Processing speed of Sony ARW vs Nikon NEF files

Software
1

I use RT 5.11, but I’m intentionally posting this in the general software category without additional tag to explicitly keep it open for input from users of other software.

About a year ago, I sold my Nikon Z 5 and got a Sony A7C II. The A7C II has 33MP vs the Nikon’s 24MP. I’m using lossless compression on the A7C II but the ARW files are nonetheless substantially larger than the NEF files.

However, I’ve found that at least in RT 5.11 (and I’d love to hear from users of both RT and other software) the processing speed on the ARW files is substantially faster than on the smaller, lower-resolution NEF files.

Is this a known issue, or maybe just my personal impression after all? How about software other than RT, is there a perceivable difference in processing speed between different manufacturers’ RAW file formats?

2

How do you measure processing performance? Did you time specific operations? I have files from Nikon Z6 and Sony A7IV and could compare too.

1 Like
3

A very broad question.

On my system, the Sigma SD9 X3F opens about twice as quick as the Lumix DC-G9 ARW in RawDigger. Each less than two seconds.

1 Like
4

I haven’t done any direct speed comparisons but using darktable and many picture files from many cameras in the PlayRaw category of this forum I haven’t noticed any camera file being exceptionally slow or fast. Maybe you could share two files for people to test on their systems.

3 Likes
5

I did some meaningless measurements of opening a couple of files with all processing off (neutral/base). All programs built from latest source code, Fedora 41, DT has OpenCL enabled. Did a few tries to warm up file system cache and average results. Numbers in seconds are approximate because it all happens so fast:

                     ART              RT               DT
NEF:                 1.3              1.0              0.8
ARW:                 1.8              1.1              0.8
1 Like
6

What kind of processing are you basing your observations on? All other things equal, the file type should only affect the speed of the image decoding step. The underlying image representation becomes the same after demosaicing. Almost every other step is influenced by other factors such as the resolution and image content.

3 Likes
7

I haven’t measured it and wouldn’t know how to, it’s just my impression when exporting edits with similar tool usage from RAWs of either type.

9

It’s probably my personal impression then. I simply had the feeling that the ARW files get processed on export a lot quicker. Maybe I’ve just become a hopeless Sony fanboy.

1 Like
10

OK, I tried exporting from ART and DT with all tools off and my default tools on:

               ART       ART Defs   DT      DT Defs
.nef           1.4       3.8        0.98    2.2
.arw           2         5.2        1.12    3.0

Smaller Nikon files are processed faster which can be expected. And DT with OpenCL is a speed demon.

1 Like