Neatimage is too slow!

[qtluong]

Yesterday, I tried Neatimage on a 100MB drum scan of a 35mm slide,

and I found the results quite impressive. However, the processing

time was over 6 hours ! This was on a Dell Inspiron 8200 (laptop)

with

a P4M 2.4GHz and 512M of RAM where the only other open application

was IE. This time sounds way over the top. Any suggestions (besides

reinstalling Windows, which I might have to do :-() ?

<p>

The image that I ran the program on had some significant uniform

areas. If that's not the case, does creating device profiles for

each scanner and film and applying them gives good results ?

[daniel_staver1]

Try setting Cr and Cb sharpening to 0. Then it's much faster, but won't remove all the color noise.

[edward_c._nemergut]

It sounds to me like something is definately wrong with your computer: I use a

Dell Insirion 4600 Laptop (1.3 GHz Celeron, 384 MB RAM) to process images all the

time. Normally, for a 200 MB scan from 6x4.5 source takes around 35-40 minutes... I

don't run any other applications at the same time--maybe that's a good place to

start.

 

Let me know if you figure it out...

[mike_lepp1]

You'll find one that sets the priority. It's default is low (or relatively) so you can run other applicatons while NI is processing. Being I run it in batch mode during the night, I set it for highest priority.

 

Another thing I found on my machine, running Win me, was that the processing time went way up when my machine went into standby. Vladimir, at NI, thought that I needed to update Win me with a patch. I did this, but at the same time he also advised me to put the machine into the highest priority mode. I've done batch processing of 10 50MB images and they were all done in the morning.

 

Sorry I'm not at home to tell you exactly what menu to look under. But it's under something like "properties" or "preferences." I'm sure you'll figure it out.

[david_bradley1]

Just a thought: are you by any chance using a CPU-intensive screen saver?

[ky2]

More RAM helps. I got 1 Gig on my Dell 2.2Ghz P4, and the thing is much faster than another similar-equipped P4 I own which has only 512Mb.

[oskar_ojala]

Go to <a href="http://www.photo.net/learn/optics/digitaloptics/">http://www.photo.net/learn/optics/digitaloptics/</a> and look at the comments, there's much discussion about NeatImage times.

 

<p>In my experience, RAM matters only so long as the program can run in the physical memory, ie. Windows doesn't start swapping. Check for this. WinXP uses a lot of memory (if that's waht you use) and I recommend closing IE.

 

<p>Assuming your 100 MB is the size of the raw image data, my 1.5 GHz Athlon does a good job on a ~72 MB image in 10 min. (I haven't had the time to test several images yet, but I wouldn't expect a huge deviation) and even with winXP I think 512 MB should be enough for 100 MB images (assuming you don't have a million useless background apps running).

 

<p>In short, you have a problem with your setup :)

[rich815]

It is WAY too slow! I tried it with image files from my 3-mp Nikon 995 and it took over 5 minutes to process on a WinXP PC with 512MB of RAM. I can do a full-res scan negative or slide in less time on my Nikon LS-4000 on the same computer. I shutter to think of what people with the big MP files are going to do. Does seem to do a decent job eliminating noise but then to my eye it leaves a weird halo-like softness/look to the images. I guess this might explain why many of Michael Reichmann's images on Luminous Landscape have that weird fuzzy glow.

[qtluong]

Thanks for the suggestions. I reinstalled windows, but

ran into system problems while doing it. Technical support

thinks my hard drive is defective and will be sending me a

new one.

<p>

As far as results are concerned, I have found them to be

excellent with no weird artefacts. True, the final files look a bit

soft, but I am not sure they are any softer than before

applying Neatimage. If I had to use Neatimage in the first place,

it is because I am over-enlarging already, hence the softness.

I found nevertheless that once the grain is removed, it is

possible to cure the softness with aggressive sharpening.

<p>

What I was also curious is if anybody had good results on

images without significant uniform areas by scanning a

different slide of the same film which does have uniform

areas, saving it's profile, and applying it to the scan without

uniform areas.

[mike_lepp1]

Hi Quang-Tuan,

 

I make profiles based on the targets provided at NI. I printed the target on white matte paper and "waste" a shot on it with every film (or digital ISO) I use. These work fairly well. Though in a few instances I had to fine tune the profile a little.

 

I also want to point out to one of the posters that there is a learning curve involved with using NI. It's easy to over do it. I tend to error on the under side. In the same vein, I use the sharpening in NI. Again, I error on the soft side because a lot of the times I'll need to resharpen after I size for print anyway.

 

Once you're past the learning curve I don't see the big deal in the time it takes to process. Just set up a batch job and walk away. Does a watched pot boil faster?

[qtluong]

As an update, it turns out that my Dell 8200 was defective.

Dell promptly replaced it with a 8500. The processing time is

now around 15mm, I continue to be pleased with Neatimage results.