Reproducible results

From a reader:

Always good to test for one’s self, but it has been known for a long time that RAW is just that….unprocessed raw data from the sensor (though many suspect Nikon does do some manipulation….different philosophy from Canon).

RGB mode will change the look of your histogram, so consistency is important.

Certain camera settings, like d-lighting, or whatever Nikon calls it, will not directly change the raw file, but may change the exposure.  General, since I always shoot raw, I turn off all the “special features” as they are really intended to enhance in-camera post processing….just to be safe.

I’m glad you (the reader) brought this up, because it gives me a chance to talk about a philosophy of this blog that’s been developing over several years. In what follows, “you” refers to everybody reading this.

Conventional wisdom is usually right. Especially in the Internet age, conventional wisdom is more than occasionally wrong. You can look at the provenance of web postings to get an idea of the likelihood that they are accurate, but sometimes you can’t find some bit of information that you need that you know you can trust. In those cases, if I have the means, I like to test for myself.

If I perform some testing and reach a conclusion, I often post the results in this blog. However, I try not to just say what I’ve concluded; I try to explain the testing that I did to reach that conclusion. Otherwise, why should you believe me? I’m just another voice on the web.

I think that reproducible results are important, so, in most cases, I don’t just ask that you believe my test results; I give you enough information to reproduce those results for yourself.

I do render judgments on the results of some testing, such as the resampling testing here and elsewhere in this blog, but in that case, I provided you JPEGed sRGB versions of the images that I used to make those judgments, so that you could see if your judgments are similar. I try to do similar things in other cases. If you really care about my results, or if you’re the least bit suspicious of my methods, do your own testing; I try to give you enough information that you can construct the entire experiment for yourself.

The above only applies to technical topics on this blog. When I’m talking about artistic things, workflows, user interfaces, or similar things, I feel free to offer my unsupported opinions, which you all are welcome to accept, reject, or debate. I especially like debate, because I almost always learn something from the discussion.