Imatest results - numbers vs. categories
I tried doing a search to see if this topic has been discussed before, but I couldn't find anything. I apologize if it has been discussed.
I'm wondering what the correlation is between the old imatest results which were categorized as "excellent", "good", "fair", and "poor", and the new results which have real numbers? I'm guess that "good" is equivalent to the 2000 lw/ph line, and "excellent" is equivalent to the 2500 lw/ph line? Is this valid, or is there no correlation between the two? Thanks!
I'm wondering what the correlation is between the old imatest results which were categorized as "excellent", "good", "fair", and "poor", and the new results which have real numbers? I'm guess that "good" is equivalent to the 2000 lw/ph line, and "excellent" is equivalent to the 2500 lw/ph line? Is this valid, or is there no correlation between the two? Thanks!
Hi SkySpades,
The old system was used in order to try and simplify the results, as different cameras (12mp vs 24mp vs 36mp etc) are capable of recording different levels of resolution.
However, we switched to showing the actual figures, so that it makes it clearer to see exactly how well a lens performs on a camera - but you also need to be aware that the highest the score can be will be dependent on what camera body was used. For the Micro Four Thirds system, where nearly all cameras use a 16mp sensor it will be much easier to compare.
In our new reviews with the actual figures, we always use the same language such as "Excellent" etc in the text, so you can read how the lenses perform.
Hope this helps.
Josh
The old system was used in order to try and simplify the results, as different cameras (12mp vs 24mp vs 36mp etc) are capable of recording different levels of resolution.
However, we switched to showing the actual figures, so that it makes it clearer to see exactly how well a lens performs on a camera - but you also need to be aware that the highest the score can be will be dependent on what camera body was used. For the Micro Four Thirds system, where nearly all cameras use a 16mp sensor it will be much easier to compare.
In our new reviews with the actual figures, we always use the same language such as "Excellent" etc in the text, so you can read how the lenses perform.
Hope this helps.
Josh
Quote:Say a lens scores 2500 at imatest for a 16mp camera, what would the approximate imatest score be for a newer 20mp lens?
In a way it does not matter.
If the lenses are of equal quality the Imatest score using the same body would be only moderately higher.
A starting point expectation with a 25% increase in MP is about a 12.5% increase increase in sensor resolution (tested in isolation) reduced to about 6% increase in image resolution when combined with a lens of similar resolution (tested in isolation) and combined with the same camera body.
Much more useful is an indication of what reading is enough for a good quality print.
The image below prints well at A3 print size from about 12 MP after image cropping.
It is the last man on the 2014 Tour de France opening stage - being encouraged to make the finish line after a crash - see his right knee and the tear in his shorts.
While good equipment can help there is usually more to getting a good image than the equipment alone.
Thanks for the reply. So in the example above, testing the same lens on a OM-D-EM 1 Mark 1 vs a Mark II would add about 150 points to the 2500 score for the Mark 1? As to your other point, I somewhat disagree. I agree that there is an acceptable level of resolution that makes a print look good, and a little more resolution beyond that won't make much difference. I feel that a lot more resolution makes a big difference. The local art gallery has a community show once a year, and I can spot the prints that were made with say the 50mp Canon or the Sony 45mp. They have a hyper realism about them. They are on the far end of the normal curve. It's the 4500 Imatest scores versus the acceptable 1800. Just based on my opinion, it seems that 2000 is a level of a decent score, 50% higher at 3000 is looking noticeably different, and 50% higher still at 4500 is hyper realism territory. The other reason why the scores might make a difference is in birding where details are everything and cropping is severe. You can fix some noise, but you can't fix loss of detail.
