Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

About Bokeh

bokehtests.com

Post reply on HN
Print view View original

51–60 of 89 posts

Re: About Bokeh

#51

Post by throwanem »

scrumbledober wrote: post #33
jiberius wrote: post #29

Earlier quoted context omitted.

The delay after taking long exposures is caused by the camera taking a dark frame with the same exposure length as your actual image. When you take a long exposure, the sensor can exhibit noise caused by "hot pixels". The dark frame is taken with the shutter closed so that the camera can get an image of just the hot pixel noise, which it then subtracts from your actual long exposure. You can turn this off in the sett…

That's great information and makes me much less frustrated with my slow a6000.

You can also try turning off long exposure NR in the menus. I don't know if it makes enough of a difference on the a6000 to be worth leaving enabled and paying the extra cost in time, but that's a way to find out.

Re: About Bokeh

#52

Post by throwanem »

lame-robot-hoax wrote: post #19
penagwin wrote: post #13

Earlier quoted context omitted.

If we're talking about non-point and shoot cameras, such as DSLRs or Mirrorless Cameras that cost $500+ such as my Sony A6000, I absolutely DO NOT want post-processing done on my camera. Portrait mode on my iPhone 7+ looks okay at a glance, but thinks like hair, wires, and twigs are TERRIBLE on closer inspection. Not to say the software couldn't be improved. For example, after taking a 20 second exposure, I have to w…

I mean, you are using an iPhone 7, iirc that’s the first iPhone to implement portrait mode. The processing time on my iPhone 11 that I just got is minimal. I upgraded from an iPhone 6S so I’m not sure what the processing time was like on iPhones with portrait mode prior to the 11 models though. As well, who’s to say what Apple could do when given an image from a high end camera sensor instead of the tiny one in an iP…

> who’s to say what Apple could do when given an image from a high end camera sensor instead of the tiny one in an iPhone?

Ruin it. But they wouldn't, because there'd be no need.

The thing about a big high-end camera sensor is that having one means you don't need to fake bokeh. You can put a good fast prime lens in front of it and get the real thing. Phone cameras have to post-process their way to it because the physics of their small sensors mean they simply aren't capable of capturing enough off-axis light to produce true bokeh.

Re: About Bokeh

#53

Post by unicornporn »

thamer wrote: post #48
unicornporn wrote: post #36

Earlier quoted context omitted.

> Portrait mode on my iPhone 7+ looks okay at a glance, but thinks like hair, wires, and twigs are TERRIBLE on closer inspection. Portrait mode on any phone looks crap. I have not yet seen a picture where fake shallow DOF does not fail and look crap.

These improvements to bokeh on the Pixel 4 and Pixel 4 XL show some pretty good results: https://ai.googleblog.com/2019/12/improvements-to-portrait-m...

Impossible to say because sample[1] is highly compressed and 800 pixels high (including a white border).

[1] https://1.bp.blogspot.com/-3Sm2XHss0lA/XffZ95s_KjI/AAAAAAAAF...

Re: About Bokeh

#54

Post by BubRoss »

linuxdude314 wrote: post #40
jfengel wrote: post #20

Earlier quoted context omitted.

Does the RAW format save the kind of information that the software uses to guess at bokeh (depth sensors, multiple lenses, multiple shots, defocused pixels)? Or is there some other format that could save all that info, so you could really use the phone but do the processing later?

There’s no, but there isn’t really a need. The R package mentioned in a comment in this thread achieves a similar effect using no additional data. Ray tracing will suffice!

It shows a hand made depth map. You can't just 'raytrace' a 2D image and magically add depth of field.

Re: About Bokeh

#55

Post by anamexis »

tpolzer wrote: post #38
jiberius wrote: post #29

Earlier quoted context omitted.

The delay after taking long exposures is caused by the camera taking a dark frame with the same exposure length as your actual image. When you take a long exposure, the sensor can exhibit noise caused by "hot pixels". The dark frame is taken with the shutter closed so that the camera can get an image of just the hot pixel noise, which it then subtracts from your actual long exposure. You can turn this off in the sett…

On the one hand that's correct. On the other hand you could completely avoid that if you did long exposures the way a Google Pixel (2+ I think) does it (in software): Take lots of short exposures and fuse them. If the device is handheld you get variability in positioning for free, if it's on a tripod, it will automatically wiggle the OIS slightly to achieve the same effect.

It seems like movement would be effective for averaging out random noise, but the described process isn’t for eliminating random noise, it’s for eliminating persistent hotter pixels on the sensor.

