DXOMark: Sony A99 II sensor is worse than the A7R II, D800E, D810, K-1

[ahsanford]

The new horsepower-spec overloaded A99 II -- which I thought had the A7R II sensor in it -- apparently laid an egg in sensor testing:

http://www.thephoblographer.com/2016/12/01/dxomark/#.WEBTt5J_3-8

Question 1: If it has the A7R II sensor in it and a year's newer internals to process the data, shouldn't we be getting the same (or slightly better high ISO) results as the A7R II?

Question 2: How the hell did PB get this data? It's not public yet.

- A

 

[ahsanford]

Now publicly available:
https://www.dxomark.com/Reviews/Sony-SLT-Alpha-99-II-sensor-review-New-super-resolution-contender

Oof. I can't explain that one at all short of (speculating):

• Perhaps the A7R II sensor isn't in the A99 II
  
  
• Would Sony's famous fine print compression / 'not 14-bit' shenanigans have affected a sensor score? The A99 II manual implies that the camera dabbles with compression to hit its 12 fps target. Maybe they accidentially switched to 12 fps mode (for whatever reason) and spiked the punch?

- A

 

[Larsskv]

What surprises me the most is the drop in ISO performance vs the A7RII. Good high ISO performance is very important in an action oriented camera.

To me, this is yet another example of Sony's impossible-to-understand approach, almost on par with the (otherwise) heavily weather sealed FE 35 f1.4 lens, which they forgot to outfit with a rear rubber gasket around the lens mount.

To me, the best reason to stay clear of Sony is their random and poorly planned product devolopment.

 

[rs]

It's an SLT. That semi translucent mirror prevents a sizeable proportion of the light from being collected by the sensor.

 

[ahsanford]

It's an SLT. That semi translucent mirror prevents a sizeable proportion of the light from being collected by the sensor.

Yep. I forgot about the T in SLT. Thanks for reminding me!

Assuming DXO's method is 100% reproducible from only one copy of a camera [comically long coughing fit] and that the A7R II sensor is indeed in the A99 II with no other improvements in hardware/processing of high ISO files, it would tell us that the translucent mirror is gobbling up... what, 2/3 of a stop? Yowza.

- A

 

[Loswr]

Assuming DXO's method is 100% reproducible from only one copy of a camera [comically long coughing fit] and that the A7R II sensor is indeed in the A99 II with no other improvements in hardware/processing of high ISO files, it would tell us that the translucent mirror is gobbling up... what, 2/3 of a stop? Yowza.

Sony said it's about 1/2 a stop. I guess uber-innovative Sony can't quite match the 1/3-stop loss of Canon's Pellix from 1965.

 

[ahsanford]

Assuming DXO's method is 100% reproducible from only one copy of a camera [comically long coughing fit] and that the A7R II sensor is indeed in the A99 II with no other improvements in hardware/processing of high ISO files, it would tell us that the translucent mirror is gobbling up... what, 2/3 of a stop? Yowza.

Sony said it's about 1/2 a stop. I guess uber-innovative Sony can't quite match the 1/3-stop loss of Canon's Pellix from 1965.

But hey, it grabs 12 fps... if you don't mind compressed RAW files.

And it's got a jillion AF points for 12 fps... that are locked after the first exposure on f/4 lenses.

I hate Sony fine print. I really do.

- A

 

[Loswr]

Assuming DXO's method is 100% reproducible from only one copy of a camera [comically long coughing fit] and that the A7R II sensor is indeed in the A99 II with no other improvements in hardware/processing of high ISO files, it would tell us that the translucent mirror is gobbling up... what, 2/3 of a stop? Yowza.

Sony said it's about 1/2 a stop. I guess uber-innovative Sony can't quite match the 1/3-stop loss of Canon's Pellix from 1965.

But hey, it grabs 12 fps...

Canon beat that in 1984 – the F-1 High Speed had a pellicle mirror and shot 14 fps.

 

[ahsanford]

Canon beat that in 1984 – the F-1 High Speed had a pellicle mirror and shot 14 fps.

At that frame rate, I'm half surprised that the film didn't come in a tommy gun like film dispensary.

Is that monstrous bottom a film feed, power motor, more battery, etc?

- A

 

[Loswr]

Canon beat that in 1984 – the F-1 High Speed had a pellicle mirror and shot 14 fps.

At that frame rate, I'm half surprised that the film didn't come in a tommy gun like film dispensary.

Is that monstrous bottom a film feed, power motor, more battery, etc?

- A

That's the battery pack on the bottom – 24 V of power.

http://www.mir.com.my/rb/photography/hardwares/classics/canonf1/html/canonf1nhighspeed.htm

 

[ahsanford]

Now you're going to stay people used to start fires in the woods with flint and tinder. I have this sweet doodad that I just press a button to make fire when camping, so flint and tinder clearly never existed.

Also, what is film?

- A

 

[Loswr]

Now you're going to stay people used to start fires in the woods with flint and tinder. I have this sweet doodad that I just press a button to make fire when camping, so flint and tinder clearly never existed.

Clearly both flynt and tinder are still around.

 

[Mikehit]

And I was under the impression that this camera was going to slay the Canon sports models.

Stupid Sony. Not listening to their customers...going backwards...crippling a potentially perfect camera. Sounds to me like they will be bust in 2 years.

 

[Frodo]

Clearly both flynt and tinder are still around.

!!

FWIW, the Canon F1 also had a 250 frame film back.
Must be showing my age but I owned the original F1 for a time. Nice camera, but the sensor was identical to much lighter cameras.
I assume that the difference between the 99 II and the A7RII is the pellicle mirror.

 

[3kramd5]

I assume that the difference [in DXO metrics] between the 99 II and the A7RII is the pellicle mirror.

The mirror, the readout electronics (almost certainly hotter), etc.

 

[Larsskv]

It's an SLT. That semi translucent mirror prevents a sizeable proportion of the light from being collected by the sensor.

That makes sence. When that is the case, enabling the 12bit raw at high fps, on top of the loss of light due to the translucent mirror, that DXO mark score would probably be noticeably worse...

 

[9VIII]

It's still a big improvement on the old A99, it's just the fixed mirror holding back ISO performance by half a stop, which everyone always knew was the case.

The design still opens up opportunities that no other camera can match, it'll be very interesting to see how the end product turns out.

 

[tron]

Assuming DXO's method is 100% reproducible from only one copy of a camera [comically long coughing fit] and that the A7R II sensor is indeed in the A99 II with no other improvements in hardware/processing of high ISO files, it would tell us that the translucent mirror is gobbling up... what, 2/3 of a stop? Yowza.

Sony said it's about 1/2 a stop. I guess uber-innovative Sony can't quite match the 1/3-stop loss of Canon's Pellix from 1965.

I had (actually have but needs servicing) the EOS RT and I was under the impression that it lost 2/3s of a stop at film level.

 

[Mikehit]

IIRC correctly is is about 1/3 stop at the sensor and 2/3 at the viewfinder?

 

[tron]

IIRC correctly is is about 1/3 stop at the sensor and 2/3 at the viewfinder

From the booklet I still have: (reflection:transparency = 35 / 65)

So practically 1/3 of the light goes to viewfinder and 2/3 of the light goes to flim. These numbers are NOT stops.

This translates differently than the numbers. For example 1/2 of the light means 1 stop loss (of course!) not 1/2.