Previous news story    Next news story

DxOMark report: Nokia Lumia 925 scores nearly as high as Lumia 1020

By dpreview staff on Nov 7, 2013 at 16:30 GMT

Nokia's Lumia 925 is PureView branded but doesn't use the Lumia 1020's large sensor and pixel-binning algorithms. However, the 8.7MP smartphone just scored nearly as well as Nokia's 41MP flagship phone in DxOMark's extensive imaging tests despite the conventional 8.7 MP CMOS sensor. Click the links below to learn what our partners at DxOMark found. 

Comments

Total comments: 22
keeponkeepingon
By keeponkeepingon (5 months ago)

"the 8.7MP smartphone just scored nearly as well as Nokia's 41MP flagship phone"

DXOMark resizes everything to 8mp before testing so yeah, surprise surprise when everything is shrunk to the same size it all looks the same.

Continually disappointed that dpreview continues to publicize DXOMarks whacked mobile tests, especially without a disclaimer as to how they jimmy the images before testing to iphone size.

0 upvotes
peevee1
By peevee1 (5 months ago)

"the 8.7MP smartphone just scored nearly as well as Nokia's 41MP flagship phone in DxOMark's extensive imaging tests"

Probably says more about DxOMark's test than about the cameras. After all, the sensor in 1020 is about 4 times the size, not only in resolution.

0 upvotes
dpfan32
By dpfan32 (5 months ago)

Contrast of many 925 is very poor.
Especially bright objects are smeared and sharpness decreases a lot. Setting exposure compensation down does not help much, darker areas then disappear in a blurry mess as well.

You have to do heavy post processing: lighten shadows and massively lower the lights. Increasing sharpness, contrast and saturation is makes the picture look finally right.

0 upvotes
wansai
By wansai (5 months ago)

I really like the 925 but it'd be really cool if it could get raw support. There's potential to do much better than the in camera processing.

I am already happy a phone actually gives me night shooting capability but it's maddening how they process it.

0 upvotes
Michael Ma
By Michael Ma (5 months ago)

This can't be right. Tim Cook said the iPhone 5s camera was good as a DSLR. This can only mean the Galaxy S4 is better than a DSLR.

7 upvotes
fuego6
By fuego6 (5 months ago)

sigh.. windows phone... oh Nokia... what could have been!

2 upvotes
Infms
By Infms (5 months ago)

Try it before you knock it... incredibly stable and slick OS.

3 upvotes
Pavel Sokolov
By Pavel Sokolov (5 months ago)

with WP since htc mozart, lumia 800 and now 925 - really good os. bauhaus know the right things in design (metro is bauhaus style).

1 upvote
yabokkie
By yabokkie (5 months ago)

pixel count doesn't necessarily affect image quality
(well it should be well known since D3 and D3X)

1 upvote
OneMoreComment
By OneMoreComment (5 months ago)

Totally agree, it never did, but it works perfectly as an advertisment nothing more.

1 upvote
dpfan32
By dpfan32 (5 months ago)

Hmmm but real world performance is lightyears worse than 1020 ... strange ...

Or maybe the had a much much better unit than dpreview here:

http://connect.dpreview.com/post/6074844014/nokia-lumia-smartphone-camera-shootout-lumia920-lumia925-lumia928?page=2

LOL

Comment edited 2 minutes after posting
0 upvotes
bigley Ling
By bigley Ling (5 months ago)

yup that is posible, maybe DXO and Connect both ned multiple test units to make fair comparisons. There seems to be such variations in quality between handsets it is not funny!

0 upvotes
oduis
By oduis (5 months ago)

As a 925 owner, I must say that the DXO Mark Mobile score does not reflect reality at all. The problem is probably that the DXO Mark take way too few camera attributes into account.

The 1020 has a much bigger sensor, so it allows a narrower depth of field. This makes a big difference in composing images (and that’s why people buy full and mid format cameras). DXO does not measure that.

The 1020 uses its much higher resolution to have digital 3x zoom, versus no zoom on my 925. Not taken into account at all by DXO Mark Mobile.