Re: About Bokeh

#56

Post by gdubs »

pyridines wrote: post #12

As a very nearsighted person, I can see bokeh if I take off my glasses and look at points of light - either nearby small lights, or far-away large lights. Also, when I look at a bright sky through a tree's canopy, the canopy transforms into a breathtaking mosaic of bokeh, which becomes animated if the tree is swaying in the wind. Can people with normal vision experience this?

For me, points of light look like the canonical star shapes you see in telescope photos — which I don’t really understand because from what I’ve read, that effect is due to the braces which hold the telescope’s lens in place.

Re: About Bokeh

#57

Post by behringer »

droitbutch wrote: post #14
behringer wrote: post #6

Earlier quoted context omitted.

As a photographer I really don't ever want to use the camera software. Even when it's perfect for my application, I want RAW photographs because you have to fix so many other things before you ever get to special effects. You need to crop, color correct, distortion correct, noise correct, exposure correct. Only after that would you ever really want to apply some kind of bokeh filter. The entire reason to use a high e…

> You need to crop, color correct, distortion correct, noise correct, exposure correct. Only after that would you ever really want to apply some kind of bokeh filter. Modern phones already do all that - and seemingly in that order. As a photographer, what I find lacking are the lenses. You can only squeeze so much light from the relatively tiny lenses. Perfect recent example was kids Christmas play. Sitting back 25+…

I've never seen any phone get even half as good results as any DLSR released this decade. Just because the phone software can do it doesn't mean it's any good. I mean my phone has photoshop on it. That's pretty good, but nothing beats a proper suite of tools on a color calibrated monitor.

Re: About Bokeh

#58

Post by s_gourichon »

pyridines wrote: post #12

As a very nearsighted person, I can see bokeh if I take off my glasses and look at points of light - either nearby small lights, or far-away large lights. Also, when I look at a bright sky through a tree's canopy, the canopy transforms into a breathtaking mosaic of bokeh, which becomes animated if the tree is swaying in the wind. Can people with normal vision experience this?

Anyone can see what it looks like to be (strongly) nearsighted by putting a magnifying lens in front of their eyes.

Anyone nearsighted can estimate the strength of their nearsightedness by looking at an object close enough to be perfectly clear, then increasing the distance until the farthest where it appears clear. Take that distance in meters and invert it. The result is the amount of nearsightedness in diopters. For example : you see clear up to 50 cm away, or 0.5m, you have 2 diopters. You see clear up to 10 cm away, 0.1m you have 10 diopters.

Anyone with normal (or corrected) vision can estimate the strength of a magnifying lens by the same method. Most magnifying glasses are 5 to 20 diopters.

Well, perhaps not strictly anyone can, but you get the idea.

Re: About Bokeh

#59

Post by zakki »

krzyk wrote: post #37
lame-robot-hoax wrote: post #19

Earlier quoted context omitted.

I mean, you are using an iPhone 7, iirc that’s the first iPhone to implement portrait mode. The processing time on my iPhone 11 that I just got is minimal. I upgraded from an iPhone 6S so I’m not sure what the processing time was like on iPhones with portrait mode prior to the 11 models though. As well, who’s to say what Apple could do when given an image from a high end camera sensor instead of the tiny one in an iP…

My wife has iPhone 8, I can't find portrait mode on it (I'm coming from Pixel 3 so I was wondering how it looks on iPhones). When I scroll the photo modes there is no such thing as "portrait". Or maybe it is only for the Plus models with two cameras?

Yeah, it is on plus model including animoji

Re: About Bokeh

#60

Post by virtualritz »

I am surprised about how biased this article is in assigning attributes like ‘good’, ‘bad’ or ‘perfect’ to Bokeh shapes. You can talk about this in these terms solely from an optical engineering pov.

The shape of Bokeh is often why a lens get picked for a shot.

What the author calls ‘bad Bokeh’ other people call ‘soap bubble Bokeh’ (google for it) and they are willing to cash out [2] for owning a lens that can produce this sort of thing. This is an artistic choice[1].

This article’s generic statement that this sort of Bokeh creates images where the out of focus BG areas distract from the FG/in focus motive shows a narrow understanding of how Bokeh can be used in composing an image.

[1] https://www.shutterbug.com/content/creating-unique-macro-ima...

[2] Caveat: I bought a Meyer Optik Trioplan 100mm f/2.8 replica on Kickstarter for this very reason: the ‘bad’ soap bubble Bokeh. I love this lens.

Post reply on HN