I have more then two gigabytes of travel photography. Each image takes up than 4 to 5 megabytes. I have to send them to some one. How can I resize all images while retaining high image quality?


5 Answers 5


As is often the case, "there is no free lunch".

JPEG is the de-facto standard image format, and uses lossy compression. That means that to get smaller file sizes, you will lose image quality. The only question is how much you lose, and whether that loss is acceptable.

That said, there are lossless compression techniques that are used by other formats (tiff and png are most common). With these formats, it is possible to reduce the file size without losing quality, but it is rare to achieve the same level of compression as with a lossy format.


I think you're going about this the wrong way. If you have 2GB of images at 4-5MB each, that's somewhere between 400 and 500 images. That's way too many. Even your close friends probably don't want to wade through all of that.

Instead, go through and pick out the very best 10%. Or 5% or even 1%. Take some care and write a meaningful caption for each one. An album of 40-50 good, interesting photos is better than a big dump of half a thousand unsorted ones.

Choose images which show the story of your trip, and which present a unique point of view not available in postcards. You may find that with this reduced data load, you don't need to resize at all. Or, if you do, since you have fewer photos to work with, you can be more careful about selecting the best parameters for each one individually rather than mashing them all down in bulk.

I'd suggest an online web gallery service for sharing, rather than e-mail. Upload once and all your friends can look.

Best of all, by choosing the best images and not bothering with the others, you are literally increasing the average image quality and saving bandwidth and storage. It's a win all around!

  • 5
    \$\begingroup\$ Nah - he's just got 40 or 50 D800 images :-). \$\endgroup\$ Commented Jun 4, 2012 at 14:28
  • 3
    \$\begingroup\$ One ought to give a bit of consideration to what is meant by "high image quality" as well. A 4-5MB JPEG is printable at a respectable PPI count on an A3+ sheet/magazine double-truck. That's great if you're sending them out for print publication, but it's just a wee bit of overkill if you're sending them to a friend. Downscaled low-compression images will have higher apparent quality that overcompressed large images. \$\endgroup\$
    – user2719
    Commented Jun 4, 2012 at 18:29
  • 1
    \$\begingroup\$ I'm NOT going to downvote this BUT I disagree with Matt and am a bit sad that so many people like this answer. This answer says that the question is invalid and proposes a new one - ie send less photos. While that is good general advice, it in no way addresses what was asked. \$\endgroup\$ Commented Jun 11, 2012 at 8:08
  • 3
    \$\begingroup\$ @Russell — but I do think it really answers the question. I think it's the best way to get 2GB of photos down to a reasonable size to share, and will actually increase quality in a meaningful way. Think of it as a lossy compression technique similar to JPEG, in that it discards the parts of the data less meaningful to humans, except unlike JPEG, it's content-aware and works across a whole set of images. :) \$\endgroup\$
    – mattdm
    Commented Jun 11, 2012 at 10:58
  • \$\begingroup\$ The only real downside to this is that it is content-aware: you can't just use software to batch apply the operation to the whole set of images. \$\endgroup\$
    – user
    Commented Jun 11, 2012 at 12:04

Summary of requirement:

  • I have more then two gigabytes of travel photography.

  • Each image takes up than 4 to 5 megabytes.

  • How can I resize all images while retaining high image quality?

  • Extra comment: Image compression and quality:

    For comment on matters related to image compression "quality", and terms such as "lossy compression", "lossy compression", "high quality" "flicker comparator" and similar,
    please see the "Compression & Quality" mini-tutorial at the end of this post.

"High image quality" is of course subjective.
The only way to compress files while retain the identical quality is to use a "lossless compression" system which does not destroy image information content but which uses some means of coding the available information into a smaller format.

If jpegmini does what you want then the
free and fabulous and well established Irfanview probably will too.
And so will many other free programs, using appropriately applied compression systems.

Irfanview jpg90 setting produces compression ratios that vary with image content and camera setting but which produce an image quality that takes intensive "pixel peeping" to tell apart from the original. For all except the very most discerning of applications * jpg90 works very well.

For explanation of the basis for this obviously subjective and probably somewhat contentious claim, and a description of how supporting measurement was made please see "Achievable compression quality at JPG90" & "Flicker Comparator" below,

Minimum compression I get except in special cases is about 2:1 and 10:1 can happen (rarely) with some images. eg Garden of small many coloured flowers may compress almost not at all and helicopter far off in blue sky may be massively compressed.

