New SONY α9 featuring full-frame stacked CMOS sensor

[hogloff]

Zero lag plus human reaction time with 36ms factored in...

10ms lag plus human reaction with ?ms factored in for shutter lag

? = the rated lag for the mirrorless camera

 

Lots of theory...have you tried the EFV in the A9?

[Bernard ODonovan]

Lots of theory...have you tried the EFV in the A9?

The point of the thread is to get an understanding of the A9. Your input is useful as is others

The technology will improve and the tech will trickle down to affordable products

Canon already have a global shutter in their cinema product. The banding is less likely in that. They are mirrorless but use cooling fans...

Given the A9 price and issues, I will not try one, but definitely tech of the future to look forward to...

I suspect SONY have implemented multiple e shutters in the sensor to reduce the scan time to multiple columns starting at the same time but still slow enough to pick up banding. I am sure they will develop global shutter when that can be done with good DR etc

Still very interesting to see who can use this as is and how those that can't react if they bought one...

 

 

[BJL]

The 1DX2 has a Shutter Lag of 36ms in fast mode... 

Thanks: that 36ms is the figure that I was vaguely remembering as "about 40ms".
Aside: what is "fast mode"? Modified AF operation?

EVF lag just adds to the delay in framing and shooting which would be more of an issue with longer lenses and say birds in flight as an example...

10ms is impressive but is also visually noticeable...

You seem to think that the EVF lag has a different and worse effect than lag after pressing the shutter release, but I do not understand why. Let me lay out an example where the total lag is the same 36ms either way:

EVF camera: EVF lag 10ms, and then another 26ms lag between pressing the shutter release and the photo bing taken (for AF and first curtain closing).
SLR/OVF camera: No VF lag, 36ms between pressing the shutter release and the photo being taken (for AF and then raising the mirror).

Either way, if you press the shutter release at the moment when the image you want is seen in the VF, the image you get will be 36ms too late, so either way, the ideal solution is to anticipate, pressing the shutter release when the image you see in the VF is 36ms before the image you want. What matters most is that the lag is consistent, and that the action photographer gets used to dealing with the lag on the camera used.

One difference though: the OVF camera has to finalize AF before mirror-lifting starts, whereas the EVF camera only has to do it before the first curtain closing starts, and curtain closing time at about 5ms, far less than mirror-lifting time. So predictive AF has an easier job with good EVF camera, predicting a far shorter time into the future.

[BernardLanguillier]

In fact it is only a matter of time until EVF have zero time lag.

All it will take is the sensor's ability to stream to a buffer at full resolution all the time. It will then be possible to capture the image that was live at the exact moment when the shutter was pressed.

Cheers,
Bernard

[scyth]

EVF camera: EVF lag 10ms, and then another 26ms lag between pressing the shutter release and the photo bing taken (for AF and first curtain closing)....
One difference though: the OVF camera has to finalize AF before mirror-lifting starts, whereas the EVF camera only has to do it before the first curtain closing starts, and curtain closing time at about 5ms, far less than mirror-lifting time. So predictive AF has an easier job with good EVF camera, predicting a far shorter time into the future.

you forgot about EFCS that is (can be) used in most cases = https://www.mhohner.de/newsitem2/efcs ... means avoiding first mechanical curtain for dSLMs with through EVF shooting ... not to mention no sensor shocks from either mirror or curtain

[BJL]

you forgot about EFCS that is (can be) used in most cases = https://www.mhohner.de/newsitem2/efcs ... means avoiding first mechanical curtain for dSLMs with through EVF shooting ... not to mention no sensor shocks from either mirror or curtain

Indeed, I deliberately ignored EFCS, since there seem to be some situations where its results are (for now) inferior, and I wanted to make the most conservative and thus most defensible argument for the lag reduction advantages of going mirrorless. (Which Leica users have been telling us about for years!)

Likewise for my decision to ignore the "high res video buffering" tactic that Bernard mentioned, and that can even provide a little bit of "negative lag" for you press the shutter release too late: store the last half second of frames, for example. I believe that some compact cameras have done this, with about 60FPS buffering. However, I would expect this to be done with compression, so maybe with direct output of HEIF stills selected from an intra-frame compressed HEVC stream stored to a single file, given that H.265/HEVC/HEIF is a better candidate for that sort of merged still-video operation that the current mix of H.264/AVC, JPEG etc.

[Bernard ODonovan]

Thanks: that 36ms is the figure that I was vaguely remembering as "about 40ms".
Aside: what is "fast mode"? Modified AF operation?

No impact. Just snappier:

"Shutter release time lag C.Fn7
Normally, stabilization control is performed to stabilize the shutter-
release time lag. By setting [Shortened], this stabilization control can
be disabled to make the shutter-release time lag shorter.
 : Standard
 : Shortened
The shutter-release time lag is normally approx. 0.055 sec. at the
shortest. By shortening it, it can be as short as approx. 0.036 sec.
The shutter-release time lag will vary depending on the shooting conditions,
lens type, aperture, etc."

