Hacker News Reader

Comment by bufferoverflow | Hacker News Reader

Yes, a tiny cell phone sensor with a tiny lens will never come close to a large sensor with a large lens, if they are both based on the same technology.

You can buy a Canon camera with ISO 4 million. Slap a f1.2 or f0.95 lens on it, and no phone in the world can compare with, what is essentially, night vision.

The optics will not compare as well. The fake bokeh is cute, but it's full of errors even in flagship devices. And it can't possibly fake the complicated situations with multiple layers, especially if some are semi-transparent.

View on HN

For me, the lens speed issue is the real killer that phones have never, and may never, come close to real cameras. Yeah, the bokeh can be simulated through software. However, for things like low light, anything far away, anything fast, (basically anything that needs information) you can't make up for it with software. <Insert CSI "enhance" joke>

I dabble in sports photography. There is no way I get the shots I get with a phone.

Bokeh is aperture and distance to subject, yes... but also sensor size. Phones have smaller sensors than even crop cameras.

Some phones actually have quite fast lenses. The iPhone 11 pro wide lens is f/1.8.

f/1.8 on an iphone and f/1.8 on a full frame camera might be the same number, but they are totally different animals.

The amount of bokeh you get between the two sensor sizes, even with the same f/1.8, exposure, and ISO, is going to be vastly different.

And the amount of light that enters the sensor through the lens is insanely different. A couple of years ago I moved from a Canon G16 with a f/1.8-2.8 6.1-30.5mm (28-140mm FF eq) lens to a Olympus OM-D EM-10 with its pancake f/3.5-5.6 14-42mm (28-84mm FF eq) lens.

Long story short, the pancake gets almost 3 times as much light as the canon integrated lens, despite the higher f number, because of the areas and sizes involved on both. And if I had gone for an APS-C for FF camera, we'd be talking in the order of 5 to 10x more light (hence, better low light images, or the ability to use shorter exposures for the same results).

Phones f/1.8 cameras are great -- comparing to other phones. If you compare them to a camera (almost any camera), they're in trouble.

Even different lenses with the same aperture on the same camera can have decently different t-stops.

Yes, because the f number is a ratio between the effective aperture vs the length of the lens. A long lens with larger f number can focus more light into the sensor than a shorter one with a smaller number, say f/4.0 400mm against a f/2.0 50mm.

The factor is not directly sensor size but rather focal length of the lens. But with smaller sensors comes a smaller field of view on the same focal length, such that you need to reduce focal length to keep it the same. I think for phone cameras the factor is ~10x compared to 35mm sensors.

How do you feel about the rise of mirrorless?

A bit skeptical. The biggest advantage for me is the burst rate and live histogram, but the newest Canon 90D, released last year, matches mirrorless FPS and also has a live histogram. The weight's advantage isn't as big of a deal as a lot of people have claimed because many of us have to deal with big lenses anyway.

The low/night light features are cool, I think, but my "mentor" who's a professional night photographer, thinks it's all a gimmick. It definitely lowers the bar for entry in this field for people who don't do a lot of night photography. The vast majority of my shots are under full sunlight or heavy stadium lights, so it's not that big of an advantage for me.

There's going to come a time where mirrorless will reach beyond what mirrors can do, but I don't think we're there yet.

The real advantage of mirrorless cameras is that there are less things that can go out of alignment, for high resolution photography everything counts (before you dismiss this, look up shimming kits for medium format backs).

That Canon has far more features than I'd expect for a DSLR - thanks for sharing.

I'm also impressed by the 90D's Handheld Night Shot mode, where it can stitch a burst of shots together like my Pixel phone.

I find that mirrorless offers another advantage from its somewhat decreased weight (a result of mirrorless camera's decreased flange distance [1], for the curious): it's relatively easy to mount most DSLR-designed lenses to any mirrorless with inexpensive adapters.

[1] https://en.wikipedia.org/wiki/Flange_focal_distance

> you can't make up for it with software

Actually that's why the Pixel gave so good results in the evening scene in the blog post. Because of the tons of AI that Google has put into improving low light conditions.

https://ai.googleblog.com/2018/11/night-sight-seeing-in-dark...

With the impressive performances of generative AI I'd raises the interesting question of what people actually want when they take a photo. Is it preferable if it looks great, even if the details aren't actually true and imagined by a neural network?

There was some discussion about it a few years ago when one of huaweis phones had a 'moon mode' that afaik basically recognized the moon and pasted in a static picture. Complete fake, but actually a reasonable idea - it's not like the moon will look any different and that's what the people want after all.

Reply on news.ycombinator.com