Tufuse Pro vs. Helicon

Discussion forum for Tawbaware's TuFuse and TuFuse Pro software
Post Reply
Chris Strobel
Posts: 1
Joined: Tue Jun 23, 2009 8:33 pm

Tufuse Pro vs. Helicon

Post by Chris Strobel » Thu Jun 25, 2009 5:35 pm

Hi, I'm new to these forums and have been very inspired to try my hand at doing mosaics with focus blending.I've downloaded both demos of Helicon focus and Tufuse Pro.In the three examples below are the results of focus stacks with both programs using their defaults.The Tufuse ones don't look quite right.Is there something I'm doing wrong here?I'm hoping to be able to get Tufuse to match the Helicon results as I can't really afford Helicon's price.Please let me know what I'm doing wrong here .Thanks!......................Chris

Image

Image

Image

dsjtecserv
Posts: 551
Joined: Sat May 05, 2007 3:25 pm
Location: Northern Virginia
Contact:

Post by dsjtecserv » Thu Jun 25, 2009 8:58 pm

After using TuFuse Pro for some time I tried the Helicon Focus demo. I did not give it an extensive test, but I pretty quickly reached the conclusion that they gave comparable results, and that neither fully solves all of the problems inherent in focus blending. So I can't provide a full comparison or decipher all of the differences you see, but I have a couple of suggestions.

The first one is pretty clearly an alignment problem. TuFuse Pro does not contain any facility for aligning photos; that has to be done externally. (If you use the version of TuFuse built in to PTAssembler, that program has the fully ability to align the images before sending them to TuFuse.) I find it simplest to just load the source images into a stack in Photoshop and use its built-in Auto-align function; there is a script to both create the stack and auto-align automatically. Because focus shifting inevitably requires the images to be resized to match each other, you need to crop off the transparent margin that is created on most of the layers. Then use the different script to save each layer as a separate file. These files can be fed to TuFuse Pro, which will then have no trouble fusing the best parts of each image in proper alignment.

For the other two, I see similar problems with both the TuFuse and Helicon versions. The second one has undefined blending around the end of the barrel, and the third has haloing. Both are inherent problems with focus fusion, it just looks like TuFuse and Helicon suffer slightly differently. The first problem occurs when the program is unable to distinguish the source of the "best focused" pixels in a given location, so it includes some of both sources. It looks like the end of the barrel may have been very slightly out of focus, so it couldn't "compete" against the contrast fabric of the table cloth at that location. I don't know about Helicon, but TuFuse uses contrast as an indicator of good focus, and the texture of the fabric has very distinct local contrast and evaluated as better-focused than the end of the barrel. It looks like the two programs rendered the result in somewhat different ways, but I think the source of the problem was the same.

Halos occur when the broad, out-of-focus fringe of a foreground object spreads onto the background. It seems to be worst when the background is intentionally left out of focus. Under those circumstances there isn't a clear "best" set of pixels for the software to choose -- all of the pixels at that location are out of focus. The program does its best to choose, but inevitably some of the pixels from the fringe of the near object get selected, which results in a halo. Again, it looks like both versions are similarly afflicted.

The alignment problem is easily solved; I just make a trip through Photoshop a routine part of my workflow. The "competing focus" issue is mainly a matter of circumstance, but you can minimize it by making sure that there is at least one perfectly focused image for each critical part of the scene. I don't have a good answer for the halos. Avoiding having sharply focused objects in front of unfocused background would minimize it, at the expense of aesthetic effect of a blurred background. It doesn't seem to always happen, but I'm not sure what I have done right when I manage to avoid it! neither of the pictures below had significant halos, despite having out of focus backgrounds.

So, I think you would find that with practice and more familiarity with TuFuse Pro you would get results as good or better than you will get with Helicon. When I tried it, I didn't see a compelling reason to change. Below are a couple of examples of image using TuFuse to expand the range of focus to bring in more of the subject, while still using a relatively large aperture to keep the background blurred.

