W3C home > Mailing lists > Public > public-respimg@w3.org > September 2013

Re: Compressive images test

From: Frédéric Kayser <f.kayser@free.fr>
Date: Thu, 19 Sep 2013 19:29:04 +0200 (CEST)
To: Jason Grigsby <jason@cloudfour.com>
Cc: public-respimg@w3.org
Message-ID: <575853326.153787339.1379611744084.JavaMail.root@zimbra71-e12.priv.proxad.net>
Smashing Magazine published a case study where this image upscaling+high compression technique was put to the test and they encountered some pitfalls. 

IJG libjpeg implemented internal downscaling a while ago:
At least Mozilla uses the turbo derivative of this lib. Beware of libjpeg versions higher than 7 since this project has been "hijacked" by a single man and many new features are not in line with the official JPEG team and are not part of the specs (a really sad story), especially regarding the new lossless mode, JPEG-XT will handle this officially in 2014.
Handling large sequential JPEGs shouldn't be a concern, I'm much more worried by simultaneous decoding of progressive JPEGs.

Frédéric Kayser
----- Mail original -----
De: "Jason Grigsby" <jason@cloudfour.com>
À: public-respimg@w3.org
Envoyé: Vendredi 13 Septembre 2013 01:17:38
Objet: Compressive images test

Compressive images[1][2]. I love the idea and fear it at the same time. 

Love it because it would be oh so easy and awesome. Fear it because even if the file size is smaller, I worry that a memory constrained UA would have trouble decompressing images that contain four times the pixels. 

I do not know if this is a rational fear or if I am simply depriving myself of the chance at a long term retina image love. 

So here's the question: 

If we wanted to construct a test that would definitively answer whether or not compressive images are safe to use or not, what would that test look like? 


[1] http://filamentgroup.com/lab/rwd_img_compression/ 
[2] http://www.netvlies.nl/blog/design-interactie/retina-revolution 

+1 (503) 290-1090 o | +1 (503) 502-7211 m | http://cloudfour.com
Received on Thursday, 19 September 2013 17:29:32 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:06:10 UTC