Page 1 of 1

picture file size test

Posted: Thu Jan 19, 2012 11:54 am
by Randy Roberts
Thought it might be good to see if there was noticable differences to the pictures posted based on various file sizes. Might help to decide what limits to file size you'll want to go with.
Don't know if this is necessarily a good test picture but it seemed to have a fair amount of detail so I used it.
I made each example a standard 600 X 450 size and then shrunk the file with Adobe Elements 6 using the "save for web" feature and optimized each for 50, 100, 150, 200, and 250 K size.

Re: picture file size test

Posted: Thu Jan 19, 2012 11:56 am
by Randy Roberts
Whoops, I see the size limit is now 100k. Well so much for the test, here's 100K.

Re: picture file size test

Posted: Thu Jan 19, 2012 11:59 am
by Randy Roberts
I think I'll try to post a bigger one anyway.

Nope, it stopped me.
I don't see much difference between 50 and 100 k myself.

Re: picture file size test

Posted: Thu Jan 19, 2012 12:55 pm
by Charlie Schultz
I could not see much (if any) difference either.

Re: picture file size test

Posted: Thu Jan 19, 2012 12:58 pm
by Mark Swanson
Actually the first photo shows better detail to me because the second one is a bit lighter and that washes some details away.

Re: picture file size test

Posted: Thu Jan 19, 2012 4:00 pm
by Tim Douglass
Is that palm in the rosette?

Oh, and I don't see much difference either.

Re: picture file size test

Posted: Thu Jan 19, 2012 4:28 pm
by Randy Roberts
Tim,
Yep, end grain black palm. It eats sawblades like cookies.

Mark,
Any differences are solely in whatever Photoshop does when it maximizes for a set file size. Both started out from the same 10Meg photo.

Re: picture file size test

Posted: Fri Jan 20, 2012 9:10 am
by Mark Day
Yeah surprisingly the 50k seems to have a little more detail.

"Yep, end grain black palm. It eats sawblades like cookies."
Figures. I like the look of that stuff too and Woodcraft has it on sale from time to time. Maybe save it for a blade that is just about used up.

Re: picture file size test

Posted: Fri Jan 20, 2012 12:44 pm
by Mario Proulx
Let's try a larger photo to see if there's much difference. These are 1000x768 pixels, one at around 84k, the other below 50k.

1000_test_lowres.jpg
1000_test.jpg

Re: picture file size test

Posted: Fri Jan 20, 2012 12:48 pm
by Mario Proulx
And here's one at 1200x800, with some fine detail. Again, one is near 100k, the other below 50k....
1200_test_lowres.jpg
1200_test.jpg

Re: picture file size test

Posted: Fri Jan 20, 2012 12:54 pm
by Mario Proulx
Opening each image in a separate tab, then toggling back and forth between them, the difference between the low res and the higher one is quite striking. Anyone else think so?

Re: picture file size test

Posted: Fri Jan 20, 2012 12:55 pm
by Mark Swanson
In each of the examples Mario, I like the first photo the best. Not sure which is which size but the first photo shows the best details to me.
When I open them in different tabs I can see the difference but in the small form here it's the first one that I like best.

Re: picture file size test

Posted: Fri Jan 20, 2012 12:55 pm
by Charlie Schultz
I thought the windings on the E string were a little clearer in the 85k picture. Tough to tell a difference with the spruce though. Maybe monitor res matters?

Re: picture file size test

Posted: Fri Jan 20, 2012 3:02 pm
by Tim Douglass
It looks to me like monitor resolution and quality have a lot to do with it, but there is (at least on my system) a clear improvement in detail at the larger size. I'm not sure it is enough to jump up and down about, but overall I think keeping the size restriction at the 50K level may be detrimental. I'd be interested in hearing from readers who are still on slow connections, though.

Re: picture file size test

Posted: Fri Jan 20, 2012 7:47 pm
by Mario Proulx
Mark, the first one in each post is the smaller one. Are you opening the images up full size to view?

In the spruce, I look at the silking(medulary rays), and with the low-res version they are washed-out, where they're much more defined in the larger one. In the original, full res image(at about 4MB), there's a good bit more detail to the rays and such, but there, definitely, it's not enough of a difference to warrant posting at that crazy-big size. Methinks a 100 or 120k limit is plenty fine. 50k is very hard to get down to and still maintain decent detailing, even in Photoshop(it can get terrible in the freeware resizing stuff) , and let's face it, we're all detail-oriented folks here, correct?

And yes, monitors make a difference!

Re: picture file size test

Posted: Sat Jan 21, 2012 12:19 pm
by Mark Swanson
Yes Mario I did open each up and toggle back and forth. I was just saying that at first glance in the smaller size the first one was my favorite.
I do agree with you on the size restriction, too.

Re: picture file size test

Posted: Wed Jan 25, 2012 2:43 am
by Chuck Tweedy
If you look at the chrome part of the capo below the neck, the reflections are all mangled in the first one. The second one is much better.
And yes the silk in the spruce is a natural pixelator for the JPEG algorithm so it shows very blocky in the smaller image.
On my Mac it is easy to see these differences.

Re: picture file size test

Posted: Wed Jan 25, 2012 2:46 am
by Chuck Tweedy
By the way Mario, what is that little gold dot near the E tuner??