interpolate or lower res?

A printer I am obliged to use and who ought to know what he is doing is suggesting using the 'Transform' tool in PS to increase the size of my panorama to 85cm at 300dpi. (it is currently 55cm)

I'm not convinced that:
a. he should be using this tool for such a critical thing
and
b. the image wouldn't print better at a lower res with les interpolation. i.e. don't resize as much, just mostly change the dimensions.

The picture relies on the details in the structures.

Has anyone done any tests of this nature?

I haven't done tests but this seems an odd way to do this. On the rare occasions I have had to increase a file size I usually do it by a series of small steps in the usual resize menu. I have often read recommendations for "Genuine Fractals" software but that seems a bit pricy for a one off picture. Perhaps someone on here has it and is willing to do a single file....

I wonder if PhotoShop does this resizing in steps anyway as part of the algorithm. Same goes with sharpening. I sharpen in steps but I'm not sure I should, it just seems right. bit off topic already

If you are using Photoshop the just resize to the appropriate size using Image>Image Size.

If using CS2 or later then you can enhance the resizing by changing the resampling method to "Bicubic Smoother" - this is reputed to be about as good as Genuine Fractals.

It is not necessary to increase the size in increments.

Thanks John, useful info on CS2 (or 3 in my case). I hadn't spotted the "bicubic sharper" for reducing images. That is the problem with PS/CS etc., so many functions that one only spots a small fraction of them!

In general, however, I think using transform (as Darkmunk's printer suggests) is unlikely to be as good as "proper" resizing. I don't have a printer here to test, and it is unlikely I'll be able to judge from the screen of a laptop!

Cheers, Kris.

I use this way and find it works for me.

It's the recommended way in PSE.

hmmm... thanks for your contributions folks.
BTW bicubic seems to be the default, not bicubic smoothing.
I have just tried bicubic, bicubic smoothing and transform tool in CS3 and they all look identical!
The image has a combination of detail and large gradients (sky). oh, and it is black and white.
It wasn't the most scientific experiment, but still, I think I might be worrying unduly in this case at least.

Bicubic is the default setting. Bicubic Smoother is for increasing the size for large prints. Bicubic Sharper is for maintaing sharpness and detail when downsizing for the web.

You will only see a difference when you make the print, view the image on the web or whatever.

You can test it a bit at home if you want to check how it looks.
Simply enlarge a very small photo (or crop of it) with a factor of 1.55 and you can print that to have a idea.

Hi Darkmunk,
Just lower the resolution to 240.(that will gain some inches.) I'll promise you,
you won't notice a difference.
It is the default setting within Lightroom,btw.

Martin.

For critical sharpness I would suggest:
1) Using the Image resize with Bicubic Smoother
2) Apply a round of sharpening (Smart Sharpen, High Pass - whatever is your preferred method) to sharpen it up after up-sizing. Best to apply to a duplicate layer and change layer blend mode to "Luminosity". Slightly oversharpen, and the reduce the layer opacity as needed.
3) When happy, flatten the image.

I've done this on 6Mp images upsized to 18x12" at 402dpi

Oh, and if the image is being inkjetted, rather than continuous half-tone (e.g. photographically printed), then view at 50% while adjusting the sharpening - it'll give the best approximation of the resultant inkjet output.

HTH!
Matt

Thanks everyone, that's great. I'll have a go at that Matt.

Just to put this into perspective, I have been using PhotoShop since 1992 so I'm fairly handy with it but you do get stuck in your ways
In all that time, I've almost never felt the need to increase the file size of an image, it just seemed wrong ('you can't get something for nothing') and in the multimedia world, downsizing is the name of the game.
But it's interesting to discover just what you can get away with nowadays.

I will still default to increasing the dimensions by lowering the resolution wherever possible tho'. Even a 72dpi image can look as sharp as a tack in the right place.

A 72ppi image (not 72dpi) can certainly look sharp in the right place - on a monitor!

Don't feel you are doing something "wrong" by resampling. Printa at normal viewing distances should be 300ppi.

Epson printers in particular actually default to print at 720dpi so that needs changing for exhibition quality, preferably to 2880dpi.

Pixels per inch refers to your image in the computer, dots per inch refers to the printing resolution of the printer.

I've never used the term ppi for anything. I just use dpi like everyone else. is there a difference? except letting the recipient know you are talking about printing, not monitors.
In printing we use lpi a lot to quickly determine the correct output resolution for a particular device. I can't recall any printer asking me what ppi my image was

2 meter banners are output at 72 er... ppi and they look fine from 6 feet away, but I suspect my panorama will be viewed at monitor distance by the curious so I need the resolution to be pushed as high as possible without being all blurry.