Focus Merge Comparison

[GracieAllen]

Windows PC, Windows 10 Pro, GTX 970, 5820 running at 4.5GHz, 32GB of DDR4 memory.  Current version (as of last Friday) of Affinity Photo...

I bought Affinity Photo when it went on sale.  To some extent because I'm hoping for a better tool than Helicon Focus for doing stack processing.

I took a simple, 4-shot series of full-sized jpegs, captured with a Nikon D810 using Helicon Remote so the intervals are consistent, and processed it through Helicon and Affinity Photo.  These were taken outside on a "windless" morning, so there may be very small differences in position between shots.  Nothing visible, but...

No processing was done on any of the images before, or after, the image stack processing, so these are as close to un-messed-with as possible.  Only thing done was to crop and resize down to a reasonable size.

Image "a" is on the left and "f" on the right

The next images are a close-up of an area with "a" on the left and "f" on the right

I'm using the Focus Merge (or image stacking) to create the raw material for further optimization in Lightroom and Photoshop, so I want the sharpest, cleanest output from the stack processing to start with.

From my examination "a" isn't as sharp and has more "haloing" (or whatever you'd call that smear on the edges in the close-up images).  Is there some setting or set of settings I can/should be using to improve the output quality of the Focus Merge?

 

 

 

 

[GracieAllen]

Hello?  No response after a week?  Is there some other channel for getting support on this product?

[John Rostron]

Could you explain what you mean by image 'a' and image 'f''? Are these input images? Is there an image representing the focus-merged final image?

Certainy halos such as you show can be an artefact of focus merging, but I think we need more information. Could you post a link to the four original images?

John

[R C-R]

It would also help if you could explain what you mean by "crop and resize." Are you talking about resampling the images or something else?

[firstdefence]

I would assume that both images have been produced by processing the set of 4 images through Helicon focus and then repeating the same process through Affinity Photo's Focus Merge to make a comparison but which image belongs to which processing?

Image 'a' looks duller and more blurred than image 'f' which is clearer and brighter colour rendition.

Also what were the camera settings?

[GracieAllen]

As the original entry said, I ran a set of 4 images through Affinity Photo and Helicon focus.  "a" and "f" are the outputs of those two processes.  The images below them are close-up views of the images above.  The halos in the image on the left do not appear in the image on the right.

" Certainy halos such as you show can be an artefact of focus merging, but I think we need more information. Could you post a link to the four original images? "

Yes, in some cases they can, but since they only appear in "a", I believe they're not the result of a problem with the input images.  The only variable is which software did the processing.  I was processing a bunch of series of images, and don't know which one I used for the original post.  I ran a new set, which may have been the same ones from the original post and placed them on my website.  I have no idea whether or not you'll be able to download them, but they're on my website at

www.dperezphoto.com

The files are  799.jpg, 800.jpg, 801.jpg, 802.jpg and affinity.tiff

 

"It would also help if you could explain what you mean by "crop and resize." Are you talking about resampling the images or something else? "

Crop and resize means I took the full-sized output of the 7360x4912 px inputs and cropped it to show the relevant area, then resized it so it wouldn't be overly large when I stuck it in the forum entry. 

"I would assume that both images have been produced by processing the set of 4 images through Helicon focus and then repeating the same process through Affinity Photo's Focus Merge to make a comparison but which image belongs to which processing?"

"a" is from Affinity.  Which is why I asked the question in this forum... 

"Image 'a' looks duller and more blurred than image 'f' which is clearer and brighter colour rendition."

Duller, flatter.  Which I presume is the difference in the two software program's default processing.  With enough additional processing I can get the Affinity output to look the way I want, it'll just take a little more clarity, sharpening, and HSL than it will using the other image.  That's not a big deal.  The halos are.

"Also what were the camera settings?"

I don't know why it would be relevant, but it was a Nikon D810, using a Tamron SP 90mm macro, on a very sturdy Gitzo 3540XLS.  Subject was in a windbox.  Nothing was moving.  I shot all the series at 1/200 @f/16 @ISO 400.  The camera does minimal processing on the jpegs.

As I asked originally - I don't see any settings, parameters or adjustments I can change for Affinity to optimize the quality of the output of the Focus Merge, but if there IS something, what SHOULD I be setting to optimize the output?  I don't know if Affinity does this on every series 'cause I don't have the time to process them all through multiple programs and I need usable output from the image stack processing.  But it's happened more than once.  It does seem that the output from Affinity is duller and sometimes appears less sharp than the output of the other focus stacking software.  As I said, I can work around that.  But, coupled with the halos it makes Affinity difficult to use to get the best possible output to use as input for additional processing.

The images and the tif output are on the website.  Let me know.  If the answer is "there's a whole bunch of people at Serif, busting their humps to make the Focus Merge work perceptibly better than Helicon or any other image stacking software, and the improved version will be out in xxx days/weeks" great.  If it's "TS, it's not going to get better any time in the appreciable future", I wasted $40, which certainly isn't the first time.

