Huawei Caught Faking Smartphone Photo with a DSLR | Hacker News Reader

Huawei Caught Faking Smartphone Photo with a DSLR

petapixel.com

232 pointsnafizh96 comments

Open articleView on HN

Not the first time Huawei fakes something, they were cheating on benchmarks:

https://www.phonearena.com/news/Huawei-Ascend-P7-found-to-be...

https://www.anandtech.com/show/8403/examining-huaweis-benchm...

Samsung and others also did it:

https://www.anandtech.com/show/7384/state-of-cheating-in-and...

I own the Ascend P7, and my GF is also a Huawei user, and they're good devices. But they feel the need to cheat.

Looks like xiaomi is adding some pressure on the low end market, where people are a little more feature conscious.

The p20 lite was brilliant my mother owns one and for what it cost it’s a great little device, but the tech content pales compared to the last xiaomi round

I think the low end pressure is applying to xiaomi as well, from Oppo, Ding-Ding, etc...it's really hard to keep up with the cheap phone providers these days!

I thought they also cheated on the RF emission test.

I think that the over clocking cheat is a bit different if they are honest about and say the phone, when overclocked, are actually capable of this. The DSLR image is not possible in any case.

Unless the goal of cheating the benchmark was to target user's who overclock their phones (honestly, who does this?, like %0.0001?), then I would still call it dishonest.

Shouldn't benchmarks reflect what your delivering to consumers?

Honest related question. Do car manufacturers do this with their car's speed tests? I know you can squeeze another 50+ HP out of a VW Golf for example with a chip and a higher octane fuel, but AFAICT the specs you see on their site and at the dealer don't show this extra HP. (I realize VW is a bad example of a company that doesn't cheat their books!).

This reminded me of one of my favorite neologisms, used to describe faked screenshots that make video games look better than they are: bullshots

Photos aside, can you trust Huawei in the first place in general?

will after installing Lineage it's relatively safe, but you are too late to party after bootloader unlock page gone, now it's up to finding friendly customer service representative or giving try to DC unlocker which sometimes work even on phones they officially don't support

Would you trust Google more or less than Huawei?

Google 100% more. But the question was trust Huawei.

While I'd be absolutely certain they both spy on me (for various degrees of "spying"), I trust Google to better safeguard the data it collects about me.

Huawei is more likely to do a poor job of keeping said data to itself, either intentionally (selling it outright) or by getting hacked.

I'm convinced Google wouldn't sell my data - it is immensely more valuable to them if they're the only ones with it. That said, they'd still use it - to target and profile me, and of course I expect them both to share it with the respective governments.

Moreover, if I use Google services through a Huawei phone, they both end up with my data, while if I use a google device at least it's just Google. I'm not enthusiastic about either, but, hey, lesser of two evils.

(For the record I have an iPhone and try to avoid giving Google too much data about me)

Someone made the same comment to me the other day.

I said "Either it's the Americans or the Chinese spying on my, either way I'm fucked". I feel like it's safe to assume that most consumer hardware is backdoored/cracked by one intelligence agency or another.

Do we have proof that's the case with Apple though? Not a fanboy, but I casually remember reading that they weren't willing to sacrifice their security and capitulate to the gov on this point.

This shouldn't shock anyone given Chinese cultural attitudes towards honesty:

https://www.scmp.com/lifestyle/article/1879850/chinese-most-...

Why aren't the actors/models refusing to participate in such deceptive practices? Surely they'd be thinking "Why are we pretending to take a selfie?"

Cuz it's not like they're claiming to cure cancer. It's just an ad. Most people don't care.

I once saw a Microsoft Cloud poster in a train station that did claim they were curing cancer. I'd rather be the Huawei model than the designer of that poster.

Cause they are paid actors/models.

There job is to put on a show and do their part, not think about the grander things like deception.

Most claim they just show up for the part.

