Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie
Hi there,
There is an issue with role permissions that is being worked on at the moment.
If you are having trouble with access or permissions on regional forums please post here to get access: https://www.boards.ie/discussion/2058365403/you-do-not-have-permission-for-that#latest

JPEGmini Shrinks Photos’ File Sizes Online, Without Reducing Quality

Comments

  • Closed Accounts Posts: 2,131 ✭✭✭oshead


    It's very good, pity there isn't a stand alone app though.


  • Registered Users, Registered Users 2 Posts: 201 ✭✭shemwhistler


    oshead wrote: »
    It's very good, pity there isn't a stand alone app though.

    I was wondering about that, my guess is they want to keep whatever voodoo they're using hidden on their servers, rather than give out the secret source code.


  • Closed Accounts Posts: 2,131 ✭✭✭oshead


    or maybe they are looking to licence the technology to the big online photo websites like flickr etc...


  • Registered Users, Registered Users 2 Posts: 6,393 ✭✭✭AnCatDubh


    But, hold on a minute.....

    JPG is already a highly compressed format so if you are to go from a size as the screen shot says of 2200k to 750k, then you have to be loosing something? No?

    Like RAW to JPG in the first place?

    From the screen shot of the article on lifehacker, to my eyes and non calibrated display, it looks as if there is less contrast - the after JPEGmini shot looks quite washed out.

    Even if my suspicions have any validity, I think it still sound like a handy/nifty wee tool to have.


  • Registered Users, Registered Users 2 Posts: 6,713 ✭✭✭DaireQuinlan


    AnCatDubh wrote: »
    But, hold on a minute.....

    JPG is already a highly compressed format so if you are to go from a size as the screen shot says of 2200k to 750k, then you have to be loosing something? No?

    Like RAW to JPG in the first place?

    From the screen shot of the article on lifehacker, to my eyes and non calibrated display, it looks as if there is less contrast - the after JPEGmini shot looks quite washed out.

    Even if my suspicions have any validity, I think it still sound like a handy/nifty wee tool to have.

    It actually uses the same encoding mechanism as JPEG (as it has to, given that it produces a spec-compliant JPEG file), but they claim to have some super smarty way of deciding which bits of the image to compress more heavily than others. It also works better with larger files apparently (according to the slightly more detailed techcrunch article I read about it). T&Cs on the site are worth a read though if you're concerned about these things, they reserve the right to use your image for promotional purposes etc etc, which seems a bit of a steep price to pay for eking another few KB out of your compressed file.

    Also, dude. It's 'lose', not 'loose'. I never thought I'd see the day :(


  • Advertisement
  • Registered Users, Registered Users 2 Posts: 6,393 ✭✭✭AnCatDubh


    Actually, a good explanation of what is happening here:

    http://www.jpegmini.com/main/technology

    A lot of talk about perceptive and human visual perception. Thus where they are getting their bang from is probably at a technical level not distinguishable to the human eye.

    Also ran an image through it and the quality is way better than the lifehacker example shown. from 1.2Mb to about 250k. Nice one.

    Only issue for them maybe their market in that storage becomes ever cheaper albeit none the easier to manage.


  • Registered Users, Registered Users 2 Posts: 6,393 ✭✭✭AnCatDubh


    Also, dude. It's 'lose', not 'loose'. I never thought I'd see the day :(

    :D


  • Registered Users, Registered Users 2 Posts: 201 ✭✭shemwhistler


    For anyone that's interest I've done a wee test here;

    http://www.flickr.com/photos/mattjeanes/sets/72157627498739727/

    And to be honest, to my untrained eye, I can't see a difference.


    I was thinking about uses for this, and yes with HDD space being dirt cheap, I think I'd always keep a higher quality image at home. I can see it's value being in mobile web usage.

    All these mobile devices; tablets, phones etc are likely running off a 3g network, and with the carriers beginning to get rid of unlimited data plans, the ability to send higher quality images to those devices with little discernible lack in quality is extremely valuable.

    As has already been said, if they license this to any platform that uploads pictures, Flickr, Picassa, wordpress etc, could be pretty huge.

    Edit: Not sure why I didn't think of this first;

    6127075652_84836d1c75_z.jpg
    Original by Matt Jeanes, on Flickr

    6126528209_9b9a6e055a_z.jpg
    Processed by Matt Jeanes, on Flickr


  • Registered Users, Registered Users 2 Posts: 6,713 ✭✭✭DaireQuinlan


    AnCatDubh wrote: »
    Only issue for them maybe their market in that storage becomes ever cheaper albeit none the easier to manage.

    Mobile devices. Probably the biggest growth area in the technology sphere at the moment, and one in which both storage capacity (on the devices) and the cost of bandwidth are both big issues again.


  • Registered Users, Registered Users 2 Posts: 6,713 ✭✭✭DaireQuinlan


    For anyone that's interest I've done a wee test here;

    And to be honest, to my untrained eye, I can't see a difference.


    I was thinking about uses for this, and yes with HDD space being dirt cheap, I think I'd always keep a higher quality image at home. I can see it's value being in mobile web usage.

