I’ve been watching a few videos from someone who calls himself the angry photographer. He is quite popular and has a lot of subscribers. He posts rambling videos that explain his views on lenses and cameras in extremely repetitive fashion, but also with a lot of useful anecdotal information that comes from long experience. He clearly knows his cameras and lenses and so I tough it out as long as I can when I am in the mood to go exploring the world of Nikon.
But I have recently discovered that he also posts A:B tests, and that these are very poorly constructed, executed, and documented. (As in no documentation at all.)
Here is one specific example: a test of high ISO between the D500 crop sensor camera and the D750 full frame sensor camera. I noticed when I downloaded the samples that there is a serious flaw in his methodology. The flaw was aggravated by the complete lack of any explanatory text. People with a modicum of skills and experience should be able to see how he constructed and executed the test. This is basic scientific method where you explain clearly what you tested and how you performed the test so that others can replicate and thus validate your work.
If you download the test images (I looked at the raw images), you will find that the lens focal length and the distance to subject is the same, which renders the subject (a test chart) 1.5 times bigger in the frame for the D500 crop sensor camera.
In my opinion, this renders a noise test completely useless, as the D500 is given a 1.5x field of view advantage. Why do I say that? Because the effects of noise and noise reduction that we care about are the destruction of fine details and textures. On sensors with very similar pixel counts, filling the frame at the same angle of view is important so that the sensor size advantage is given its due. Else, you are just crippling the larger sensor, as was done in this test.
In other words, in real world scenarios the frame is always filled by the subject and thus either a different lens must be chosen for each sensor size, or a different position must be taken in order to equalize field of view. The above test artificially penalized the full frame sensor by equalizing the subject size by cropping the image. This uses a small subset of the available pixels unfairly, thus magnifying the noise.
He obviously has an agenda as mentioned in some of the comments, and this is the trick he uses to push it. But that does not make the test useful for anyone that might actually want to shoot these two cameras in similar scenarios.
So my opening comment was:
Yes, I did flirt with ad hominem there and I regret it. Because it obviously eliminated any possibility that an actual discussion would ensue.
He took exception to that comment and came back with a couple of rather lazy ad hominem arguments that are also devoid of any clarification.
I agree that I had conflated angle of view and crop factor and so he was probably standing his ground on some tiny detail or other, but choosing not to elucidate. As in speaking in sound bites without meaning. So I tried another tack:
Well, he began to shout and doubled down on the ad hominem arguments while again avoiding clarification or explanation:
Somewhere in there is apparently an argument based on a conflation of pixel pitch (the actual size of an individual photosite) with some sort of concept where a lens dumps light (to which I can find no explanatory reference anywhere on the Internet) and with a repeat of “crop is a crop is a crop” (which adds nothing and frankly makes almost no sense without appropriate context). In other words, this seems to be nothing but an angry outburst with ad hominem overtones. (As in, a rambling insult.)
So once more into the breach, but with a higher quality argument (I hope).
And when I hit enter on the best quality argument I could make, I find out that he has blocked me
Seriously? I had hoped to at least get him to explain to me how pixel pitch,light dumping (whatever that is) and crop is a crop is a crop somehow mitigate the mismatched angles of view in his test images. But he chose to leave his insults stand as the last word and create the impression that I bowed to the superior ego.
I don’t normally engage with trolls, but I really wanted to figure out what he was talking about behind all the shouting and insults. He chose to troll me and then block me to retain the last word.
So why did I go through the effort of putting this out there?
- He blocked me to get the last word, and the last word was a stream of ad hominem arguments with no substance. I tried for a discussion and did not shy away from further discussion, so I am certainly not going to let that record stand on its own.
- Readers who want to perform tests have to understand the equalization of variables before wasting time on creation of a test with no useful meaning or conclusions.
- Readers who want to publish tests must learn to explain their work and not savagely attack people who question it. I was a bit too dismissive in my opening comments, but I tried to clarify several times only to be met with insults. That is not the way for people to learn. And only rank beginners will tolerate the lack of clarity on the assumption that the video creator knows what he or she is doing.
So for those who want to get into testing and vlogging, the above is an example of what not to do. Remember the lesson, lest you be dropped onto the heap of failed vlogs.