The 1020 to The 950

-




The cameras on the Lumia's are what made me switch from an iPhone, and has been a great pocket companion for the last couple years but as with any phone it started to show it's age. 

 The new 950 seemed like it was a reasonable replacement, but with fewer pixels and a smaller sensor I was curious if it would be too much of a compromise.

 So it seemed appropriate to go over the specs and see exactly how much of a change to expect.



At first it seems that the 1020 still has an advantage. It has a larger sensor, with a higher resolution. But generally the rule is for all things being equal total light gathered is most important.

 So how much light do both of these cameras gather?

 The 950 has a larger aperture ratio and a lower base ISO. Allowing about half a stops more light to each pixel, but with less the half the total pixels. So if we assume the same final resulting size the 1020 still has half a stop advantage.

 But the 950 should allow to shoot with 1 stop extra light using ISO 50 for only half a stop slower shutter speed getting you to the same total amount of light gathered. When shooting that way we could expect similar results.

TESTING SETUP

 To test it out, I took a 18% gray card in sunlight and measured at 15ev and in the shade at 9ev. The gray card filled approx the same FOV between both.

The 1020 took photos at ISO 100 from 1/8000 down to 1/125th. 

 This was repeated with the 950 at both ISO 50 and ISO 100

 The 950 was unreliable at choosing shutter speeds and so some speeds where missed (Appears to be a bug in the UI?)

 Each photo was then brought up to a ~50% value for comparison and scaled to the same output size.

 Since visual inspection of noise is hard, a 64x64 square of the same region for each of the reduced sized photos and calculated the Standard Deviation and Mean. The "noise" is represented as the Mean/STD.

 [if there are two SNR numbers the first is the gray card and the second is the dark spot]

 *note: This is a general test without proper controls for variation. The numbers that follow are to assist with comparing and are not definitive. Personally I would consider anything within 20% of each other effectively identical due to my own induced variation, but without reproduction it could be greater.
Baseline Sunlight

950 - ISO 50 - f1.9 - 1/5000s : 1020 - ISO 100 - f2.2 - 1/6400s
SNR 24.8 : SNR 20.3
Before Over Exposure

(This is the image before detail started to be lost due to over exposure and not a measure of equal exposure. Increasing the exposure would cause clipping)

950 - ISO 100 - f1.9 - 1/800s : 1020 - ISO 100 - f2.2 - 1/500s
SNR 31.4 : SNR - 13.2 
950 - ISO 50 - f1.9 - 1/320s : 1020 - ISO 100 - f2.2 - 1/500s

SNR 31.7 : SNR - 13.2 
Past Over Exposure

(This is the image pushed one stop (950 didn't take 1/400s so 1/320s was used) to clipping)

950 - ISO 100 - f1.9 - 1/320s : 1020 - ISO 100 - f2.2 - 1/250s
SNR - N/A
950 - ISO 50 - f1.9 - 1/250s : 1020 - ISO 100 - f2.2 - 1/250s
SNR - N/A
The two cameras are close at first but oddly the 1020 gets surprisingly noisy with over exposure. I would suspect that somehow the light changed bring out texture of the gray card while shooting. It only became noticeable after reducing exposure in light room.

 Below are stops until clipping for both of them as tested:

 950 -   18% gray to clipping (ISO 50   f1.9 1/5000s -> ISO 50   f1.9 1/320):  +4.0ev
1020 - 18% gray to clipping (ISO 100 f2.2 1/6400s -> ISO 100 f2.2 1/320):  +4.3ev


Underexposure Baseline

950 - ISO 50 - f1.9 - 1/80s : 1020 - ISO 100 - f2.2 - 1/100s
SNR 27.5/7.2 : SNR 23.4/8.0

950 - ISO 100 - f1.9 - 1/80s : 1020 - ISO100 -f2.2 - 1/100s
SNR 25.6/7.4 : SNR 23.4/8.0
Underexposure limit

For underexposure the 1020 shows a color pattern even 1 stop below, but it became much more pronounced at -3 stops which was past my personal acceptable limit. The 950 was then matched for the same under exposure.
950 - ISO 50 - f1.9 - 1/640s : 1020 - ISO 100 - f2.2 - 1/800s
SNR 13.5/3.6 : SNR 10.9/3.0
950 - ISO 100 - f1.9 - 1/640s : 1020 - ISO 100 - f2.2 - 1/800s
SNR 10.4/2.6 : SNR 10.9/3.0

 At both ISO 50 and ISO 100 the 950 looks better but I might consider the results interchangeable. They both are okay but not great. The 1020 simply looked awful. Which has generally been my experience when needing to bring up exposure in post.

