There has been some debate in rhe RedBubble community regarding the printability of RedBubble preview images. It is generally accepted that normal resampling methods will produce a vastly inferior print to that which an original image could.
However, it has been raised in discussion that Genuine Fractals Print Pro , a piece of software sold by onOne Software, employs a resampling algorithm that may be able to produce saleable prints from RedBubble previews.
In order to test this theory, I have taken three of my RedBubble images and compared the originals with resampled versions of their RB preview images.
These comparisons allow you, the viewer, to objectively judge the validity of claims that RedBubble previews can be turned into saleable prints.
Comparison: Original Resolution Images to Resampled RB Previews
Comments are most welcome and appreciated.
Mike Finley, 3 months ago
As expected the major part of the image deterioration comes from the compression artefacts introduced by the size reduction applied by redbubble, and not from the upsampling applied by Genuine Fractals.
Unexpected is how varaiable the damage is depending on the image.
Shaun Johnston, 3 months ago
I would say the major part of the image deterioration comes from the size reduction itself, and not from compression artifacts introduced after the fact. In the end, any upsampling tries to compensate for information that isn’t there, by making a best guess as to what was. Genuine Fractals just makes a better guess than most other software out there does.
RB applies the same compression (or even less I would suspect) to its previews as any other image-centric website out there.
Randy Monteith, 3 months ago
Hi Shaun – How about doing a comparison with our own Buddy Sears enlargement software here is the link to his Journal entry about it
Click Here
Natalie Perkins, 3 months ago
Thank you so much for doing this little experiment Shaun, you rock.
Insofar as Genuine Fractals? Doesn’t pass muster with me. You’ve got to be blind to accept a print enlarged with this as acceptable. I still stand by my argument, if someone is going to go to all this trouble (and Photoshop CS3 can do some pretty impressive enlargement nowadays too!) they were never going to buy a professionally printed work!
I am the natural enemy of the artifact though, they make me shudder.
Martin Tanton, 3 months ago
Software like Genuine Fractals is a godsend when you’re up against a print publishing deadline and all your client can give you is a low resolution image. It turns something totally unusable into something just about acceptable. IMHO it will never turn it into anything good. The old saying ‘you can’t polish a turd’ applies here. All GF will do is give the turn a slight surface gloss.
Helen Bascom, 3 months ago
The re- sampled versions are stinky bad. Thank you for doing all that work for us.
Shaun Johnston in reply to Randy Monteith’s comment, 3 months ago
Hi Randy,
I just gave Pixel Terminator a shot. Pretty impressive stuff. It seems to produce a less jaggy image than the Genuine Fractals resample, however it’s still a case of guesstimating what information was in the original image.
Thanks a lot for the link!
Paul McClintock, 3 months ago
If this is the best that can be done, I’m not terribly worried. Those re-sampled images look piss poor.
Although, I’d be interested to see how line art goes – I’d assume it would be much easier to enlarge an illustration with better results.
Julie Langford
,
3 months ago
Great info Shaun, and proves what I stated in the forum – these programs are ok for small increases, but increasing to a range over about 150% = loss, and sometimes, lots of it. Brilliant comparisons here.