Because they are some of the lowest paid workers in the project. Why should they risk their precarious jobs when higher-paid, and more secure workers are also aware of the deception?

How many software developers work on adware?

People gotta eat somehow.

They're probably happy they can put food on the table. Actors for ads are not really at the top of the food pyramid.

While the Nigerian prince is? It's all just a scam.

What makes you think they even know the product that they are posing for?

>Surely they'd be thinking "Why are we pretending to take a selfie?"

Why? Their job is making you believe something fake. It's right there in the name of the job: "actor."

There's a million reasons. "Pretend to take a selfie so we can check the lighting." "Let's try a few different poses so the director has some options to choose from for the layout." "Let me take one with my other camera for a wardrobe reference."

Have you never made up a dummy report in HTML just so the designer could check the colors and layout? Do you refuse to work on any prototype that doesn't use live data?

This is an outrage! Next you're going to tell me that the pictures of happy creatives and handsome businesspersons in every service or product website are not actual users who are just so damn happy that they can't stop smiling, but actors/models or even (gasp) stock photos!

My favorite is when someone in an ad says "I'm not an actor"

Given that Apple set the standard by posting photos that are explicitly labeled as being shot on iPhone, it think it’s totally fine to expect an image with selfie-arm extending into the frame to have actually been shot on the device.

At least on billboards here in Berlin with a small print that the pictures are (heavily) post processed - 10cm size on a 10m high print.

Obviously you haven't read the article:

> Photos from Apple's "Shot on iPhone" ads are indeed taken with iPhones, albeit with additional equipment like special lenses attached to the phone, and they've been touched up with professional photo editing software.

Yep, just a standard iPhone here as any regular user would have: https://petapixel.com/assets/uploads/2017/06/shotoniphonefea...

Yeah, shot using an iPhone. Using the iPhone’s minuscule sized image sensor, through its tiny lens, and built in image processoring engine. With an attached lens.

Huawei used a large image sensor DSLR with dedicated processing chips and lenses designed for the body. The product they’re trying to pass off as having produced the image needed have been within 29 square miles.

> Everything was fine until people took a closer look at the behind-the-scenes photos (which have since been deleted) from the ad shoot posted by Elshamy to her Instagram account.

I doubt it's an accident. It's probably just a PR stunt by huawei, otherwise not many people would even talk about this phone. Unless whoever is runnigng that actress's instagram account is incredibly stupid to not notice that clearly controversial photo: [https://petapixel.com/assets/uploads/2018/08/bts0.jpg]

Or alternatively the controversy increases Elshamy's profile. Given that the account is either controlled by her, or her agent, I'd say that's the more likely explanation.

to be fair it should be mentioned Nokia did this in 2012 as well, it's not like Huawei doing it alone

https://www.theverge.com/2012/9/5/3294545/nokias-pureview-ad...

For their N97 promo video Nokia faked the whole UI of the phone: https://youtu.be/vJpEuMidcSU

Does anyone know of recent instances of such false advertisements in the US? Or are there severe enough punishments to deter them?

The last time I remember something like this happening was from 2012[1].

[1] https://www.cnet.com/news/nokia-forced-to-apologise-for-fake...

I might sound jaded, but: isn't this happening constantly anyway? Usually with just a small bit of "fig-leaf" credibility (like the aforementioned "shot on iPhone" pieces, but especially supertiny illegible disclaimers) - but the effect is the same.

I mean, watch this ad: https://www.youtube.com/watch?v=wDjGnWO_fG4

Can you spot the "screen image simulated" lines? Would you even have noticed it, had you not been looking for it?

fakes aside, my wife bought this phone (the pro) and I have to say it has the best camera (+post processing) I've seen on a phone. My wife sucks at taking photos but with that thing they look kinda nice. Their 3 lens composite photo system seems like it actually, really helps. And their added post processing of "portrait" mode gives a nice instagramish look, It's too much for my tastes (I don't do instagram soo....) but it's interesting it does some depth-of-field stuff too (probably with the help of the 3 lenses)

The only downside is I hate the $850 price tag (I refuse to pay more than $200 on a phone...) it'll be interesting when this sort of tech trickles down in the next few years.