[Grazie]

I don’t know your camera, but aren’t JPEGS a lossy format? Is it possible that this halo was captured during shooting? Can your camera shoot in RAW? Would that help?

[John Rostron]

Thanks @GracieAllen for your very informative reply. You are quite right in saying that the Affinity Focus Merging has no parameters. Nor does it specify the algorithm. (BTW, which Helicon Focus algorithm did you use?) My impression is that the AP focus merging has improved over the years, but it can still produce haloes. If you look at the tutorial video on Focus Merging, then it explains how you can lose the haloes. It essentially involves cloning from one of the original images.

When I do focus merging, I use AP first, and then Helicon if necessary.  It really depends on the subject matter.

I do not think that @Grazie's suggestion, that it is due to a lossy jpeg, is likely to be the reason for the halos, although starting from raw would probably give a better overall result.

John

[firstdefence]

Quote

I don't know why it would be relevant, but it was a Nikon D810, using a Tamron SP 90mm macro, on a very sturdy Gitzo 3540XLS.  Subject was in a windbox.  Nothing was moving.  I shot all the series at 1/200 @f/16 @ISO 400.  The camera does minimal processing on the jpegs.

I wanted to know what settings you were using because they are important in relation to the end product. If I take stack images i generally shoot at f/8, I know this sounds a bit backwards but I have found going to f/16 can deteriorate an image, I would normally not go any higher than f/11 and would rather take more images, maybe 7 to 10. I would try taking the images at f/8 and reduce the ISO to keep to the 1/200. An added benefit to f/8 is a more out of focus background and sharper flower.

[Grazie]

9 minutes ago, firstdefence said:

 An added benefit to f/8 is a more out of focus background and sharper flower.

Yes, that makes a whole load of sense. I like it.

[John Rostron]

 

18 minutes ago, firstdefence said:

 An added benefit to f/8 is a more out of focus background and sharper flower.

 

7 minutes ago, Grazie said:

Yes, that makes a whole load of sense. I like it.

f/8 is generally regarded as the 'sweet spot' for most lenses, so if you are not constrained by vibration or depth of field, then this is the one to use. (By sweet spot, it is the one with little, if any, diffraction effect that degrades the image, whilst giving a good depth of field.

I will have a play with your original images to see if I can get rid of the halos.

John

[John Rostron]

I downloaded your three images and used Affinity to stack them. I then cloned the haloed stamens using original image 801. To do this double-click on each image  in the source stack, then wait a moment for it to render. You can then see which one has the sharpest focus. I used 801 and just cloned the stamens. Here is the cropped image.

This took just a few minutes.

I would suggest that Affinity will do your focus stacking for you. Helicon Focus might do it with less extra editing, but then you are paying for this!

John

[Grazie]

Nice one John. I come from a Video background, where focus<>lighting is paramount. I’ve used NDs to keep my iris wide as poss. 

[John Rostron]

Another member, @Friedster, has recently posted here with a problem in Focus Merging. I merged his images in Affinity, Zerene Stacker and in Helicon Focus and found that Helicon did not properly align the images. This is most unusual, I normally find HF very effective.

John

[GracieAllen]

Thanks John...

Yes, the source of THIS set is jpeg, because Helicon Remote on the tablet CREATES jpegs. I have other sets captured in RAW, both from a Nikond D810 and a Nikon D850, and they work fine in Helicon Focus and even Photoshop.  Admittedly Photoshop is crude, but it doesn't produce halos.

And yes, John, I know I can clone from one of the images, but that's not a practical thing to do in the real world.  Cloning a half inch of stem may work, but on any normal focus stack it's not going to be reasonable to clone all thing things Affinity is handling poorly...

In this case I rendered the stack with Method C in Helicon Focus.  And yes, I paid for it a few years ago.  I keep HOPING there will be a significant improvement in Affinity so I can STOP using Helicon, but from what I'm understanding here, there isn't...  And having to spend time trying to clone out the problem areas from Affinity, that don't occur in other products, isn't a practical answer.

Is there any practical way to "encourage" the people making Affinity Photo to improve their Focus Merge 'cause having to do extra work just to use Affinity isn't reasonable.

[John Rostron]

I have AP, Zerene Stacker and Helicon Focus (not HeliconCapture). For a Focus Merge, I start off with Affinity for a 'quick and dirty' run, then examine the result. If it satisfies me, I stick with it, otherwise I try Zerene or Helicon.  I find that AP suits about half the time. 

The Focus Stacking on AP could be improved, in particular it should offer a choice of algorithms. I find that, for a given stack, the different algorithms can result in different appearances of the backgrond and the prevalence of halos.

You can encourage Affinity to improve the process by posting a request in the Feature Request forum. If you do I will add my support.

John