Hello Community,
Why is it that the jpgs are not gzipped and fails, but all the other images (gif, png, etc) are gzipped?
http://www.webpagetest.org/result/140103_59_RW0/1/performance_optimization/#compress_images
Thoughts?
Thanks!
TRX
Hello Community,
Why is it that the jpgs are not gzipped and fails, but all the other images (gif, png, etc) are gzipped?
http://www.webpagetest.org/result/140103_59_RW0/1/performance_optimization/#compress_images
Thoughts?
Thanks!
TRX
That’s not gzip, it re-compresses the jpeg’s at quality level 85 to see if they were compressed at too high of a quality.
Sorry I forgot gzip related to text components ONLY. Can you elaborate on the process and logic it performs flagged jpg (other image types) as a compression pass or NOT?
Also is there a thumb rule for the size of a text component as when to gzip or not?
CDN’s like EdgeCast states “Note that files larger than 1MB will never be served with compression regardless of whether their content-type is enabled for compression.”. So it sounds like compressing text components larger than 1MB does not optimize the data transfer. Is that correct?
Gzip EVERYTHING that is text based, especially if it’s 1MB+
As an added bonus, it’s always good to minify the content, CSS, JS, HTML.
Also, trying to gzip an image has two negative effects, #1, it’s gonna bloat the size, #2 uses system resources for nothing.
As for you images that got flagged, try compressing them in a lossless fashion. You can do it with photoshop or jpegtran.
I also recommend trying JPEGMini on your JPeGs. Maybe you’ll get a better file size
Thanks for the clarification.
What would be your size rule for sending text base files over the network in chunks?
Perhaps because files like HTML, CSS and Javascript rarely reach 1MB, and compression of any file above 1MB (though this is not some magic file size, just an imposed limit) is much more costly in terms of CPU-time than your average text files.
Yeah I guess it varies depending on the config as to apply what rules to when to gzip text components.
What’s would be the thumb rules on using chunk encoding?
Chunk encoding just send the webpages in pieces, if you have css or js file chunked, the page will have to load the complete file in case their is code that needs to be parsed.
Chunked when combined with flush(); (if using apache). It’ll send part of your website as soon as the buffer is ready. But be carefull, if not used properly, visitors will see your page go from text mode, to finished mode. It’s usually a good idea to inline some css of the top part of your site, and put a flush right after it. People will see the banner part and the rest of the site to follow.
That’s why sometimes you see the googlebar on top of page before the rest of the content is loaded.
Thanks.
In what situations are best to use chunk encoding vs. not chunk encoding?
I would use chunk for anything that is above the fold content. Other than that, I would recommend you refactor your site to not use it.