I'm intrigued to know what "Instagramish portraits" look like!

smoothed, like they're wearing more makeup then they really are. like a cheap version of standard photo touch ups.

I mean even the resulting DSLR photo is touched up so heavily. How would anyone believe this is representative of the camera quality?

Because the phone camera does post-processing. That was the point of their ad pitch.

> How would anyone believe this is representative of the camera quality?

By being a typical consumer, without photography expertise. Also, they claim to use "AI" to post process the photos.

if you believe ANY advertising, you lose.

iirc, nokia also tried that in the past...

Correct

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.

Do you get much mileage out of RAW photos coming from phones? I haven't had much luck with them.

I 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.

The shallow DoF can be approximated if you have depth information.

Yeah, but it looks crappy because you can't extract information behind the edges of foreground objects.

If that were true, they wouldn't have faked the picture.

Professional cameras are easier to use, and the shallow DoF looks more professional (especially to inexperienced photographers).

I 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?

>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.

Fair enough, yeah - proper lighting even with a smartphone on a tripod would go a long way

This 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.

"Misrepresenting the product" means to fabricate a result that misleads potential customers to think that they will get a similar result when using the product. If (for the sake of argument) a DSLR and phone camera produced identical results, they would not be misrepresenting the product because customers will happily get the results they saw on TV. The implication or direct statement that the result is actually produced from the phone is a lie, but it's harmless to the customer. It is harmful if the result is vastly different than the normal circumstances of operating a phone camera because of lighting, photographer's skill, and hours of editing. What I am claiming is that DSLRs and phone cameras are in fact very identical, so this (in itself) is not misrepresenting the product.

I insist on the truth you are satisfied as long as its "truthy"

https://www.nytimes.com/2010/10/17/magazine/17FOB-onlanguage...

I would pay extra for a smartphone that let me do aperture/shutter control so that I could get the DoF I wanted. People say there's no point in compact cameras anymore and that's mostly true as far as pointing and shooting, but I can still take more interesting photos with my 5-year-old compact that gives me those controls than I can on my phone, even if my phone has more megapixels.

Chances are, a smartphone isn't going to give you the DoF you want (except by "faking it" using post-processing). There's a direct relationship between aperture for a given focal length and the front diameter of the lens. With a tiny lens and a low f-number, you're going to have a small focal length and therefore a wide DoF.

The controls don't matter - you can't get shallow DoF on a camera with a lens/sensor the size of the one in your phone. This is why the new iPhones mock it in software.

For 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...

Sure 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?

>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.

Agreed 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.

DoF 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.

Perhaps you should think about how f-number is defined? https://en.wikipedia.org/wiki/F-number

Then read up on 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...

Did 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?

In that case the aperture diameter will decrease. f-number does not directly affect depth of field, except through the aperture diameter.

F-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.

This discussion is about shallow DoF. You cannot have shallow DoF with a lens with a small aperture diameter. Both the aperture diameter and sensor must be large.

With everything else constant (f-number, sensor size, magnification, etc), increasing the aperture diameter will increase the circle of confusion and therefore decrease the depth of field.

So what? Camera companies, television companies, have been using simulated images in advertisements since forever. It's an ad. Of course it's fake. What world are you living in?

It's not about what world we're living in -- else we'd bend over and take any other kind of stuff that happens as well.

It's about what world we want to live in, and that (among more important things) doesn't include fake images passed as real.

Exactly. This is advertising people.

Since the dawn of time advertisers have over-exaggerated and over-sold products. Call it an art (I might, but not in this case), call it a sham. It is what it is. No surprises here.

Seriously. Who would even assume that an ad is not shot with professional equipment in the first place?

Reply on news.ycombinator.com