Sony e pz 18-105 mm f4 so sánh 16-70 f4 năm 2024

This will be a nerdy post on MTF and other sharpness scores, so if you have some aversion to these then stop reading now.

One of the disadvantages of review sites is that they usually test only one copy of each lens, so you are never sure if the results are really indicative or not. Lensrentals is an exception, but their testing of APS-C lenses is pretty limited so far.

Three of the lens review sites have published test results for two of the most discussed zoom lenses for Sony E mount APS-C cameras, the Zeiss 16-70mm and the Sony 18-105mm. There are a bazillion threads on 'which is better'. I hope to add a bit of semi-objective evidence.

Again, if you think that sharpness scores are junk, then stop reading now.

Of course, 'which is better' is not solely determined by sharpness. For some, the extra 2mm at the wide end is crucial and that will decide in favour of the Zeiss, for others the extra reach means the Sony is more suitable. Then there is the question of the Zeiss 'pop', the power zoom of the Sony and lots of other stuff yada yada yada and where is the aps-c f/2.8 zoom Sony blah blah??? But this post is about sharpness of these two lenses.

Photozone.de, ephotozine.com and dxomark have all tested the two lenses on an NEX-7 and published results so we have a good comparison point. Unfortunately, I don't think that there are any other lenses which are tested on the same camera by all three!

I've 'normalised' the MTF scores from each site, simply by assuming that they all produce the same average score. This could be wrong (and probably is) but as long as it is not too wrong then the results should be ok. The point is to look at the relativities between the two lenses and use the three sites to find an average.

What follows is a lot of charts and a bit of text.

Here are the results for the wide end (16mm for the Zeiss and 18mm for the Sony), with the Zeiss first (center then corner) and the Sony next (center then corner) and there is a summary chart at the end. The numbers on the vertical axis don't mean anything at all - its just a way to compare the heights of the bars between the charts. You can see how the results are similar (and different) between the three review sites.

Now the mid range. For the Zeiss, photozone.de chose 40mm and the other two 35mm. For the Sony, ephotozine.com chose 45mm and the other two 35mm. Go figure.

Finally the telephoto (long) end. You can see from the second of these charts why photozone.de did not like the 16-70mm - the results are awful in the corners at 70mm. In their review they say, "In our disbelief we sent the lens to the local Sony service TWICE and they confirmed that it is within the factory specs."

But clearly the other two sites got a better copy.

And here is two charts showing the average scores for the lenses wide open and one stop down. Commentary follows.

The results were pretty consistent across the three sites. Dxomark tends to have a 'flatter' profile across the frame, but its not that much different.

What can we conclude here? The scores between the lenses are pretty similar but at the margin I think that the Sony is a little better for a 'standard' photographer, which I will try to explain.

First I should point out the obvious, that we would all love great IQ from wide open across the frame at all focal lengths. Given that is not possible, the manufacturers have to make choices on where to make the compromises - how the lens is 'tuned' - where are you prepared to accept lower IQ?

I am going to assume that you are going to generally shoot landscapes with the wide angle and portraits with the telephoto end. There are obviously exceptions to this.

With this assumption, at the wide end you are looking for decent sharpness across the frame - the corners are often just as important as the center - and are willing to stop down a bit to get it. That is, comparing sharpness for landscape lenses is a matter of asking which has the best corner sharpness regardless of aperture. Having said that, don't shout your objections too loud, its only a general rule-of-thumb.

The corner sharpness at the wide end is best in the Sony at f/5.6, making it the landscape winner. If you want to do low-light hand-held indoor wide-angle architecture (e.g. cathedrals) then neither of these lenses will help that much.

At the mid-range focal lengths you want the lens to be performing at its most even, with good center and corner performance, preferably from wide open. The Sony wins here.

For portraits, you are after great center sharpness at the long end wide open (for subject-background separation and bokeh) without worrying too much about the corners.

Again, the Sony is the winner here.

The Zeiss seems oddly 'tuned' to me. It is almost the reverse of what I just described, where center sharpness gets worse at longer focal lengths and corner sharpness is no better at the wide end than at the telephoto end.

Of course, if your preference is for having the best center sharpness at the wide end then the Zeiss would be your pick here.

But really, the IQ differences are very small between the lenses and I doubt you would notice them in most situations. I think that the Sony is the better buy, mainly because it is much cheaper and has a much longer focal range, with the only downsides being size and the famous distortion at the telephoto end (which can be corrected in camera or in post). For IQ, the two lenses are probably very close on average. Its just that the Sony seems to be better in the more commonly used situations.