The 950 seems to have a better looking noise then the 1020 which I would consider an advantage. A cell phone really shouldn't be expected to be pushed much anyways.

 This last one for fun was to take the last most underexposed of the set and push it 4.67 stops. It's not great. But it's also from a camera phone. The left is before, the right is after.
950 - ISO 50 - f1.9 - 1/2000s ( -4 2/3 ev)
SNR 5.1/1.8

Take Away


This was only a test using the base ISO where light wasn't really limited and the 950's ability to allow more light helped.

 I might try and do another look at higher ISO settings. But from what I can see the 950 is a viable replacement for the 1020 for how I used the phone.

 You don't get the benefits that go along with a 38mp photograph, but on such a small camera it was rarely a viable spec. 

 Much of the advantage was that the 1020 captured a lot of light and gave flexible raw files for a phone. The 950 continues that and adds more important features such as multi exposure RAW capture and faster shot to shot time.





Notes:

*All exposure estimations where cross checked using the calculator tools at http://www.scantips.com/

*Lumia 1020 photos I've taken can be found here: flickr album

*Lumia 1020 specs: https://en.wikipedia.org/wiki/Nokia_Lumia_1020

 *lumia 950 specs: https://en.wikipedia.org/wiki/Microsoft_Lumia_950





Samples

Since gray cards are literally one of the least interesting photographic subjects and don't really help compare a camera these are pairs of photos from of both the 1020 and 950. The order is arbitrary (Alt text will reveal the camera) and are hardly objective representations. But I think they do a good job proving that the 950 holds up to the 1020 for photographs in a variety of situations.

 I might try taking both phones out and shoot/edit some identically before finally retiring the 1020.


950

1020





950





950

1020





1020
950

Comments

Post a Comment

Popular posts from this blog

Nikon D800e Vs. Nokia Lumia 1020

-
Image
I will be the first to admit that the choice of title is a bit absurd. The D800 sits near the top of DSLRs with astounding dynamic range, interchangeable lens, and a sensor that's nearly 15 times the size of the 1020. So why even consider comparing such different machines? One word: Resolution The Lumia 1020 sports a 41 mega pixel "Pureview" sensor and optical image stabilization (OIS). In 4:3 this gives a working resolution of 38 mpix, and at 16:9 a working resolution of 34 mpix. This puts it 2 mpix on either side of the D800’s very impressive 36 mpix sensor. These two cameras which are worlds apart share nearly identical pixel counts.  While not all megapixels are created equal, we can’t help but wonder in a practical sense, how does a 36mpix image on a phone stack up when compared to the likes of a dedicated SLR of similar resolution? But exactly what would be fair wasn't easy. The Nikon is a very versatile machine producing images no smart phone cou...
Read more

Film Photography & being mindful of Scanning

-
Image
(Note: I am giving Medium a try as I find the layouts look nice. It's here ) Film often gets the short sell. Digital came onto the scene and it no longer made sense for the vast majority of photographers and snapshot shooters to keep buying rolls. But often I observe digital shooters taking cheap shots at the medium. They mistake soft scans for a soft medium. And while the nature of film inherently lacks the sterile clarity of digital, much of that softness comes from a different source. Which is why I wanted to attempt a visualization the impact scanning can have. But before that comparison we need to understand that the resolution of a camera or a scanner are not all created equal. The Resolution in MP(megapixels) or dpi(dots per inch) listed on the box often isn’t a literal translation to what is resolved (the fine detail we can see).  Shooting digital it’s very easy to take this distinction for granted but it still applies. Am image from a 24MP camera might onl...
Read more

CineStill 800T: General Impressions

-
Image
Generally there isn't too much exciting happening with films. Mostly the less popular ones go away, but the people at CineStill have done something fun. They've taken a few of the newest stocks from Kodak and did a good job repackaging it for normal color chemistry and use. The one that had the most interest to me was the  800Tungsten . It's pretty much the VISION3 500T stock use for cinema (You can see the 5219 markings on the negatives), but it's had the REMJET removed making it safe for C-41 processing. The question I had was what is it like to shoot, and is it worth the extra expense a roll. I had a hard time finding details of the sorts of quirks of this film, so I thought I would share my own experience. ISO 800 or ISO 400 First I would say I am not sure I agree with the speed rating. The very first roll I shot at 800 and felt generally the results were on the underexposed side. It doesn't make or break a shot, but generally shooting at ...
Read more