6 ms·
The sensor and lens of phones are good enough today to look exactly like the picture in the video (besides the shallow DoF that is only possible with large lens
by vortico 8y ago
The sensor and lens of phones are good enough today to look exactly like the picture in the video (besides the shallow DoF that is only possible with large lenses). The quality is primarily in the skill of lighting the room (in this case with $1000 of softboxes), and secondarily how the image is developed from RAW and retouched.
- mehrdadn 8y agoDo you get much mileage out of RAW photos coming from phones? I haven't had much luck with them.
- vortico 8y agoI don't have a phone capable of RAW images, but it's not a focus of the camera's design for those that have it, unlike professional cameras, so the bit depth will be low and it will expose the high ISO noise before "smart" denoising is applied when taking JPEG.
- jobigoud 8y agoThe shallow DoF can be approximated if you have depth information.
- vortico 8y agoYeah, but it looks crappy because you can't extract information behind the edges of foreground objects.
- utopcell 8y agoIf that were true, they wouldn't have faked the picture.
- vortico 8y agoProfessional cameras are easier to use, and the shallow DoF looks more professional (especially to inexperienced photographers).
- sethhochberg 8y agoI completely agree with you with regard to professional cameras and real DoF vs synthetic DoF that some of the nicer phones are doing these days.... but it sounds like you don't view using a DSLR here as misrepresenting the product in the advertisement? If we have to have "*enlarged to show texture" on cereal boxes so that people know they aren't buying a box full of 6"-wide Wheaties or say "professional driver on a closed course" when that new Audi on the TV does a 360 on the salt flat, surely there would be some expectation that there would be a disclaimer that the photo of the actor who was pretending to take a selfie with this hot new phone wasn't actually using the product being promoted?
- vortico 8y ago>it sounds like you don't view using a DSLR here as misrepresenting the product in the advertisement? Not really. I'm saying that it's already largely misrepresented through 99% other factors. The use of the DSLR is such a small factor that no one should really care whether it is used or the phone camera.
- sethhochberg 8y agoFair enough, yeah - proper lighting even with a smartphone on a tripod would go a long way
- michaelmrose 8y agoThis is terrible logic. You are acting as if the statement was the quality of the camera and if a similar enough picture could in theory have been taken it is almost true to the extent that if an identical image could have been taken it would magically transform into a truth. The actual statement was that particular image had been taken by that device which was 100% false. In fact it rarely makes sense to talk about degrees of truth people are either misleading you and nothing they have to say is worth anything or they are being truthful.
- vortico 8y ago
- pcurve 8y agoFor bright outdoor shots, I think it's possible. But there's no way you can get shot like 0:18 or 0:24 using camera phone...
- vortico 8y agoSure you can. For those images, start with properly placed lights. Then murder the images in your editor by increasing shadows beyond what is natural, and tune with the curve editor. What objective limitation does a phone camera have that would prevent you from doing that?
- coldtea 8y ago>What objective limitation does a phone camera have that would prevent you from doing that? Huge noise when you raise the shadows, even at ISO 100.
- vortico 8y agoAgreed that would be the largest factor, besides DoF. Although, I've been surprised by iPhone 8 or greater noise levels. In the article's video, you wouldn't notice the noise at that resolution if taken with a modern iPhone. Also, the shadows are mainly in the background, so you could retouch it easily with additional manual blurring. For Huawei, I have no idea, but I assume it's not much worse.
- rimliu 8y agoDoF has little to do with how big the lens is (unless that’s your way to say large aperture, but you can have 1.8 with small lenses too). But it has some to do with how large the sensor is.
- dboreham 8y agoPerhaps you should think about how f-number is defined? https://en.wikipedia.org/wiki/F-number https://en.wikipedia.org/wiki/F-number Then read up on depth of field : https://en.wikipedia.org/wiki/Depth_of_field https://en.wikipedia.org/wiki/Depth_of_field It should be clear that DOF has everything to do with the physical size of the lens because aperture for a given focal length defines the front element diameter. Your example very small f1.8 lens has to have a very small focal length and therefore deep DOF. Edit : found the math I was looking for : https://en.wikipedia.org/wiki/Depth_of_field#Use_of_absolute_aperture_diameter https://en.wikipedia.org/wiki/Depth_of_field#Use_of_absolute...
- rimliu 8y agoDid you miss the part where I introduced F-number into discussion instead of the lens size? What do you think will happen to DoF when I take my huge 70-200 and close it down to 22?
- vortico 8y agoIn that case the aperture diameter will decrease. f-number does not directly affect depth of field, except through the aperture diameter.
- rimliu 8y agoF-number describes the aperture diameter, it has no way to affect anything. The same as meters do not affect the distance. However, when you close the aperture on the large lens, the lens still stays large. That was exactly my point: it is about aperture, not the lens size. I get what you say about physically larger lenses having larger apertures in absolute dimensions, not the ratio, however that does not help smartphones, because of the distance of the focal plane and the sensor size. And this where the limiting factor is.