Noise comparison 1.09 vs 1.091

[thomi]

Advertisement (gone after registration)

I see a slight improvement from 1.09 to 1.091.

 

Fotos taken at 1 Sec, 2400 ISO.

[thomi]

On the floor it can be seen even better.

[dratt]

are you sure that there is a difference?!?

[rami G]

When Epson got their firmware update (R-D1s) the difference in high iso performance was dramatic. The 1600 iso become better than the 800iso before. It was not about searching for the minute differences. It was striking. And the white ballance (which was alraedy better, to my taste at least, than the 5D for artificial light) was also improved.

[scott kirkpatrick]

How about checking whether the vertical line, slightly more than half way across the frame, is gone in 1.091? The appears pretty reliably if you shoot at 1250 or 2500 and have dark areas which you then pull up a stop or more in a raw developer. Shutter speed has to be > 1/30 so that dark frame subtraction doesn't remove it, and this is a RAW artifact, not seen in the in camera jpegs.

 

scott

[ChrisC]

I see a slight improvement from 1.09 to 1.091.

 

Tonal developments are different, as is colour. From these examples, I cannot confidently share your conclusion. In any case, I am viewing screen resolution at over 17 times the print output size for 300 dpi; any print difference in these examples would be inconsequential.

 

...............Chris

[Venkman]

Advertisement (gone after registration)

How about checking whether the vertical line, slightly more than half way across the frame, is gone in 1.091? The appears pretty reliably if you shoot at 1250 or 2500 and have dark areas which you then pull up a stop or more in a raw developer. Shutter speed has to be > 1/30 so that dark frame subtraction doesn't remove it, and this is a RAW artifact, not seen in the in camera jpegs.scott

 

Scott,

 

I tried to reproduce it and where I was successful in the past, I can't provoke it with 1.091.

(the white thing left of center is a cable duct for my beamer)

 

I just shot a wall in my room @ 2500 1/16 and can't see it anymore (saw it in the past when I tried that). JPP from RAW pushed to +2.

 

Would be nice if it was true - maybe someone else more talented could confirm.

 

Dirk

[rosuna]

I don't know how much noise can be corrected on the RAW file by means of a firmware change.

 

CMOS sensors, in theory, have a lower signal-to-noise ratio, but Canon was able to offer superlow noise RAW files. How they do that?

 

The "full-frame transfer" CCDs show great quality (tonal gradation, dynamic range, low noise) at low ISOs, but that quality is not scalable. Why is it?

 

(The "fill factor" of FFT CCDs is near of 100%, but CMOS are around 60%).

 

.

[scott kirkpatrick]

Scott,

 

I tried to reproduce it and where I was successful in the past, I can't provoke it with 1.091.

(the white thing left of center is a cable duct for my beamer)

 

I just shot a wall in my room @ 2500 1/16 and can't see it anymore (saw it in the past when I tried that). JPP from RAW pushed to +2.

 

Would be nice if it was true - maybe someone else more talented could confirm.

 

Dirk

 

I've been shooting with 1.091 since it leaked, some low light stuff, and I can't see it either.

 

Let's hope they killed that one.

 

scott