[Bernard ODonovan]

...You seem to think that the EVF lag has a different and worse effect than lag after pressing the shutter release, but I do not understand why. Let me lay out an example where the total lag is the same 36ms either way:

EVF camera: EVF lag 10ms, and then another 26ms lag between pressing the shutter release and the photo bing taken (for AF and first curtain closing).
SLR/OVF camera: No VF lag, 36ms between pressing the shutter release and the photo being taken (for AF and then raising the mirror).

Either way, if you press the shutter release at the moment when the image you want is seen in the VF, the image you get will be 36ms too late, so either way, the ideal solution is to anticipate, pressing the shutter release when the image you see in the VF is 36ms before the image you want. What matters most is that the lag is consistent, and that the action photographer gets used to dealing with the lag on the camera used.

One difference though: the OVF camera has to finalize AF before mirror-lifting starts, whereas the EVF camera only has to do it before the first curtain closing starts, and curtain closing time at about 5ms, far less than mirror-lifting time. So predictive AF has an easier job with good EVF camera, predicting a far shorter time into the future.

I do not want to convince you. It either takes a leap of faith or willingness to understand...

People who never shoot subjects that move quickly and erratically with longer lenses will not understand the importance of zero lag viewfinders...

In your case if you enjoy EVFs then you will be a happy camper... 

I think the options and developments are great but equally understand the OVFs still have a place in photography...

 

[hogloff]

I think the options and developments are great but equally understand the OVFs still have a place in photography...

Yep, right next to Tri-x film and the cube flash in the photography history museum. 

[BJL]

I do not want to convince you. It either takes a leap of faith or willingness to understand...

People who never shoot subjects that move quickly and erratically with longer lenses will not understand the importance of zero lag viewfinders...

I am willing to understand if someone offers valid arguments and evidence, but so far all you offer is the repeated claim that viewfinder lag matters more than shutter lag, even if the later is several times longer. I instead have attempted to explain why—as fas as I can tell—it is the total lag that counts, by which I mean how late the image will be taken if you press the shutter release at the moment that the scene you want is seen in the viewfinder. VF lag can be corrected for by anticipation in exactly the same way as release lag can be, so long as the photographer can learn to adjust for the amount of lag, which requires the lag time to be fairly consistent.

Thanks for the details on "fast mode":
Normally, stabilization control is performed to stabilize the shutter-release time lag. By setting [Shortened], this stabilization control can be disabled to make the shutter-release time lag shorter.
The shutter-release time lag is normally approx. 0.055 sec. at the shortest. By shortening it, it can be as short as approx. 0.036 sec. The shutter-release time lag will vary depending on the shooting conditions, lens type, aperture, etc.

If i am reading this right, the normal mode is stabilized to have a fairly reliable lag of 55ms, perhaps to help photographers who anticipate a fast changing scene by learning to hit the shutter release 55ms before the the desired scene will arrive in the VF, as I discuss above. On the other hand fast mode is faster but variable "as short as approx. 0.036 sec. The shutter-release time lag will vary depending on the shooting conditions, lens type, aperture, etc.". This sound more suitable for shooting without "lag anticipation", as with less predictable motion where the best one can do is hit the shutter release when the scene looks right in the VF. (And for duffers like me who do no have the skill to anticipate motion.)

I think the options and developments are great but equally understand the OVFs still have a place in photography...

I agree—SLRs with OVFs still have some advantages, but no longer the former EVF problem of greater total lag. (Rangefinders still win for overall lag!)

[Bernard ODonovan]

So just looking at the A9 again:

"When the auto or electronic shutter mode is selected the viewfinder frame rate is fixed at 60 fps"

That is 1000/60 = 16.6 Milliseconds refresh, and if we are generous, we will assume that there is no frame padding...

Also:

"When [Shoot. Start Disp.] is set to [On], blackout occurs only when the first image is shot (page 48)."

And:

"Notes on shooting with an external
flash
When shooting with an external flash,
light and dark stripes may appear on
the image if the shutter speed is set
to faster than 1/4000 seconds."

This relates to SONY developing some uncorrected lenses and then using a lens profile to correct the final image. They are not alone but the latest wide angle zoom has more distortion than the competition. Confirms the viewfinder is the uncorrected image so not 100%:

"Notes on shooting with the
viewfinder
• The image may be slightly distorted
near the corners of the viewfinder.
This is not a malfunction. When you
want to see the full composition
with all its details, you can also use
the monitor.

• If you pan the camera while looking
into the viewfinder or move your
eyes around, the image in the
viewfinder may be distorted or the
color of the image may change.
This is a characteristic of the
lens or display device and is not
a malfunction. When you shoot
an image, we recommend that
you look at the center area of the
viewfinder."

