Pixel Perfect wrote:
I find I can get results from LR about 5x faster than from DPP and in most cases the output is similar. DPP has better colour, I think LR produces better fine detail, DPP handles noise a bit better, but I try to do as little luma NR as possible and leave that to a better tool like Topaz or imagenomic or DxO's prime NR.
+1 @ expedience vs. results bang for buck can leave DPP in the dust for a lot of folks. Then when we start talking about capabilities, DPP lags there again (imo). I spent a fair amount of time tweaking my PS workflow for sharpening, in lieu of using DPP or ACR. Noise OTOH is a different thing that I've not developed better than ACR yet.
My thought was just that OEM + OEM tends to mate well and since I found DPP to be better @ sharpening that ACR (back when), it might stand to reason that OEM NR would again have an edge over ACR NR (DXO might be interesting though).
My general experience has been that better results = more processing time = slower throughput vs. faster throughput = less processing time = lesser results. Classic GIGO vs. you get what you pay for (in time / power), no real surprise there @ no such thing as a free lunch.
After posting the DPP shots last night I realized that it probably isn't a fair comparison with the LR conversion. For the LR shots default sharpening was on and output sharpening was on. I'll try to post a couple LR conversions with no sharpening applied to see if that evens the playing field. Also I'll check what profile LR was using and switch it to one as close to DPP (Neutral) as I can. The conversions may be a lot more similar after I adjust those variables in LR.
Here are the LR conversions at 6400 ISO with Camera Neutral selected and no sharpening applied.
Thanks for posting the test. I haven’t tested newer bodies, but high iso is WAY better on the 30D than XT, imo, even though both are 8mp. I think they are different sensors, but I suspect that the 20D was better, too, and that was supposedly the same sensor.
That and some other stuff make me think in camera processing can make a big difference.
Also, smaller pixels have to be amplified more, so you get more noise. Although 20 to 24MP isn't that big of a deal imo.
When I look at performance, I evaluate the amount of detail retained at high ISO as part of a comparison. Just looking at Chroma is only part of the story.
Unfortunately, by using f/16, there will be no difference in retained detail due to diffraction, so its not possible to see if more NR could be applied to the 80D and still keep usable detail. What we really want is the maximum detail possible after NR is applied.
Its difficult to judge the 6400 shots a couple of posts up because of the apparent motion/defocus blur on the 80D shot. I have not compared high-ISO on them other than a quick 3200 comparison of books on a shelf in my computer room. I noted a slight advantage to the 7D2 at 3200, but I also noted some differences in the WB of the two cameras, despite using a gray card for each.
I would venture to say, prematurely at least, that the advantage from 100-400 ISO goes to the 80D, and they are close in the 800-1600 and maybe 3200 range, but the 7D2 holds a modest advantage beyond that.
Thanks for doing these tests and showing your results.
Edit - I guess I'd better get on my desktop and produce some results to share....
After Shooting with the 80D for a few days, the 7D2 is going to be sold.
Not a huge improvement, but an incremental improvement in 400-1600
ISO noise. Image quality is slightly better in terms of contrast. Autofocus
seems very capable. I like the build and speed of the 7D2, but image quality
and low light autofocus is important.
pointbob wrote:
Is there anyway to use dpp raw conversion in conjunction with lightroom editing?
Sure, tweak and export as 16-bit TIFF and open in LR. However, a DPP/PS combo makes more sense: PS has more tools and features and DPP can be setup to open the file in PS after exporting (that's what I do).
