navmannz
Offline
• • •
Upload & Sell: On
|
| p.1 #1 · p.1 #1 · Sony's star-eater algorithm not fixed by f/w 4.0 | |
Contrary to some initial enthusiasm (including my own), it appears that Sony's v. 4.0 firmware for the Sony A7RII (and SII) has not fixed the spatial filtering that they introduced with f/w v. 3.3. Testing it rigorously in the real world has been difficult, with a full moon holding sway over the southern skies, along with a lot of cloudy weather. However, I managed to shoot some frames during the weekend, targeting part of the Milky Way. With iso set at 6400 I shot four frames in APS-C mode with a ZM Distagon 35, altering the aperture settings to give equivalent exposure, i.e., F/2.8 for the 2.5 second exposures and F/4 for the 5 second exposures - long exposure noise reduction was on for one pair and off for the other.
When I imported them into LR, I had to drop the black slider to around -65 to compensate for the moon illumination of the atmosphere and raise the color temp to 8000 to get a reasonably neutral background, but the difference in star capture was immediately apparent, indicating that significant elimination of stars was still occurring in the longer exposures - see below...
To test this further, I loaded all four files into DeepSkyStacker, and using the debayering option for reading raw files, and using a star detection threshold of 10% I got the following results when registering the photos:
Exposure and numbers of stars detected
2.5 second LENR off - 9761
5.0 second LENR off - 3114
2.5 second LENR on - 9760
5.0 second LENR on - 3003
Similar sky background figures indicate that I was reasonably close to getting equivalent exposures across the four images. Unfortunately the big difference in numbers of stars detected between the 2.5 and 5 second exposures indicates that substantial elimination of stars (70%) is still occurring despite this upgrade.
Others are now reporting similar results, with a post by Mark Shelley (here) indicating that the spatial filtering algorithm has been modified so that the green channel is now unaffected, but overall it is still firmly in play. Others reports documenting the continuation of spatial filtering are also in that thread and in other threads on this site.
I've spoken to a Sony support person, who indicated that the Sony engineers 'are still working on a solution and that a firmware update will be issued in due course'. However, I think that would be ill-advised to slacken off our pressure on Sony to fix what is a significant issue for a number of users.
-John
2.5 sec (left) and 5.0 sec (exposures) at 2:1 - note green cast to stars in longer exposure
|
