I’ve been using the Nikon 32mm F1.2 prime lens for many months now and I have owned and used the Olympus 45mm F1.8 lens for much, much longer. I think it’s safe to say that I know both lenses inside and out. Looking at their properties – like focal length, suitability for beautiful bokeh, and so on – one inevitably arrives at the conclusion that these two lenses serve the same purpose inside their respective ecosystems and that therefore they are very much comparable. In my opinion both of them are great lenses in their own right. There are huge differences in the image quality department, however, which I’m going to discuss in the following article.
Considerations and Processing
Comparing sharpness and other properties of lenses across camera systems with different sensor sizes, resolutions and aspect ratios can never be done in a manner that is going to satisfy everyone. People not knowledgeable enough are going to assume that different numbers – like 45mm vs 32mm or F1.8 vs F1.2 – equate to comparing apples to oranges, when in fact we are comparing the same thing here. When you look through the electronic viewfinders of the Olympus OM-D E-M10 and Nikon 1 V1 – which are the two cameras I have used these lenses on – you are hard pressed to find meaningful differences in field of view and background and foreground blur respectively. For all intents and purposes these two lenses can be used to photograph the same thing, from the same spot, with very similar results. But, as is often the case, the devil is in the details. There is no way around the 6 megapixel advantage of the OM-D E-M10 and the fact that it has a narrower sensor, which enables it to utilize the higher resolution and less prone to imaging artifacts (such as vignetting) part of the image circle the Olympus 45mm F1.8 is projecting out of its rear end.
So how should we compare the sharpness of these two lenses? I think we can all agree that 100% crops are the way to go, but should we a) compare 16MP vs 10MP with crops from the OM-D E-M10 showing more magnified but less acute detail or should we b) resize the OM-D E-M10’s images to 10MP and compare like for like?
After spending hours in Lightroom looking at the images at 100% I’ve arrived at the conclusion that there is no point in doing a). The size difference is so huge, that it’s virtually impossible to compare the sharpness of the two lenses. The alternative b), on the other hand, has problems of its own: how do you resize and sharpen the images shot with the OM-D E-M10 and 45mm F1.8, without destroying or artificially boosting detail? Because of this conundrum I’ve decided to upload the following images to flickr, so you can take a look and judge for yourself:
- Full resolution 16MP images shot with the Olympus 45mm F1.8 on the OM-D E-M10
- 10MP versions of photos shot with the Olympus gear
- Full resolution 10MP images shot with the Nikon 32mm F1.2 on the Nikon 1 V1
You can find all these images, including crops embedded throughout the article, in my Flickr set. Don’t forget to click on the crops to view them in full resolution on Flickr!
Note also that I did all the processing in Lightroom 5.7. No corrections were made in regard to sharpness, vignetting, CA and purple fringing. Sharpening and noise reduction settings were left as is, which means 25 for sharpening and 25 for color noise reduction. I didn’t apply any additional sharpening during export, which means that the OM-D’s/45mm’s images, which were resized to 10MP, were NOT sharpened for the second time.
When you think of portraiture lenses, sharpness is not the first thing on the list of desired attributes that comes to mind. Sharpness can even be detrimental in portrait photography, since it makes skin imperfections more obvious. That being said, it is always easier to apply a blur filter and get rid of excess sharpness than to transform a blurry, muddy mess into a respectable photograph. Fortunately there is very little difference in the sharpness department between the Olympus 45mm F1.8 and Nikon 32mm F1.2. The 32mm F1.2 is a tiny little bit sharper in the middle across all aperture values – an amount that can easily be rectified with added sharpening. I believe this is either due to the resizing of the photos shot with the Olympus gear or due to the Nikon 1 V1 having a weaker antialiasing filter in front of its sensor.
Now let’s take a look at the edge sharpness. Here it is impossible to compare virtually identical crops, because of the different aspect ratios of the sensors (4:3 vs 3:2). So keep that in mind.
Upper left corner
Lower right corner
The 45mm F1.8 is sharper at the edges of the frame, but only by a very small amount. In my opinion neither of the two advantages is meaningful enough to warrant choosing one lens over the other. Especially not when you consider that they are almost indistinguishable once you stop down a couple of aperture stops, in which case both can be used as special purpose lenses for landscape photography.
Purple fringing is one of those imaging artifacts that have become less and less of a problem as image processing software and computer hardware became more and more powerful and affordable. Nowadays purple fringing is no longer a deal-breaker as it used to be, since it can be remedied with just a few mouse clicks in many RAW processing applications – of course, as long as it’s not a case of extreme purple fringing. That being said, it’s always better not have to deal with it in the first place, since it saves a lot of time – time which can be invested in other areas of image processing or in shooting more pictures.
The Olympus 45mm F1.8 has a sizable lead in this department. It displays a low tendency to purple fringing wide open and you really don’t have to stop down the aperture too far to get rid of PF entirely. In fact stopping down by 1/3 of an aperture stop to F2 eliminates PF completely. With Nikon 32mm F1.2, however, purple fringing is more of an issue. Wide open PF is strong enough to warrant the use of the CA removal tool in Lightroom. And you have to stop down more than two aperture stops to F2.8, to get rid of it the old-fashioned way – as in without software correction.
