75/1.8 samples & full specs on four-thirds.org

Started May 24, 2012 | Discussions thread
Anders W
Forum ProPosts: 15,516Gear list
Re: Thanks for the link.
In reply to fermy, May 26, 2012

fermy wrote:

Anders W wrote:

The assumption required for the comparison is simply that the resolution of the sensors compared is effectively the same (same MP count, same AA filter strength). Under that assumption, the comparison is valid for the sensor resolution in question. For a pair of sensors with lower resolution, the observed performance difference would be smaller. For a pair of sensors with higher resolution, it would be bigger.

That is not what you did several times when you tried to compare 85/1.8D tested on 10mp DX to the 45/1.8 Zuiko tested on 12MP m4/3. So now you want to restrict the comparisons to the sensors with the same mp count? That's definitely a step forward.

No, I don't want to restrict comparisons but, as already indicated several times, make sure that they are as correct as the data allow.

What is clear is that 5D+35/1.4L combo resolves more than E-PL1+20mm pany combo as in order to compare system performance you can safely multiply lpmm by the sensor size.

Hah. So that's what you think? In fact, the opposite is clear. As I have alredy pointed out, the figures in the little tables I gave you have already been multiplied by the square root of the ratio of the sensor areas, i.e., in this case 2. The raw lp/mm values reported by Lenstip are only half as high for the Canon as in the table I gave you.

I haven't use any of your figures for comparison at all. I went to lenstip english site and took figures from their reviews directly.

OK. It would have been better if you had used my figures and thereby ended up with the correct outcome rather than try to do things yourself and mess it up, e.g., by using the graph for the 20D rather than the 5D and still multiply by 2 rather than 1.28. So let me show you how to do it right.

1. Here are the links to the relevant pages of the reviews of the 35/1.4L and the 20/1.7



2. Here are the figures in lp/mm, center/edge. The camera we are comparing with, to keep the pixel count as similar as possible and not bias the results against the Canon, is the 5D, not the 20D. So make sure that you look at the right graph.

Canon 35/1.4L on 5D
1.4 23/12
2.0 26/14

Panasonic 20/1.7 on E-PL1
1.7 61/45
2.0 69/47

3. To convert to your preferred unit of lp/ih (line pairs per image height), we have to multiply the figures for the Canon by 24 (since the FF sensor is 24 mm high) and those for the Panasonic by 13 (since the MFT sensor is 13 mm high). We then have:

Canon 35/1.4L on 5D
1.4 552/288
2.0 624/336

Panasonic 20/1.7 on E-PL1
1.7 793/585
2.0 897/611

4. Alternatively, to keep the aspect ratio from affecting the comparison and to simplify the calculations, we might just as well just multiply the figures for the Canon by 2 for comparability, which yields the results I posted already on page 2 of this thread:

Canon EF 35/1.4L on 5D
1.4 46/24
2.0 52/28

Panasonic 20/1.7 on E-PL1
1.7 61/45
2.0 69/47

As I have already told you, what can be done is to compare the same lenses on different sensors and this is what Lenstip do when they replace one test camera by another within a system, e.g., Canon, Nikon or MFT. In some cases, such as when replacing the E-P1 by the E-PL1 it may be possible to isolate the effect of the difference in AA filter since the pixel count is exactly the same. But in most cases, it is only possible to say something about the effect of the sensor change. The point here is that the impact of such changes tend to be quite limited, and does not approach in magnitude the large performance differences we are talking about in this case. Again, this makes it highly unlikely that they could be explained away in the manner you suggest.

Where is the evidence that effect is limited? And what is "limited" in numbers? I don't see on their site lens retests in the quantity that would allow drawing the conclusions regarding the effect of AA filter. If there is a retest on E-P1 and E-PL1 please link to it.

You find comparisons of the kind I was talking about in the reviews conducted of new lenses about the time they switch from one camera to another. Here is an example from the time they switched from the Canon 20D (8 MP) to the Canon 50D (15 MP). In spite of the newer camera having nearly twice the pixel count of the older, the observed resolution figures do not, as you can see, increase by more than some 10 to 15 percent when you move from one camera-lens combination to the other:


 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH Panasonic Lumix G 20mm F1.7 ASPH +18 more
Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow