What's the secret behind extra large depth of field

[Curious Fox]

While most of the topics here are about image quality, I am more interested into large depth of field with lower image quality and lower magnifications. Is there any info how Keyence and Hirox with some of their lenses achieve large DOFs (without stacking), like this https://www.youtube.com/watch?v=MrEOhUIcnoM ?

I tested an old Keyence and got a large DOF but the image quality was terrible.

Is it possible to get similar DOF with low magnification microscope lenses, like 2x lens with raynox tube lens? I am not able to achieve it (I am familiar with aperture and diffraction).

Thanks.

[Macro_Cosmos]

What's the secret to our hobby? The DoF of a 20x Mit is less than 2um, but we can make the entire subject in focus.

Yeah, it's focus stacking. Keyence uses some fancy algorithms to generate the large DOF image you're seeing on the monitor. Without their bundled system and software, their lenses are frankly abysmal, riddled with overall softness and CA.Those lenses probably do have larger DoF, since they are just so soft.

It's all about computer algorithms. What they are essentially doing is capturing large amounts of information to generate the interactive 3D virtual objects.

If you want larger DoF, you'll require lowering the NA. Software wise, unless you are able to write your own software to perform this almost real time trickery, and you have a system that's able to create the numerous images for processing, you're out of luck.

[Curious Fox]

It's not focus stacking in the video I provided, I am interested in similar DOF and quality. Both Keyence and Hirox have high resolution lenses (one of them use Nikon objectives) and lenses with less resolution and large DOF. I am trying to achieve this DOF but I am not even close to it.

[Curious Fox]

This is sugar, no stacking.






Bug




These are good examples, maybe I am close to their DOF, I am not sure.

[rjlittlefield]

It's not focus stacking in the video I provided, I am interested in similar DOF and quality.

The options for DOF and resolution are limited.

With normal lenses, free of aberrations, there is a strict tradeoff between DOF and resolution. The only way to get around that is with classic focus stacking: take a bunch of pictures and rummage around to find the sharp bits.

There is an alternative technique, sometimes called "wavefront coding", that combines specially aberrated optics with subsequent digital processing. The trick is that the optics generate an image that looks pretty blurred by itselt, but it's equally blurred over a relatively large depth. Post-processing along the lines of deconvolution then removes the blur and restores the original sharpness at only the cost of some increased noise.

I don't have any specific knowledge of what goes on inside Keyence systems. From the above descriptions of its behavior I suspect that they're using something like wavefront coding.

That's presuming that in fact they're doing anything special at all. When I look at the video, I actually don't see anything that strikes me as an unusual combination of DOF and resolution. I think it's quite possible that what the video shows is just stopping down and taking the resolution hit from diffraction, but making that unnoticeable by presenting only moving images at video resolution.

--Rik

[Macro_Cosmos]

It's not focus stacking in the video I provided, I am interested in similar DOF and quality. Both Keyence and Hirox have high resolution lenses (one of them use Nikon objectives) and lenses with less resolution and large DOF. I am trying to achieve this DOF but I am not even close to it.

It is, the "stacking" is performed in their bundled software, and then showcased to you. It's purely algorithmic and it's a digital processing trick. It's not necessarily the same as the type of Z-stacking we use, but it works in a rather similar way. Some use a thing called wavefront coding which again is digital processing, utilising some kind of phase masking. I would guess there's systems that combine the two.

There's also a way to increase DoF by adding some sort of secondary lens element in front of the objective, I read that somewhere. It sacrifices image quality.

I don't know much about all this digital trickery to achieve greater DoF. Can't really explain well, I'm chewing on words here trying to sound coherent. Time to time, some full keyence setups do pop up on ebay and Japanese auctioning sites.

[dolmadis]

It's not focus stacking in the video I provided, I am interested in similar DOF and quality. Both Keyence and Hirox have high resolution lenses (one of them use Nikon objectives) and lenses with less resolution and large DOF. I am trying to achieve this DOF but I am not even close to it.

It is, the "stacking" is performed in their bundled software, and then showcased to you. It's purely algorithmic and it's a digital processing trick. It's not necessarily the same as the type of Z-stacking we use, but it works in a rather similar way. Some use a thing called wavefront coding which again is digital processing, utilising some kind of phase masking. I would guess there's systems that combine the two.

There's also a way to increase DoF by adding some sort of secondary lens element in front of the objective, I read that somewhere. It sacrifices image quality.

I don't know much about all this digital trickery to achieve greater DoF. Can't really explain well, I'm chewing on words here trying to sound coherent. Time to time, some full keyence setups do pop up on ebay and Japanese auctioning sites.

Is there any available software one can download/trial with "stacking" by wavefront coding?

[Chris S.]

I think Rik's last paragraph, above, deserves careful attention:

That's presuming that in fact they're doing anything special at all. When I look at the video, I actually don't see anything that strikes me as an unusual combination of DOF and resolution. I think it's quite possible that what the video shows is just stopping down and taking the resolution hit from diffraction, but making that unnoticeable by presenting only moving images at video resolution.

Video photography is much less revealing of diffraction than still photography. Look at an isolated frame from an otherwise outstanding video, and you'll see what I mean--it will likely look awful. Something about the merging of images into a moving picture seems to make the human eye/brain sense higher quality than exists in the individual stills.

And let us remember that a 1080p "High Definition" television only displays about two megabytes of information. What still photographer among us shoots anything like a two megabyte camera?

If one were shooting macro video even for high-end release, the simplest approach to creating enough depth of field to tell stories would be the old fashioned approach of stopping down, and leaving diffraction to become far less noticeable due to the nature of video.

--Chris S.

[rjlittlefield]

Is there any available software one can download/trial with "stacking" by wavefront coding?

Not to my knowledge. And remember that wavefront coding requires special optics to implement the unusual and extreme aberration. The software and special optics comprise a mated pair.

--Rik

[Smokedaddy]

You 'might' find this interesting to read, maybe not either.

http://confocal-manawatu.pbworks.com/w/ ... of%20Field

[Curious Fox]

It is, the "stacking" is performed in their bundled software, and then showcased to you. It's purely algorithmic and it's a digital processing trick. It's not necessarily the same as the type of Z-stacking we use, but it works in a rather similar way. Some use a thing called wavefront coding which again is digital processing, utilising some kind of phase masking. I would guess there's systems that combine the two.

They have "one click" stacking integrated in their software but it is taking images of static objects and it takes a few seconds to finish it. Let's not confuse it with video. If you are saying that they have stacking in real time in videos, keep in mind you have 25 frames in a second, you have a time barrier to take a lot of images for every frame of the video. Also can't imagine computer performance that can do it in real time.
I will learn more about wavefront coding.

That's presuming that in fact they're doing anything special at all. When I look at the video, I actually don't see anything that strikes me as an unusual combination of DOF and resolution. I think it's quite possible that what the video shows is just stopping down and taking the resolution hit from diffraction, but making that unnoticeable by presenting only moving images at video resolution.

Actually I see something unusual, to get this DOF you have to stop down that much that will result to almost completely lose the light on a sensor, their light intensity from some other videos is not that high. This is the only weird thing I see.

Video photography is much less revealing of diffraction than still photography. Look at an isolated frame from an otherwise outstanding video, and you'll see what I mean--it will likely look awful. Something about the merging of images into a moving picture seems to make the human eye/brain sense higher quality than exists in the individual stills.

This is true.