Image

Image

Dave

maxlyons
Posts: 3649
Joined: Fri Jun 20, 2003 8:55 pm
Location: USA
Contact:

Re: Tufuse Pro vs. Helicon

Post by maxlyons » Thu Jun 25, 2009 9:35 pm

I think Dave gave some excellent advice. The only thing I'd emphasize is that TuFuse Pro has no ability to align your images. And, on most lenses, changing the focus also changes the field of view (similar to zooming in or out), which means the images won't be perfectly aligned.

In addition to using Photoshop to align the images (I guess you need a fairly recent version...my Photoshop CS doesn't have that feature), you can also use PTAssembler to do that. PTAssembler has its own Auto-align feature that can be used to output a set of aligned images (and optionally invoke a stacking program like TuFuse from within PTAssembler).

Some folks have asked if I'll add the ability to auto-align to TuFuse Pro. I'm considering it, but I believe that the "auto-align" feature (at least to do it well) is considerably more complex than the actual focus blending itself. In other words, if I added this ability, then more of the program logic would be dedicated to auto-alignment than the actual stacking logic. There isn't anything wrong with that, per se, but it seems odd to me!

Max

dsjtecserv
Posts: 551
Joined: Sat May 05, 2007 3:25 pm
Location: Northern Virginia
Contact:

Post by dsjtecserv » Thu Jun 25, 2009 10:14 pm

I think the Auto-align function (and the script that creates the stack and then aligns it) were new in CS3. It works quite well and I like having the ability to inspect and (if necessary) manually micro-adjust the alignment. But is an extra step, and does create a duplicate set of images which either have to be stored or deleted. I can see where adding the alignment feature to TuFuse Pro would be like the tail wagging the dog. But if you run out of other innovative features for your software, and have nothing better to do, it would be nice to have alignment in TuFuse Pro.

Dave

maxlyons
Posts: 3649
Joined: Fri Jun 20, 2003 8:55 pm
Location: USA
Contact:

Post by maxlyons » Thu Jun 25, 2009 10:51 pm

dsjtecserv wrote:the tail wagging the dog...
That's the expression I was looking for. I was thinking that it would be like buying a car to "accessorize" your car stereo!

Max

PSHRutPark
Posts: 168
Joined: Sun Jan 21, 2007 3:28 pm
Location: New Zealand
Contact:

Re: Tufuse Pro vs. Helicon

Post by PSHRutPark » Mon Jul 27, 2009 6:42 am

maxlyons wrote: Some folks have asked if I'll add the ability to auto-align to TuFuse Pro. I'm considering it, but I believe that the "auto-align" feature (at least to do it well) is considerably more complex than the actual focus blending itself. In other words, if I added this ability, then more of the program logic would be dedicated to auto-alignment than the actual stacking logic. There isn't anything wrong with that, per se, but it seems odd to me!
Max
Max,

Just a thought: could PTA be used to auto align images and the result fed to TuFuse Pro?

PETER
http://www.pastoralsystems.co.nz/photography/

maxlyons
Posts: 3649
Joined: Fri Jun 20, 2003 8:55 pm
Location: USA
Contact:

Re: Tufuse Pro vs. Helicon

Post by maxlyons » Mon Jul 27, 2009 8:47 am

PSHRutPark wrote: Max,

Just a thought: could PTA be used to auto align images and the result fed to TuFuse Pro?

PETER
Absolutely. In fact, that is my recommended solution. I've even tried to make it easy for folks to do that with PTAssembler's Auto-Align feature.

Max

PSHRutPark
Posts: 168
Joined: Sun Jan 21, 2007 3:28 pm
Location: New Zealand
Contact:

Re: Tufuse Pro vs. Helicon

Post by PSHRutPark » Mon Jul 27, 2009 6:05 pm

maxlyons wrote:Absolutely. In fact, that is my recommended solution. I've even tried to make it easy for folks to do that with PTAssembler's Auto-Align feature.

Max
Thanks a relief because that's what I have done in the past. Didn't notice the recommendation in the posts above so was beginning to wonder if there was a reason against that.

