how good is Genuine Fractals?

[Nony Buz]

I need to be printing a series of 30x40 prints from my 35mm negs. I just
spoke with my lab and they can provide me with a 59Meg scan for $9.95
from my 35mm neg. The other options I am considering is using a film
scanner along the lines of the Nikon Coolscan 5000 ED and Genuine
Fractals to create my own "high res scans".

The scanner will have three main purposes, in this order:

1: proofing (will no longer get paper proofs)
2: cropping and touch (Senior portraits)
3: making large poster (30x40) size images

Any thoughts on how the 30x40 would look after using Genuine Fractals on
the 20Meg image that would be created from a scanner like the Coolscan
5000?

 

[Bart van der Wolf]

SNIP

Any thoughts on how the 30x40 would look after using Genuine Fractals on
the 20Meg image that would be created from a scanner like the Coolscan
5000?

Personally I'm not too impressed by GF for many images. On some images it
does a good job of faking (too) sharp edges, but on many images it does so
at the expense of 3D objects which will look flat. I also consider it
rather overpriced.
You can judge for yourself by downloading the trial (20 open or saves) of GF
PrintPro 3.0 at:
http://www.lizardtech.com/download/dl_options.php?page=trials .

An IMHO much more realistic rendering is achieved by Qimage (30 day trial
version at http://www.ddisoftware.com/qimage/), and it has a
price/performance ratio that's hard to beat.

Also check this
http://forums.dpreview.com/forums/read.asp?forum=1003&message=7642362 for
some more background on interpolation programs.

A generally valid test is performed by downscaling a good quality image
(e.g. to 25%), sharpen it, and upscale to the original size using the
different algorithms and compare the upsampled version with the original.

Bart

 

[Hecate]

I need to be printing a series of 30x40 prints from my 35mm negs. I just
spoke with my lab and they can provide me with a 59Meg scan for $9.95
from my 35mm neg. The other options I am considering is using a film
scanner along the lines of the Nikon Coolscan 5000 ED and Genuine
Fractals to create my own "high res scans".

The scanner will have three main purposes, in this order:

1: proofing (will no longer get paper proofs)
2: cropping and touch (Senior portraits)
3: making large poster (30x40) size images

Any thoughts on how the 30x40 would look after using Genuine Fractals on
the 20Meg image that would be created from a scanner like the Coolscan
5000?

Actually, there is a better, and cheaper method of increasing image
size. I know it works in Photoshop and it may work as a plug in to
other software, but you'd need to check.

Go to: www.fredmiranda.com.

His Stair Interpolation plug in (only costing about $20) works well.
And at that price you can afford to try it out ;-)

 

[Mac McDougald]

I need to be printing a series of 30x40 prints from my 35mm negs. I just
spoke with my lab and they can provide me with a 59Meg scan for $9.95
from my 35mm neg. The other options I am considering is using a film
scanner along the lines of the Nikon Coolscan 5000 ED and Genuine
Fractals to create my own "high res scans".

The scanner will have three main purposes, in this order:

1: proofing (will no longer get paper proofs)
2: cropping and touch (Senior portraits)
3: making large poster (30x40) size images

Any thoughts on how the 30x40 would look after using Genuine Fractals on
the 20Meg image that would be created from a scanner like the Coolscan
5000?

The Nikon will make 61MB TIFF file all by itself, and with *real pixels*,
too.

 

[David J. Littleboy]

The Nikon will make 61MB TIFF file all by itself, and with *real pixels*,
too.

Yep. But that's 3600 x 5400 pixels or so, which is 120 dpi at 30x40. Since
4000 dpi scans look like mush at 300 dpi, that's going to be seriously soft.
That's a 32x enlargement from the film, which is, basically, seriously
unreasonable.

David J. Littleboy
Tokyo, Japan

 

[Bart van der Wolf]

SNIP

Yep. But that's 3600 x 5400 pixels or so, which is 120 dpi at 30x40. Since
4000 dpi scans look like mush at 300 dpi, that's going to be seriously soft.
That's a 32x enlargement from the film, which is, basically, seriously
unreasonable.

Yes, it is stretching it to the limits, but whether the result is acceptable
depends on the subject/situation. I don't think (hope) people expect that
such an image will have as high a quality as an 8x10in contact print from
large format film.

If we assume that the limiting resolution we can extract from a fine grain
color slide film by scanning, seems to max out at approx. 85 lp/mm, then any
physical magnification over 10-15x will suffer visibly at close inspection
under adequate light. However, that doesn't mean that an 'acceptable' print
can't be made at an even higher magnification...

We may need to trick the eye into seeing more edge sharpness than actually
is available in the image. The printing step itself also introduces losses,
and if we need to interpolate we know that a compromise has to be made
between accurate pixel definition or a smoother rendition. Some algorithms
are robust enough to differentiate between real edges and noise, most of the
time, or at least produce a plausible solution. For many subjects I think
Qimage's Pyramid interpolation does a great job.

Bart

 

[Josef G.]

Excellentttrote:

Actually, there is a better, and cheaper method of increasing image
size. I know it works in Photoshop and it may work as a plug in to
other software, but you'd need to check.

Go to: www.fredmiranda.com.

His Stair Interpolation plug in (only costing about $20) works well.
And at that price you can afford to try it out ;-)

--

Hecate
(e-mail address removed)
veni, vidi, reliqui

Give a try to "Pictura" that can be found at:
http://www.dmmd.net/products/products.htm
Excellent results

 

[GEO Me]

Gevataim, Israel

K. Holim
Rehov HaMeorer 24
1980-81

Geo