The 1020 will get DNG raw support, which opens a whole world of highlight recovery, noise reduction fine tuning, precision in shadows etc. in Lightroom. The DXO Mark Mobile does not measure that.

So to sum up, DXO Mark Mobile looks scientific but is pretty useless to determine real world camera capabilities.

Comment edited 11 minutes after posting
3 upvotes
bigley Ling
By bigley Ling (5 months ago)

I think DXO analysis is designed to cater for the "full auto" user.

The full auto user will not be capturing DNG and post processing using a PC/MAC later.

Also a full auto user, may want more of an image sharp rather than having narrow DOF.

Lastly the full auto use would probably use the smartphone camera for instagram shots, where higher resolutions are not important.

They should make two DXO results, one for "full auto" users and another for "enthusiast smart phone photographers" I guess connect does look into features like zoom, quality of MP, sensor size, software features more in detail than DXO

0 upvotes
joe6pack
By joe6pack (5 months ago)

With a short focal length, you can't do much background blur except for subjects within 1 ft or so. I don't see much of an advantage here.

I would even argue that most people using smartphone camera want deep DOF with everything in focus. So what you claim advantage may actually be a disadvantage.

Comment edited 3 minutes after posting
0 upvotes
oduis
By oduis (5 months ago)

I'd say the opposite: Many smartphone snapshots are friends and family, and especially portraits benefit from lower DOF. I compared the 925 and 1020 in a shop, and the DOF difference is clearly visible.
I agree with Bigley that DXOMark should be more comprehensive and maybe differentiate between serious and simple users.

But DXO Mark Mobile is even not comprehensive enough to judge just simple full-auto users. For instance they do not take optical image stabilizations for still shots into account at all. A big practical plus for the Lumias in general, but not measured by DXO Mark.

And there are more holes in the DXO Mark Mobile like this.

0 upvotes
vv50
By vv50 (5 months ago)

the DXO Mobile protocol explains what their "reality" is - http://www.dxomark.com/Mobiles/How-we-test-smartphones-The-DxOMark-Mobile-protocol
Every website's objective criteria for measurement is subjective to everyone's interests. Only if you care about exactly the same things would you agree with the points they assign to devices.

1 upvote
oduis
By oduis (5 months ago)

Thanks for the Link, vv50.
The measurement description opens up even more big holes.

For instance they measure low light and bright light according to Lux values (so a fixed brightness). But the ISO value that the camera takes is largely depend on what the camera automatic decides for that scene and brightness of the lens. So the ISO values vary greatly, and ISO influences image quality a lot.
So if a camera automatic decides on lower ISO, resulting in higher motion blur on moving objects, it will get much better DXO Marks, but real world results will be unsharp.
That is why the DXO Mark for DLSRs always measures all curves by ISO, not by lux brightness.

So the DXO Mark Mobile test results are not comparable between cameras at all. But the fancy charts try to fool you into thinking so.

I could dig up many more holes ;-)

0 upvotes
bigley Ling
By bigley Ling (5 months ago)

It does bring back the point that DXO is really testing "full auto" on smart phones as dependant on how the automatic setting is programmed in relation to what is an acceptable minimum shutter speed in a low light test.

I believe DXO will have a hard time taking into account ISO as well since on some handsets the ISO cannot be set manually like the iPhone, hence possibly the simple set lux rates for measuring performance of camera optics.

Comment edited 6 minutes after posting
0 upvotes
oduis
By oduis (5 months ago)

If your measurements are seriously wrong by structure, then you cannot say “but I don’t know a correct way to measure it, so I present you the wrong values anyway”. Especially when you try to give yourself a scientific look.

And as outlined above, DXO Mark Mobile is full of flaws, even by reduced point-and-shoot standards. Could give you some more if you need ;-)

0 upvotes
bigley Ling
By bigley Ling (5 months ago)

there must have been a severe penalty taken by the 1020 to result in DXO score that is that close to the 925! We know the textural detail and xenon flash is superior on the 1020 vs the 925, so somehow the 1020 must have had some major issue to bring it's score that close to the 925.

0 upvotes
MistyFog
By MistyFog (5 months ago)

That is assuming the test or measurement is correctly done. :)

0 upvotes
Total comments: 22