Like purple fringing vignetting can be taken care of with software correction, but unlike PF there are specific kinds of vignetting, which are very hard or very labor intensive to correct. I’m talking about the kinds of vignetting, where corners are not only darker, but also have a very different color cast than the middle of the image. And unfortunately the Nikon 32mm F1.2 has exactly that kind of vignetting. As you can see in the side-by-side images down below, the 32mm has a strong green cast in the corners at aperture values between F1.2 and F2. Furthermore it has a much stronger tendency to vignetting than the Olympus 45mm F1.8. You have to stop down all the way to F4 to get rid of vignetting entirely with the 32mm.
And while you can use the embedded profile in Lightroom to brighten up the corners, when processing 32mm F1.2’s shots, the color cast remains. And this, in my mind, is the worse of the two problems. The only way to fix this, would be to shoot a blank surface with all aperture values and to use these images to take care of the problem in appropriate image editing apps.
Regarding the Olympus 45mm F1.8, it appears that correction data is embedded into metadata of each image and that most image editing apps use this data to automatically fix vignetting. Purists might find this unsettling, but I have to ask: is this really a problem, if the results look as good as they do? Would you prefer a “hands off” approach like Nikon’s with the 32mm F1.2? I think, in the end, every photographer has to answer this question for himself or herself. For me, what Olympus is doing makes a whole lot of sense. It has no drawbacks that I could find and it saves me a lot of time, which I would otherwise spend fixing the color cast.
Bokeh is one of the most important attributes for portraiture lenses and it’s also one that is the most subjective. No other image quality attribute of a lens is subject to taste to such a degree, as is bokeh. And its quality is not only the result of focal length and aperture, as you can see in the images down below. On paper, Olympus 45mm F1.8 and Nikon 32mm F1.2 are equally matched. The 45mm F1.8 is equivalent to a 90mm F3.6 full frame lens, while the 32mm F1.2 behaves more like an 86mm F3.24 full frame lens. The Olympus lens is therefore a tiny bit longer, which is good for bokeh, but it’s also a tiny bit slower, which in turn is bad for bokeh. All in all, the differences in focal lengths and apertures are negligible. It seems to be the specific constructions of the lenses, which make all the difference. And the most prominent one is that the 45mm F1.8 has perfectly round bokeh balls even towards the edges, while the 32mm F1.2 has a strong tendency to so called “cat eyes”, or in layman’s terms almond-shaped bokeh balls. Some people like round bokeh balls – the more “neutral”, “scientific” interpretation – while others prefer the more “artistic” interpretation of cat eyes. I belong to the former camp.
One more difference between these two lenses lies in how stopping down affects bokeh. Olympus 45mm F1.8 has an advantage here as well. It manages to keep its bokeh balls round for longer. Unlike 32mm, which already has edgy bokeh balls at F2, the 45mm’s bokeh balls remain almost perfectly circular at F2.8 – in both cases this equates to stopping down the aperture by 11/3 stops.
In the following images, which were taken from the same spot, Olympus 45mm F1.8 seems to have a very slight edge in terms of quantity of the bokeh wide open. Whether this is due to the additional focal length having a larger influence than the smaller aperture, I cannot say. But the difference is there, however insignificant it may be.
Sharpness is, somewhat unexpectedly, not one of the image quality aspects, in which the Olympus 45mm F1.8 and Nikon 32mm F1.2 differ the most. In fact, in this regard, they seem to be quite equally matched. 45mm is a bit better at the edges, while the 32mm is a tad better in the middle – in both cases across the entire aperture range. Even so, the differences are minor. Differences in purple fringing are a bit more pronounced, but since this particular issue is the easiest one to fix in post processing, I wouldn’t make myself crazy about it. In the vignetting and bokeh department things are more serious, though. The Olympus is simply a lot less prone to vignetting. Whether this is due to a better lens construction or software correction, it doesn’t matter – at least to me it doesn’t. What’s important is the end result and it’s vastly superior with the Olympus’ lens. Regarding bokeh, there is more room for interpretation. But simply put, if you prefer more scientific/neutral interpretation of bokeh, 45mm F1.8 is the clear winner. If you prefer a more “artistic” interpretation, 32mm F1.2 is probably the better choice for you. But, when talking about choosing, there is none without considering the price first. And the price is the 32mm F1.2’s Achilles’ heel. In my opinion it is simply way too high, considering the fact that the 32mm is worse than the 45mm in most image quality aspects.
This is how you can support BigStreetGuns
If you like my work and intend to buy anything at amazon, please consider using my affiliate links: AmazonUS, AmazonUK, AmazonDE
You won’t pay a penny more compared to going to amazon directly, but I will receive a small commission. You will be supporting BigStreetGuns at no cost to you. Thanks for the support! 🙂
If you are not a regular amazon customer but would still like to support my work, you can do so by making a modest donation through PayPal. It would be most appreciated.
Follow me on:
[…] ← Previous […]