Corrected "real world analysis of Google’s webp versus jpg"
michael.terretta.com
michael.terretta.com
What the colour gradients (1Mcolors) example in the referred to article (http://englishhard.com/2010/10/01/real-world-analysis-of-goo...) shows is that WebP, at the limits of quality, prefers to blur. The problem is that most real-world scenes are of objects that have texture, and an artificial texture is psychologically more pleasing than blurring away all these high-frequency small detail patterns. Dark Shikari’s article (http://x264dev.multimedia.cx/?p=541) shows this really clearly: in a real-world landscape, WebP loses badly to both jpg and x264 with too much blurring.
Thing is, if you just do the "obvious" quality metric and do a subtraction of the encoded picture from the original, and sum up all the differences, that doesn't encode the psychological aspect.
A blurred image might indeed do well here, because the blur colour will be somewhere in the middle between the highs and lows of the high-frequency pattern, so the error won't be too large.
But a psychologically-oriented encoding may have encoded a pattern that resembles the original frequencies, but not precisely aligned, so the highs and lows may not coincide exactly between encoded image and original. It may thus simultaneously give a worse difference on a subtraction, but please the eye more.
At such a high degree of compression, Jpeg wouldn't be able to get to any level of quality because it boils down to a pixelation in the size of 8x8 (or 16x16). You'd definitely needs a different kind of compression to preserve quality at such high level of compression. Realistically, many image editors offer default compression at 80% so the artefacts are less visible. I don't know how WebP cocmpare to Jpeg at the 80% setting.
I don't know what magical encoder the author is using, but I was easily able reproduce the original post's results (for JPEG, I didn't try WebP). Steps:
1. Download the original image from < http://fc07.deviantart.net/fs21/f/2007/275/f/4/f4e91d2565442... >
2. Open in your favorite image editor.
3. Save it with whatever optimization settings you want, until it gets down to the goal size (in this case, 44.5KiB).
4. Compare to the original.
For example, I used GIMP, with these settings < http://imgur.com/SIasU.png > , which produced this result < http://imgur.com/8Sclu.jpg >. Although the author is using a different encoder, I doubt it's this much better than GIMP's. (edit: with better settings, GIMP's result is < http://imgur.com/8aEPJ.jpg >).
The author updated comments on reddit saying he hadn't been clear, but he was talking about full 1920x1200 in the 45KB file. That's a different story, so I repeated my tests at full resolution.
Using Photoshop's "Save for Web and Devices" option, Photoshop JPEG compression (at "quality zero") introduces a moire pattern, but still does not generate the extreme banding issue shown on the blog, and replicated in your results.
Here is what Photoshop generates in 44,962 bytes, at 1920x1200 resolution:
http://michael.terretta.com/update-corrected-real-world-anal...
I apologize, I'm not familiar with GIMP or why its results might show such dramatic banding.
Updated version: < http://imgur.com/8aEPJ.jpg >
"Save for web" is probably a shortcut to remove those extra data. I wonder if Photoshop is also adding additional smoothing, beyond GIMP's maximum? That would explain why the PS image's background is smoother, but the inner details (like specular highlights) are not as distinct.
Metadata would help explain this, since as a percentage of file size, it'd be a lot more detrimental on the first image at 45K than the BMW at 575K.
But sure seems there's something else going on with the first image's circular gradient; PS seems to be dithering its blocks in a moire pattern while GIMP is showing strong banding even on files w/o metadata.
EXIF (APP1) -> 5328 bytes
Photoshop IRB (APP13) -> 5896 bytes
Adobe XMP (APP1) -> 26284 bytes
Adobe (APP14) -> 12 bytes
Total -> 37520 bytes (36.6 kB)
That's 36 kB of metadata, leaving only 8 kB for compressed image data. This explains the banding seen in the English Hard post. The BMW image on the other hand (http://fc01.deviantart.net/fs51/f/2009/313/1/c/Jesus_Mobile_...): Photoshop Ducky (APP12) -> 15 bytes
Adobe (APP14) -> 12 bytes
Total -> 27 bytes
The existence of the "Ducky" section means that the image was "saved for web" in Photoshop.EDIT: Added numbers for the BMW image.
According to jhead[1], GIMP is not including metadata in the saved JPEG.
At quality "13", the file was 46302 bytes. Using the "purejpg" option on jhead, it dropped to 45592, or -710 bytes.
This 45592 bytes is exactly, to the byte, what the original article found.
Convert with image magick got me only down to 72,634 bytes http://public.adifferentengine.com/files/lossy.magick.jpg
So yeah the encoder does matter and Photoshop's is that much better than GIMPs.
convert testImage.jpg -quality 1% lossy.magick.jpg
I'm in the process of updating the images on the original blog post (I'm the author) using photoshop's jpg compression algorithm, and it should change the results a bit.