PETER
http://www.pastoralsystems.co.nz/photography/

waters
Posts: 96
Joined: Mon Dec 26, 2005 10:38 am

Post by waters » Thu Sep 10, 2009 10:40 am

Regarding Auto align in CS3 etc., I have been doing monopod panoramas, which would seem to eliminate bracketing and any TuFuse possibilities. How much tolerance for movement would Auto align handle? If I could bracket on a monopod it would be interesting, but I suspect that the slightest movement would cause problems.

maxlyons
Posts: 3649
Joined: Fri Jun 20, 2003 8:55 pm
Location: USA
Contact:

Post by maxlyons » Thu Sep 10, 2009 7:39 pm

waters wrote:How much tolerance for movement would Auto align handle?
Autoalign (part of PTAssembler) should tolerate movement reasonably well...at least, it ought to be able to find control points and align the images as well as they can be aligned. But, PTAssembler can't fix parallax errors. If the camera moved enough that parallax errors are visible, then the stacking (in any program) won't work very well.
If I could bracket on a monopod it would be interesting, but I suspect that the slightest movement would cause problems.
If the camera was reasonably stable between shots (using the camera's autobracketing feature and shooting as quickly as possible while keeping the monopod as stable as possible), and the subject matter doesn't contain a lot of foreground, then you may find that this works quite well. I have even had luck doing this hand-held, albeit with a scene containing only distant elements.

Max

dsjtecserv
Posts: 551
Joined: Sat May 05, 2007 3:25 pm
Location: Northern Virginia
Contact:

Post by dsjtecserv » Fri Sep 11, 2009 8:26 am

I shoot handheld exposure bracketed shots fairly routinely. Like Max, I avoid doing that when there is strong foreground against background, but I have still had successful single frame (not panorama) exposure-fused pictures with a strong foreground. You have to sacrifice mirror lockup (which on my camera requires a separate shutter-press for each exposure) and instead go with automatic bracketing. The pictures will definitely need to be aligned before fusion, but that is usually done successfully.

I would think that with the additional stability of a monopod, single frame exposure fusions would be not problem, and if you can make that work for a regular panorama you can proably make it work for an exposure-fused panorama.

Dave

munchmeister
Posts: 38
Joined: Sun Feb 18, 2007 12:52 am
Location: Fort Collins, Colorado

Post by munchmeister » Mon Dec 28, 2009 6:38 pm

You guys are way ahead of me as usual. WAAAYYYY ahead, but I'm replying here to link to this very, very helpful discussion on alignment and fusing.

WWG
Posts: 92
Joined: Sun Nov 26, 2006 11:53 am
Location: Maine
Contact:

Re: Tufuse Pro vs. Helicon

Post by WWG » Wed Dec 08, 2010 3:22 pm

maxlyons wrote:And, on most lenses, changing the focus also changes the field of view (similar to zooming in or out), which means the images won't be perfectly aligned.
Max
I am resurrecting this old thread again because after using PTA autoalign and TuFuse Pro to do many handheld exposure blends I am finally trying to do some focus blends. I am using a tripod and using the delayed shutter feature on my Nikon D90. I used a fixed normal lens, 35mm on the D90, and moved the focus only a about 1 inch across a small cluster of flowers. I am getting haloing, enough to ruin the shot. If I manually rotate the focus ring I can see significant focal length shifts on my consumer level lens. It is clearly visible just by looking through the viewfinder.

I understand that PTA will correct for lateral motion caused by handheld jitter but do you think it should also correct for focal length changes as well? If so I'm not getting there. Since I was on a tripod I initially went directly into TuFuse Pro and got haloing so I put them into PTA and autoaligned them. PTA's output tif also shows haloing.
Werner

WWG
Posts: 92
Joined: Sun Nov 26, 2006 11:53 am
Location: Maine
Contact:

I found the problem

Post by WWG » Fri Dec 10, 2010 11:20 am

The image I was using was a flower that covered about 40% of the image area, everything else was foliage that was out of focus on all the component images. The in-focus area on all 4 shots only ranged over the flower itself. I looked at the control points in the preview screen and realized that most of them were in the intentionally out-of-focus area. I deleted most of the control points except those that covered the intended in-focus area and re-optimized. The haloing disappeared.

In the end I cropped the image heavily afterward. I should have cropped them all before blending.
Werner

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest