Home / Opinions / HTC One M8 Camera vs A Proper Camera: Fake Bokeh On Trial

HTC One M8 Camera vs A Proper Camera: We Put Fake Bokeh On Trial

Andrew Williams

by

HTC One M8 Camera vs A Proper Camera: Fake Bokeh On Trial

HTC One M8 vs Fuji X-M1 35mm: Can HTC fake it to make it?

HTC talks big about the camera of its new HTC One M8 phone. It uses two rear lenses and two rear sensors in order to let the phone tell between objects that are close and far away.

The main use for this feature is to create a fake 'bokeh' effect. Bokeh is the nature of the blurriness in backgrounds seen in photos with a shallow depth of field. Normally phones can't successfully replicate this sort of effect – not compared with a 'proper' camera – as their sensors and lenses are just too small and basic.

To see whether the HTC One M8's fake bokeh can really compete with the real deal, we took the phone and the FujiFilm X-M1 compact system camera out for a spin. We equipped the Fuji with the Fujinon 35mm f1.4 lens, which is renowned for its all-round bokeh-licious-ness. Here's what we found.

Basic Bokeh 101

The way to get the most marked shallow depth of field effect is to shoot with a fairly close-by subject. When the in-focus plane is quite near, the background has plenty of room to get that creamy out-of-focus look that is what the HTC One M8 camera is all about.

Unlike a true optical effect, though, the HTC One M8 is a bit limited in what it can shoot. You can't use the uFocus bokeh mode if you're out of the standard mode. If you use HDR, you're out of luck. And if the phone reverts to night mode due to low-light conditions, it's also game over for uFocus.

To work within these limitations, we first went for a classic outdoors portrait shot – with the help of a furry friend.

Test 1: Easy Portrait

HTC vs Fuji 7

Original HTC One M8 shot

Without the Duo camera working its magic, the HTC One 8 can only produce a slight shallow depth of field effect. The flowers behind the blue monster are only a tad out of focus. But is it enough to give the subject an extra half-ounce of pop. The colour contrast helps too, though.

HTC vs Fuji 5

Processed HTC One M8

With the uFocus mode engaged, we see a distinct defocussing of the background. However, there are several weird inconsistencies. There's a flower to the left of the monster's head that has been judged by the camera's algorithm as being in the foreground (it wasn't), and there's a halo of 'crispiness' around the subject. This is where the M8 hasn't been able to judge the edges of the monster accurately.

FujiFilm X-M1 shot

HTC vs Fuji 6

The first Fuji shot immediately shows the benefits of natural, purely optical bokeh effects. Diffraction causes the background twigs to appear expanded, not only blurred. And there are no issues separating the furry edges of the monster from the background.

You may also like:

Test 2: Flora

For our second test we're going to try something a little trickier. It's a close-up of a flower with much less colour contrast than the first test, and with a less well-defined subject – a flower head is pretty small.

HTC One M8 camera sample

Original HTC One M8 shot

The fairly wide-angle of the HTC One M8 lens and (like most phones) limited close-up focusing abilities make it tricky to get good macro-style photos with the M8. The natural depth of field properties of the f/2.0 lens are not really enough to make the yellow flower stand out in this shot.

HTC One M8 camera shot

Processed HTC One M8

The HTC One M8's handling of this shot is quite bizarre. Note how the uFocus mode has de-focused the stem above the 'main' yellow flower even while it is in practical terms at the same depth as the flower and the stem below it. There is a marginal increase in the blurring effect of the far background, but it doesn't really compensate enough for the ways uFocus has essentially broken this shot.

FujiFilm X-M1 shot

FujiFilm X-M1 flower

As with the processed HTC shot, we see some of the flower stem going out of focus, but here it's consistent with the depth of field and the object's position in the field. Fake bokeh simply can't reliably create truly narrow depth of field effects without making the wrong decisions - even with a second lens, apparently.

Test 3: Subject Separation

Here we're going to see how well the HTC One M8 can separate subjects when they're not all that far apart, in terms of depth. Once again we're calling on a few furry friends. We've lined up four in a row and are going to try and focus on each consecutively to see how the HTC One M8 deals with the task.

It's here that we come up against another limitation of the camera, though. uFocus can only de-focus items, not re-focus them. So to help out, the HTC One M8's photos here are made using two 'source' shots - one focused on the nearest toy, the other on the furthest.

Original HTC One M8 shot

HTC original

HTC One M8 on Totoro (nearest)

HTC vs Fuji

This is perhaps the most successful HTC One M8 shot yet. Aside from some errant in-focus bits just to the right of the subject, the other toys are clearly blown out of focus. There is also some gradation in the effect -- the toys at the back are more strongly affected than the tiger cub next to the subject. This would definitely do for Instagram ('likes' not guaranteed, mind).

Fuji X-M1 on Totoro

HTC vs Fuji 4

Look closely between this photo and its HTC alternative and you'll see the subtle difference between being 'out of focus' and the simpler blurring effect used by the HTC One M8's uFocus mode.

HTC One M8 on tiger

HTC vs Fuji 14

This second separation test if less successful. The tiger has been brought into focus, but it hasn't been separated from the front Totoro toy. This seems odd when the two were successfully separated in the last photo. HTC's blurring algorithm is not very consistent. Also note the in-focus (ish) flowers just to the right of the subject.

FujiFilm X-M1 on tiger

HTC vs Fuji 3

Here the Fuji has done what the HTC could not – bring the tiger into focus while blurring out the nearby Totoro toy, our first subject. It's tests like this (and the flower test) that show that while hardware is involved, the HTC One M8 still relies hugely on software.

HTC One M8 on zebra

HTC vs Fuji 13

Selecting the zebra has brought it into focus and taken the front toy out of focus. But the Tiger is still pretty sharp, as is the back toy. Oh well.

HTC vs Fuji 2

FujiFilm X-M1 on zebra

Here the depth of field is enough to keep the rear blue monster largely sharp, but the Totoro toy and tiger are clearly blurred out, letting the zebra become the clear focal point of the shot.

HTC One M8 on blue monster

HTC rear

This photo is an interesting one. The front two toys have been successfully blurred-out, and there's a difference in blur between those toys and the flowers behind them. It's more evidence that uFocus doesn't apply a single-level blur to its photos. However, there are some weird aberrations again.

Part of the zebra is out of focus, but there are still in-focus parts on its arms. This may be the patterning of the stripes confusing the uFocus algorithm, Hopefully, HTC will be able fix these sorts of issues with a software update.

FujiFilm X-M1 on blue monster

HTC vs Fuji 1

As with the zebra shot, there's not a hugely marked difference in focus between the zebra and monster. But it is 'correct' optically.

Test 4: Messing with the Foreground

For our final test we're looking at how the HTC One M8 is at blurring out the foreground when it interferes with the background – which is in this case where the photo will be focused. This is probably the M8's trickiest challenge yet. Unfortunately, the actual scene has to be a little different between the two cameras because of their different focal lengths. Let's have a look.

HTC One M8 original shot

HTC vs Fuji 10

The HTC One M8 actually produces enough of a 'bokeh' effect to get the effect we're after here. The difference in field positions is so great that you can get pretty good effects without any processing. Granted, it's not the most interesting of scenes. Pretend it's a giant wasp instead of a Sky dish, or something.

HTC One M8 processed shot

HTC vs Fuji 9

On first glance, this shot might appear to be successful. However, it's deeply problematic and doesn't stand up to any closer scrutiny. There's really no separation between the leaves and the wall directly behind it, and there's actually little depth-related focusing nearby the dish. Also note the inexplicably in-focus row of bricks by the roof to the right of the photo.

This shot would actually be better off with the sort of zone focusing you get with one of Nokia's Lumia phones.

Fuji X-M1 shot

HTC vs Fuji 11

It's another easy success for the Fuji. No surprise there. However, this shot also exposes another bokeh feature the HTC One M8 is incapable of creating. The extreme blurring of the leaves has actually caused some of them to become translucent in this shot. To call bokeh effects blurring is really an oversimplification – it's all about how light is diffracted through a lens. And that's not easy to fake.

Verdict: Is the HTC One M8 uFocus feature any good?

These tests have proved, sadly, that in a lot of conditions the HTC One M8's uFocus Duo camera feature is not much use. It often spoils what would be a decent shot using the 'natural bokeh' of the phone's lens, and its algorithm needs work.

However, for simple portraits it can be a great extra, making people's faces pop more than they otherwise would. You need to manage your expectations and learn how to use it, but it is worthwhile. We just wish it was a bit better.

Next, read our guide to Android 4.4 tips, tricks and secrets

Philip Xenakis

March 26, 2014, 7:57 pm

Nice article! It's a shame HTC opted for a somewhat-gimmicky camera setup - I think having one good rear camera would have been better, in the end.

xmrawesomex

March 26, 2014, 9:46 pm

While these trials are nice, the depth sensor technology is still too new to expect superb results. I suspect that once the depth sensor is updated and the algorithms updated, the depth of field tricks will be a lot more useful. I took a picture in the store today that was actually very successful. The pictures may be a hit or miss so far, I see a lot of promise in this direction.

Indigo

March 26, 2014, 11:37 pm

Good to see members of the Trustedreviews team helping out by modelling in the photos. Seeing the personalities behind the reviews really brings them to life.

Stephen

March 27, 2014, 9:08 am

Is this really a fair comparison? You are comparing a phone camera to an €600 camera with a €500 lens if I checked prices correctly. The lens is a prime 35mm 1.4 aperature lens specifcaly designed and dedicated for this type of shot??
From the shots shown I think the HTC ONE M8 does a good job for a camera phone, Not sure what is the expectation but it is never going to match dedicated equipment.

andyvan

March 27, 2014, 9:41 am

The point is really to illustrate the difference between what the HTC does with its two cameras and what an actual camera does. Some might believe they're getting the same effect, but the reality is a little less extraordinary.

As Andrew points out in the conclusion, it's a nice extra but people need to check their expectations a little.

Andrew_TR

March 27, 2014, 1:19 pm

This made me actually laugh out loud.

