Good idea, but the idea itself might have issues

  • davince21

    (@davince21)


    6 years, 7 months ago

    This seems like a great idea to me, but somehow it doesn’t seem to be doing something quite right. The page size increases from ~2 MB to around 7 MB for me when Base64 encoding the images, and trying to view the page source will hang the entire tab. Also, the amount of requests increases from 42 requests to 44 according to the Pingdom Tools.

    It’s cool that this could improve general SEO scores, but it seems pointless if the page is actually much slower in practice. Maybe selective encoding where it would actually improve things would be better.

Viewing 3 replies - 1 through 3 (of 3 total)
  • Plugin Author macprawn

    (@macprawn)

    6 years, 7 months ago

    Thanks for the feedback,

    So the debate around base64 encoding is one that is still going strong. It will definitely make your page heavier, as base64 encoded images are not compressed or anything. I am a bit puzzled by the increased number of requests, though. Would it be possible to have the before and after pingdom reports? Or to have a live test page to test with?

    Someone else suggested having a setting to only encode images up to a certain size, which I am working on. I believe this would help in this case too.

    Thanks!

    Thread Starter davince21

    (@davince21)

    6 years, 7 months ago

    Thanks for the response! The website I’m using for testing is my own – http://vincentbeers.nl/. I use Pingdom’s Stockholm test server as it’s closer to the web server, which is located in the Netherlands. My site uses a caching plugin but I make sure to clear the cache between tests.

    The results:
    With base64 enabled: https://tools.pingdom.com/#!/bqdSKF/https://vincentbeers.nl/
    With base64 disabled: https://tools.pingdom.com/#!/daIrqt/https://vincentbeers.nl/

    It’s interesting how big the difference in the perceived performance grade is while actually performing worse (size-wise anyway).

    In all of this though, what’s most baffling to me is that the amount of requests doesn’t decrease whatsoever. That’s exactly what would have the biggest impact normally. I can’t imagine it’s counting the data URIs as separate requests, but perhaps that explains that.

    Plugin Author macprawn

    (@macprawn)

    6 years, 7 months ago

    Thanks for this – I’ll definitely check it out. The good news is, your score with base64 enabled is much better than without… As I said, the page weight being much bigger is totally normal, since base64 encoding is not a compact way of delivering images… That’s why I think having a threshold above which images are not encoded based on size makes sense, so you can work out the best compromise for your own website.

    The number of requests is a bit weird – I’ll have to compare both waterfall lists and try to isolate the additional network call. But aside from this one puzzling detail, the rest is quite as expected I think, when using base64 encoding for all your images. (and I now agree that might not be wise)

    Thanks for this – really appreciate it!

Viewing 3 replies - 1 through 3 (of 3 total)
  • The topic ‘Good idea, but the idea itself might have issues’ is closed to new replies.