 91 or 75. Where do these appear and why haven't I see them??

Second exporting with too much sharpening. That can make jaggies more obvious as the software tries to increase the fine contrast between areas that change sharply in tone and shade. Cure - if using Lightroom then output at either low or standard sharpening for screen depending on the nature of the image.

I believe I have been guilty of over-sharpening images (probably still am!) I think I'm overcoming this slowly.

Third allowing Flickr to re-size an image for embedding. Sometimes the resizing algorithms don't favour already slightly over-stressed images, and what looks fine at 1200 X 800 gets really messed up when resized to 900X600 on the fly in a browser. Cure - for uploading to flickr keep image sizes a bit larger (to give more data for smoother resizing) and avoid deliberately embedding the images at smaller resolutions on sites like TP where we'll analyse each others pictures critically. If you want a smaller image to look its best on the web then output it at required resolution and don't allow any further resizing.

Again, I'm a bit lost here, particular the word "embedding" - what exactly is that and when is it applied/used? I shoot in RAW/jpg, although I rarely utilise the RAW images unless the jpg is not satisfactory. Numbers confuse me and I've always taken the view that I need to reduce sizes to save space with saved images.

Looking at a recent photo - RAW image, 6000x4000, 23,414kb; jpg 14,006kb. Obviously I won't want to store saved images that big or space will soon disappear, so, as a rule of thumb (don't know why, just how I've worked), I resize worked jpg to 2000x1333. There is the option in Photoshop Elements to save the image at different levels and I use 9 (high), which, in this case, gave me 1,525kb. At 12 (maximum) it would be 2,893kb. Much more manageable sizes, I would have thought. The former (9) is the image I save at and use on flickr (my photo site!).

I'm not sure how this relates to your work, if at all! Probably not. Am I totally wrong here (if so, which way to go)?

Apologies if this is off-topic, but it's how discussions seem to evolve on forums like this.
 
Apologies if this is off-topic, but it's how discussions seem to evolve on forums like this.

This made me chuckle a little.

I'm totally lost on this - I haven't come across "jpg compression", or numbers like 91 or 75. Where do these appear and why haven't I see them??
And
There is the option in Photoshop Elements to save the image at different levels and I use 9 (high), which, in this case, gave me 1,525kb. At 12 (maximum) it would be 2,893kb.

The number itself doesn't really matter, but you already use jpg compression as the second sentence I quoted described. It seems that Adobe Elephants goes up to 12 - Lightroom rejigs the scale up to 100 - but it's the same thing.

I believe I have been guilty of over-sharpening images (probably still am!) I think I'm overcoming this slowly.

The 'right amount' of sharpening is almost impossible to define - for me, it's enough to crisp up fine detail without the image becoming all hard and obviously sharpened. You should know that there are different times to sharpen too - when processing an image at full size, and secondly when exporting it for presentation at a non-native size. When images are resized, the original fine detail is lost in the scale change and along with it the sharpening, so it's normal to sharpen on export IN ADDITION to the time of processing IF your workflow requires resizing. Sharpening is always the very last part of processing before presentation.

Again, I'm a bit lost here, particular the word "embedding" - what exactly is that and when is it applied/used?

Embedding is just putting a picture into a document like a web page. Any time you post an image to TP you are embedding it unless you only give a link to the image on another server.

Looking at a recent photo - RAW image, 6000x4000, 23,414kb; jpg 14,006kb. Obviously I won't want to store saved images that big or space will soon disappear, so, as a rule of thumb (don't know why, just how I've worked), I resize worked jpg to 2000x1333. There is the option in Photoshop Elements to save the image at different levels and I use 9 (high), which, in this case, gave me 1,525kb. At 12 (maximum) it would be 2,893kb. Much more manageable sizes, I would have thought. The former (9) is the image I save at and use on flickr (my photo site!).

My personal take on image storage, and that of many of the guys here, is that you want to store the very highest quality images you can as a future starting point. For most of us that's RAW format. An application like Lightroom, On1 Photoraw, DXO Photolab, Capture1 etc lets us process the image to make it look the way we want without making ANY changes to the RAW file underneath - non-destructive processing. That means that in the future, if we want to print an image big or reprocess it because our technique has improved or tastes changed, the original image is there untouched allowing limiteless reworks and ensuring that if we want to make an enormous print then all the fine detail and subtle tonality is still there in the image.

Your approach to editing and resizing is fine if you never want to re-work your images or make a high-resolution print for the wall. However you're saving in jpg at reduced resolution, which means that a lot of the original data from the image has been irrecoverably lost. That may not bother you one iota, but IF it does then you might reconsider your approach. To me, storage is relatively cheap, so I have a 2TB hard drive in my laptop, plus a couple of external 4TB hard drives for extra image storage and backup. I used to use a destructive image editor as you do, and sincerly regret doing so, carefully cropping and processing images to save space, thinking it didn't matter, losing part of the records that I can never recover of people no longer alive.
 
Last edited:
Thanks again Toni. I feel somewhat happier now! I take it you're not a fan of Photoshop Elements (Elephants!) I think I misled you when I wrote "store saved images" - I should have written "store edited images" as I always retain the original (RAW or jpg) together with the edit
Cheers!
 
Elements is fine, but I always enjoy a little word play. :) Image editors are just tools, and it's daft to be all tribal about which one to use.
 
Back
Top