Evan

March 27, 2014, 1:42 pm

What concerns me most is that @Andrew_TR is a 30 year old man with so many stuffed animals

marmo7ade

March 27, 2014, 1:50 pm

So now the camera isn't even good? It's certainly not the best smartphone camera that exists (Lumia 1020) but it's definitely "good".

mathesar

March 27, 2014, 3:45 pm

It's good but not really 2014 good, It's using the same 4.1MP "Ultrapixel" camera as the previous One (M7) and any phone camera comparisons you find online you'll see that once again the limited 4.1MP resolution is the biggest issue which puts it at the back of the pack in overall image details. I'm just surprised HTC didn't at least bump up the rear camera resolution.

Recent comparison: http://www.phonearena.com/revi...

Evan

March 27, 2014, 6:18 pm

We'll update this article if/when HTC make improvements to the depth sensor technology. You may well be right and algorithm updates become much better at finding the edges, for example, of the in focus item.

gesmith

March 27, 2014, 6:35 pm

Your test has a rather serious flaw, perspective distortion (google it). The lens on the Fuji has an effective focal length of 56mm due to crop factor (1.6x). The lens on the HTC One is equivalent to a 28mm. A 28mm on the same Fuji camera wouldn't produce as much image depth. You should have matched the focal lengths between the two cameras you were comparing. You wasted your time.

illregal

March 28, 2014, 1:59 am

Learn to use the one before trying to do a comparison.

Keith

March 28, 2014, 11:29 am

He is not trying to compare a camera to a phone, but comparing what real bokeh, and artificial bokeh that HTC is using, I thought that was obvious. So from my POV he certainly didn't waste his time, so thanks Andrew.

Keith

March 28, 2014, 11:33 am

Yes, because after learning to use HTC the algorithm used to create bokeh will magically get better.

illregal

March 28, 2014, 3:18 pm

None of the shots are even framed the same.. look at the flower pic.. Totally different perspective, kinda hard to do a comparison. Again on the building with the satellite dish. It's a totally different shot to begin with. I'm not saying the fuji isn't much more capable.. But they've intentionally gimped the comparison to make it look worse then it is. And don't you think within the next month there will be 1 to 2 updates to improve its algorithm.. If you don't you're dense.

Keith

March 28, 2014, 3:33 pm

What does it even matter if there not framed exactly the same?.
I think your missing the point. I currently have the HTC One, and I think it's a great phone, and I'm sure the M8 will be too, but I also think it's nice knowing the limitations of a phone I might potentially buy, been so defensive as you currently are is not informative or helpful. And yes, HTC hopefully will update the algorithm but this has nothing to do with "Learn to use the one" has it? (unless you're dense).

Willy

March 28, 2014, 4:15 pm

Why cant they just use a good camera? I have an M4, love it but the cam is a let down!

Nokia cameras are the best!

illregal

March 28, 2014, 4:40 pm

It's just sad when a reviewer on "trusted reviews" tries to paint something in bad light, when they just don't know how to use it. This goes for other sites as well.
It's like a noob hopping in a ferrari, and wrecking it. then trying to blame the car. Hint.. It wasn't the cars fault.
Like i said, it will improve with updates so it's really no big deal. But had they used it properly in the first place, maybe there wouldn't have been the issues in the first place. Its cool if you don't see their bias. but its pretty evident even in the heading of the article.
"we put fake bokeh on trial"
"Can HTC fake it to make it"

Stanley Yen

March 28, 2014, 8:00 pm

Totally agree with you. Reviews should eliminate personal bias as much as possible. Plus, why would you even compare phone camera to a professional grade camera?

Keith

March 31, 2014, 9:59 am

Lets get this straight!!, I have the original HTC One and like I said really like it. And so do TR, have you not seen the scoring, so the bias comment is certainly not justified. But liking a £600 gadget and not knowing it's weaknesses is really sticking your head in the sand. The fake bokeh has nothing to do with how the phone is used, or what focal length lens was used etc. Although it does a nice job is does have some weaknesses, and in fact if reviewers don't point these out, HTC in the long run won't have any incentive to update there firmware to improve it. Is that something you really want?. It's cool that like me you obviously like the HTC, but what would be even cooler is if people keep the pressure on making a great product even greater!!. That in my eyes would be the ideal world. To much fanboyism is not helpful.

Evan

April 1, 2014, 11:42 am

I think you're missing the point here. HTC are marketing this as a major feature that provides the bokeh effect you get with top cameras and lenses. "Give your photos an artistic touch with built-in Bokeh effect." Andrew has tested to see whether the claim is true.

Bokeh is not always a good effect - you can have good and bad bokeh even in proper cameras. Also this is a fake bokeh effect, it is not happening naturally through the lens. It is an emulation.

Also your comparison regarding a noob jumping into a Ferrari couldn't be more off. Andrew has reviewed hundreds of phones and cameras and has a deep knowledge of both subjects. The comparison is more like Lewis Hamilton spending a day on a track with a new Ferrari and offering an opinion on its cornering.

comments powered by Disqus