    All these mobile devices; tablets, phones etc are likely running off a 3g network, and with the carriers beginning to get rid of unlimited data plans, the ability to send higher quality images to those devices with little discernible lack in quality is extremely valuable.

    As has already been said, if they license this to any platform that uploads pictures, Flickr, Picassa, wordpress etc, could be pretty huge.

    r

    It's kinda hard to tell, as that image is pretty soft all over, but if you pixel peep enough you can kind of tell that some of the high-frequency image information is being lost. For all practical purposes I'd say you wouldn't notice it. It'd be interesting to see it done on a sharp image though.


  • Advertisement
  • Registered Users, Registered Users 2 Posts: 201 ✭✭shemwhistler


    Apologies, I'm not entirely sure what you mean by soft and sharp, so I racked my brain for what I considered to be the sharpest image I had on flickr, and came up with this. I hope it's of more use.

    http://www.flickr.com/photos/mattjeanes/sets/72157627498739727/with/6127171534/

    6127169578_a8924a0cc6_z.jpg
    Lego - Original by Matt Jeanes, on Flickr

    6127171534_f24d4e6d0e_z.jpg
    Lego - Processed by Matt Jeanes, on Flickr


  • Closed Accounts Posts: 760 ✭✭✭hbr


    Apologies, I'm not entirely sure what you mean by soft and sharp, so I racked my brain for what I considered to be the sharpest image I had on flickr, and came up with this. I hope it's of more use.

    Rescaled images (smaller than the original) are heavily sharpened on flickr
    and most other photo sharing sites. It would be better to compare a couple
    of 100% crops.


  • Registered Users, Registered Users 2 Posts: 201 ✭✭shemwhistler


    hbr wrote: »
    Rescaled images (smaller than the original) are heavily sharpened on flickr
    and most other photo sharing sites. It would be better to compare a couple
    of 100% crops.

    Sorry, you've completely lost me with 100% crops.

    I'll get me coat.


  • Registered Users, Registered Users 2 Posts: 6,713 ✭✭✭DaireQuinlan


    hbr wrote: »
    Rescaled images (smaller than the original) are heavily sharpened on flickr
    and most other photo sharing sites. It would be better to compare a couple
    of 100% crops.

    The originals are available of the above, if you click through the (rather annoying) all sizes thing that flickr have.


  • Closed Accounts Posts: 760 ✭✭✭hbr


    The originals are available of the above, if you click through the (rather annoying) all sizes thing that flickr have.


    That should be a valid comparison. Looking at the original and processed versions,
    I can see some differences in contrast and noise. I tried using Gimp to re-compress
    the original to approximately the same size as the JPEGmini. The result looks a bit noisier than the JPEGmini version.

    The original is at the top, JPEGmini middle and Gimp bottom.

    173694.png


  • Registered Users, Registered Users 2 Posts: 201 ✭✭shemwhistler


    hbr wrote: »
    That should be a valid comparison. Looking at the original and processed versions,
    I can see some differences in contrast and noise. I tried using Gimp to re-compress
    the original to approximately the same size as the JPEGmini. The result looks a bit noisier than the JPEGmini version.

    The original is at the top, JPEGmini middle and Gimp bottom.

    6034073

    Hi hbr, I'm only getting a broken link, is it me or you?


  • Closed Accounts Posts: 760 ✭✭✭hbr


    Hi hbr, I'm only getting a broken link, is it me or you?

    It was displaying ok here, but that may be because it is in my browser cache.
    The file seemed to be missing so I have uploaded it again. Is it ok now?


  • Registered Users, Registered Users 2 Posts: 201 ✭✭shemwhistler


    hbr wrote: »
    It was displaying ok here, but that may be because it is in my browser cache.
    The file seemed to be missing so I have uploaded it again. Is it ok now?

    Working now thanks.

    I'm curious to hear your assessment, my eye is rubbish, the mini'd one seems pretty good, maybe a bit lighter, and as you say, less noise than the gimp'd version.


  • Moderators, Arts Moderators Posts: 10,520 Mod ✭✭✭✭5uspect


    Perhaps overlay them in Photoshop and change the blend mode to difference.
    I'll make up a test image later and try myself.

    EDIT:
    Okay so here's a quick test.
    I made up this test image:
    6034073

    Nothing special, just a few sharp lines and come colour gradients.
    Saved it as a jpeg in PS at quality 10 resulting in a 260KB file.

    Ran it through JPEGmini and got a 152KB file.
    Saved the original again in PS at quality 4 and got a 144KB file.

    Then to compare I overlaid both the heavily compressed images over the original lightly compressed JPEG and changed the blend mode to difference. There was no immediate difference and a black image resulted. I had to rescale the histogram to bring out the small differences.

    As you can see here for JPEG mini
    6034073

    and the resaved image:
    6034073

    These are pretty tiny differences that you have to go looking for. Both methods seem to give very similar results so I would say form this quick, crude, test you're probably better off just compressing in PS once.
    Would be good to look at some real world examples too.


Advertisement