Dissapointing Sony A77 II High ISO performance

Started 6 months ago | Discussions thread
TrojMacReady
Senior MemberPosts: 8,517
Like?
2010 called....
In reply to fishywisht, 6 months ago

fishywisht wrote:

philbot wrote:

Since we know it has the same sensor as the A6000, SLT = 1/3 stop of light loss, so obviously whilst it's going to be 'worse', we know how much worse..

So you think you 'know' this? Sounds like you 'know' what Sony tell you. I 'know' from the results that the A77 noise was one stop away from its rivals. I didn't rely on a manufacturer making a figure up to impress me to judge the images. People can step back and reason that if 1/3 stop loss were true, it would give only 1/3 stop of light to the autofocus system, which would put it far behind its competitors.

Perhaps Sony are simply lying and that the loss in the mirror is at least 1/2 stop. Their claims would be easy to prove if anyone cares to remove their mirror and shine a torch through it.

This has been done in countless tests and comparisons. DXO measured binary data and compared the A55 and NEX 5N: 0.5 EV difference. And for example the A77 and NEX 7: 0.3 EV difference.

People have removed their mirrors and compared image brightness: about 0.5 EV difference (compensation required in a converter to equal brightness). The actual transmission values of the A77 mirror/beamsplitter (without camera) have been measured  in a lab at different angles. At the angle used, that turned out to be 70 to 80%. The former translates to 0.5 EV light "loss" and corresponds with values mentioned in the various Sony patents.

As for AF comment, that makes no sense whatsoever. The AF and metering sensors in classic DSLR designs also get about a third of the light (through the semi silvered center of the main mirror, bounced via a small secondary mirror behind it) while the rest is reflected to the OVF. Meaning, those AF sensors have to work with only a smaller portion of the light too.

The A77 understates ISO's, a few competitors overstate ISO's, that alone leads to wrong conclusions about signal to noise ratios of these cameras and the effect of the beamsplitter (between 0.3 and 0.5 EV as described above).

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
OKNew
Huh?New
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow