Thanks for the issue.
Tested the issue on windows 7 using chrome version 57.0.2987.110 with the below steps

1.Opened the URL https://research.googleblog.com/2017/03/announcing-guetzli-new-open-source-jpeg.html
2.Observed the image as like in screen shot.

Please find the attached screen shot for reference.
Request you once please try to upgrade chrome to latest stable and if the issue still persists please provide us the actual and expected behaviour to triage the issue further.

Thanks,

Deleted: 702284.png 354 KB

	702284.png 354 KB View Download

Sorry but I don't understand what you have tested.
The issue is a request about the algorithm which is used by Chrome when composing jpg using Canvas API. Is it Guetzli or not? 

Thank you for providing more feedback. Adding requester "kavvaru@chromium.org" to the cc list and removing "Needs-Feedback" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

As per comment #2 considering this issue is a feature and marking it as untriaged.

Thanks!

I am just hearing about this now (it'a a big company, LOL).  Sounds interesting, as does the zopfli encoder for png.  Would make sense to use these as long as the encoder libraries have all the features we need.

@noel: Going forward, are we going to be updating the blink encoder integrations, or are we consolidating with skia?

> I am just hearing about this now (it'a a big company, LOL).

:)

> Sounds interesting, as does the zopfli encoder for png.  Would make sense to use these as long as the encoder libraries have all the features we need.

Note zopfli encoding, and guetzli's computation, are currently slow, which limits its applicability to compressing static content, and that is typically done off-line (rather than in-browser).  Good compression rate improvements, in general, but there's no free lunch (compression time).

> @noel: Going forward, are we going to be updating the blink encoder integrations, or are we consolidating with skia?

Blink _decoders_ are being consolidated with skia, no word on what's happening with the _encoders_.  +cblume@ +scroggo@ might have comments.

There has been some speculation of consolidating encoders, but no active development.

I agree with noel@ that Guetzli is not likely to be something we want in the browser. I've heard of 20 minute compression times for regular-sized jpegs.

It is fantastic at compressing while considering what a human will see. But it is definitely currently most fitting for off-line. 

Maybe it will speed up in the future and we could reconsider then. But for now, it might not be what the users actually want.


I also agree with scroggo@ about encoders. It may be a bridge we eventually cross, when we get there. But for now, consolidating encoders might be a little ways out.

re: encoders, cblume@ scoggo@ a short note about them in your docs would be decent enough reminder of future work.

re: this bug report.  20 minute compression times is not really suitable for the browser user-case.  OK as an idea, but these new algorithms are designed off-line use.