On 6000 x 4000 = 24 mp Sony A77 images JPG90 compression gives typically 4:1. If you can tolerate a 1/2 res = 4240 x 2830 = 12mp image then you get typically 8:1 compression for sending to people. This will print in A4 at over 300 dpi which is more than can genuinely be used by most colour printers = a long and loud discussion for another forum :-). So eg 2 GB of travel photos would typically reduce to 256 MB with an image size filling more than 4 x 1080p monitor screens.

  • If you really really care then maybe JPG90 will destroy data you'd like to keep. As long as you keep the master file JPG90 is usually good for most other purposes.

Achievable compression quality at JPG90

JPEGMINI is not an obviously useful improvement on eg Irfanview at JPG90 setting.

I downloaded a jpegmini trial image from their site (creative commons licence) and their compressed version. I save dthe original using Irfanviews at jpg90 and jpg85 settings. At jpg90 the files size is modestly greater than jpgmini achieves and at jpg80 it is essentially idential (!1.9 MB down from 9.97 MB for the original.)

jpgmini and irfanview jpg85 were indistinguishable in result using a flicker comparison test and very close scrutiny at 100% crop (ie monitor pixel per image pixel).

The original / jpgmini / irfanview jpg85 produced an exceedingly minor variation in visible quality in a few places using 3 image flicker comparison. The difference in each case was so small as to be virtually undetectable anywhere.


All the jpgmini examples that I saw on their site were horrendously contrived and were clearly going to offer massive compressioin ratios with light jpg compression. They contained repetitive elements that were amnenable to compression - this statement based on my experience in using jpg compression. The one image that I tested was the "dart board on wall" shot but the "man on crossing" and "strawberries" photos would behave similarly.

Flicker comparator:

A "flicker comparator" is a tool used by astronomers to allow very very minor changes in images to be spotted. Two images to be compared are viewed alternately so that if they were identiucal a static image would appear. Any differences "flicker" as the image changes to and fro. The same principle can be easily achieved using a browser which will step on to the next image when a key is pressed or can be set to do so at a timed rate. You effectively set up a 2 slide slide-show.

Load two images into a directory. Point browser at directory. Step on seqentially. Crude. Effective. Image toggles between two images ANY differences that my eye can see are visible as a flicker. If you look for a flicker that is so minor as to almost be imperceptible then when you try static side by side viewing it is impossible (for me) to see the differences by an means. ie flicker comparison is better than straight pixel peeping.

As a test I took a 6000 x 4000 image and viewed it on an LCD screen with about 2000 x 1200 resolution. I expanded the picture to pixel per pixel (100% view) so I was seeing the top left of the image. Changing one pixel in dark hair to red produced a tiny but unmissable flicker. Adding the one pixel dot to orange brown skin made it hard but not impossible to spot. Increasing size to 4x4 = 16 pixels made an ugly blur that would be flicker visible with about any colour combination.

Compression & Quality

"Lossless compresssion:"

As a trivially simple example the sequence
"1000000000000000000001" could be expressed as
Even in this "ASCII" type encoding using 7 bits per character "file" size is reduced from 22 characters to 8 characters.
Obviously such a trivial scheme would not be very effective in mnay cases, but it demonstrates what is meant by "lossless compression". Systems such as TIF offer lossless compression.

"Lossy Compression:" system offer a tradeoff between loss of visual quality and/or of technically measurable quality and file size. Initially an image can usually be substantially reduced in size while removing only small amounts of information so that the majority of users in the majority of cases will not notice the changes. Experts usually will. As the % of files size reduction is increased more and more information is lost, the formal measures of quality get worse (eg more signal to noise ratio) and the number of people who notice and and the number of images which are visually affected increases.

Some compression systems offer variable degrees of compression to suit various degrees of quality/size tradeoff and a number of different means of compression are often used are within such systems - often the user is not made aware of which systems are being used internally.

Measuring or discerning quality. There are many formal means of assessing quality and, to some extent, the affect of a given degradation is subjective. Some systems make use of the characteristics of human eye-brain or ear-brain operation to perform compression iun a manner which aims to be less perceptible for a give compression size by "hiding" the compression in less noticed places. An excellent example if MP3 audio compression which performs lossy compression of certain lower level signals when other higher level signals are also present which are deemed likely to mask the changes in the low level signal. eg when the French Horn is producing 110 dBz signals at 2 kHz how well will you notice lossy compression of 60 dBz Violin harmonics at 16 kHz? (dBz = some relevant logarithmic measurement system).

Subjective quality / Flicker comparator: One subjective method of assessing quality is to compare two images by displaying them sequentially on the same screen etc and swapping between at a rate of say a few times per second. If the images are absolutely identical no changes will occur. If there are significant differences in some locations large "flickering" effects will occur as the displayed image changes. If there are a few extremely minor differences then "flickering" will still occur but it will be more subtle. The eye-brain system is very good as detetcingh very small changes which occur in this manner and 'flickering' due to very small changes can be detected when a few pixels change by a few percent in a single location in an image of say 20 million pixels. How large "few" is in each case varies with user and experience, but such a system will reliably detect differences far below the limits which are liable to be detected by te same user using a simple side by side comparison.

This method is so effective that it has long been used by astronomers for examining start field images taken days weeks or months apart to look for changes caused by events such as the passage of a distant planet across the stellar backdrop (which is how eg Pluto was finally located), or the first faint appearance of a new comet - often detectable weeks to months prior to being immediately obvious by direct observation.

So "flicker comparison" is often a good way of determining if a compression system has affected an image significantly from a given viewer's point of view. If a flicker comparison shows little or no difference then a compression system can be reasonably described as "high quality" for that user. There will lways be effects which are less obvious on flicker comparison thatn they are to the eye-brain, but generally it's a commendably effective system given its ease and simplicity.

  • \$\begingroup\$ The phantom downvoters have given this two downvotes. Would somebody like to have the courage of their convictions and be useful as well and say why? The advice is good, it addresses the original question better than Matt's +15 "You don't want to do what you want to do - do something else instead" answer. Problem = ? \$\endgroup\$ Commented Jun 7, 2012 at 4:06
  • \$\begingroup\$ I dunno. I voted up. :) Maybe people take issue with your characterization of the amount of loss in the "jpg90" setting? Your point about the contrived examples is particularly helpful. \$\endgroup\$
    – mattdm
    Commented Jun 7, 2012 at 11:19
  • \$\begingroup\$ @mattdm - +3/-2 which is ludicrous. I'm not sure how I can 'characterise' the differences better in subjective terms. Load two images into a directory. Point browser at directory. Step on seqentially. Crude. Effective. Image toggles between two images ANY differences that my eye can see are visible as a flicker. If you look for a flicker that is so minor as to almost be imperceptible then when you try static side by side viewing it is impossible (for me) to see the differences by an means. ie flicker comparison is better than straight pixel peeping. \$\endgroup\$ Commented Jun 7, 2012 at 13:01
  • \$\begingroup\$ You may also be getting "fallout" votes from people who disagreed fundamentally with the premise of the question — which also explains why my non-answer got so many upvotes. \$\endgroup\$
    – mattdm
    Commented Jun 7, 2012 at 15:17

There is a freeware tool called PackJPG based on some recent image research which is able to compress JPEGs losslessly by about 20% with no loss in resolution or quality (basically it replaces the simplistic entropy coding stage after quantization with a more sophisticated context-sensitive one). I've often used it for transporting images to bandwidth-sensitive friends, or for online backup where I pay by the gig. This can be combined with tools mentioned in other responses which produce JPEGs. Certain commercial compressors like Infima's JPACK and StuffIt have similar functionality.


Fast Image Resizer

Get on with your life.

The 0/-2 votes when I saw this answer probably relate to the apparently flippant response. It may not be obvious that there was a link to an extremely useful free tool. So, consider something like:

Fast Image Resizer - free, flexible, very easy to use:

A program which is liable to meet your requirement is "Fast Image Resizer".

You can download a fully free copy from here

This program allows drag and drop resizing with flexibility in image sizes produced,
destination (folder, email, ...) and much more.

They say:

  • Multicore processor compatible
    Resize images to any size quickly and in high quality
    Create thumbnails for your website images
    Resize Algorithm quality and JPEG quality configurable
    Create resized files in new folder, or in the same folder as the source
    Use EXIF information to rotate your pictures to the correct orientation
    Copy EXIF info from source image
    Easy resizing by using windows explorer Send To menu
    Automatic cropping option
    Can read JPG, BMP, GIF, PNG, TIFF and HD Photo (.wdp, .hdp) files
    Writes JPG, BMP or PNG files
    Compatible with Windows XP, Windows Vista, Windows 7 and Mac OS X 10.6 Freeware


  • \$\begingroup\$ This is better, but: a) how does it compare to other options, and b) the question asks for no quality compromise. We all know that's not possible, but the answer should address that directly. \$\endgroup\$
    – mattdm
    Commented Jun 10, 2012 at 12:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.