More on the EVF:

"When shooting with the viewfinder,
you may experience symptoms
such as eyestrain, fatigue,
travel sickness, or nausea. We
recommend that you take a break
at regular intervals when you are
shooting with the viewfinder.
In case you may feel
uncomfortable, refrain from using
the viewfinder until your condition
recovers, and consult your doctor
as necessary.
Notes on continuous shooting
During continuous shooting, the
monitor or viewfinder may flash
between the shooting screen and a
black screen. If you keep watching
the screen in this situation, you may
experience uncomfortable symptoms
such as feeling of unwellness. If you
experience uncomfortable symptoms,
stop using the camera, and consult
your doctor as necessary."

4K shooting, just out of interest:

"Notes on recording for long
periods of time or recording 4K
movies
• Depending on the camera and
battery temperature, you may be
unable to record movies or the
power may turn off automatically
to protect the camera. A message
will be displayed on the screen
before the power turns off or you
can no longer record movies. In
this case, leave the power off and
wait until the camera and battery
temperature goes down. If you turn
on the power without letting the
camera and battery cool enough,
the power may turn off again
or you may be unable to record
movies.
• Under high ambient temperatures,
the temperature of the camera rises
quickly.
• When the temperature of the
camera rises, the image quality may
deteriorate. It is recommended that
you wait until the temperature of
the camera drops before continuing
to shoot.
• The surface of the camera may get
warm. This is not a malfunction.
• If the same part of your skin
touches the camera for a long
period of time while using the
camera, even if the camera does
not feel hot to you, it may cause
symptoms of a low-temperature
burn such as redness or blistering.

Pay special attention in the following
situations and use a tripod, etc.
– When using the camera in a high-
temperature environment
– When someone with poor
circulation or impaired skin
sensation uses the camera
– When using the camera with
the [Auto Pwr OFF Temp.] set to
[High].
• Especially during 4K movie
shooting, the recording time may
be shorter under low temperature
conditions. Warm up the battery
pack or replace it with a new
battery."

[Bernard ODonovan]

...
If i am reading this right, the normal mode is stabilized to have a fairly reliable lag of 55ms, perhaps to help photographers who anticipate a fast changing scene by learning to hit the shutter release 55ms before the the desired scene will arrive in the VF, as I discuss above. On the other hand fast mode is faster but variable "as short as approx. 0.036 sec. The shutter-release time lag will vary depending on the shooting conditions, lens type, aperture, etc.". This sound more suitable for shooting without "lag anticipation", as with less predictable motion where the best one can do is hit the shutter release when the scene looks right in the VF. (And for duffers like me who do no have the skill to anticipate motion.)
...

Correct, but the difference is tiny in most cases of the first frame... Hence my comments on the other benefits of zero lag...

As I mentioned before if you have a camera that can do spray and pray style, then this is all less relevant...

 

[BJL]

So just looking at the A9 again:

"When the auto or electronic shutter mode is selected the viewfinder frame rate is fixed at 60 fps"

That is 1000/60 = 16.6 Milliseconds refresh, and if we are generous, we will assume that there is no frame padding...

I am not primarily interested in the specs of the A9 itself, since my original comment was to dispute the broad claim that EVF lag will always put EVF cameras at a disadvantage to SLRs.
But to set the record straight:
- that 60fps is with EFC, which in turn has essentially no shutter release lag, for a total lag of about that 16.6ms.
- with mechanical shutter, the EVF can operate at 120fps for a lag of maybe 8.3ms—then adding another 5ms or so to close the first curtain gives a total of about 13.3ms.
These two lags can then be compared to the minimum of 36ms for the fastest SLR shutter release that we know about.

[Bernard ODonovan]

I am not primarily interested in the specs of the A9 itself, since my original comment was to dispute the broad claim that EVF lag will always put EVF cameras at a disadvantage to SLRs.
But to set the record straight:
- that 60fps is with EFC, which in turn has essentially no shutter release lag, for a total lag of about that 16.6ms.
- with mechanical shutter, the EVF can operate at 120fps for a lag of maybe 8.3ms—then adding another 5ms or so to close the first curtain gives a total of about 13.3ms.
These two lags can then be compared to the minimum of 36ms for the fastest SLR shutter release that we know about.

The processing strain will affect the static specs as they hint here after the shooting starts:

"Notes on continuous shooting
During continuous shooting, the
monitor or viewfinder may flash
between the shooting screen and a
black screen. If you keep watching
the screen in this situation, you may
experience uncomfortable symptoms
such as feeling of unwellness. If you
experience uncomfortable symptoms,
stop using the camera, and consult
your doctor as necessary."

So will panning as the viewfinder needs to refresh lots of data... I am sure it is very good but they have been a bit coy. It takes time to read all these little facts and then finally paint a picture of what this camera is. I suspect it will be flashing the odd unchanged or blank frame as you move it round or hold the shutter down

It is great alternative for some but it is clear it does not replace an optical viewfinder for others...

[BJL]

Bernard, now that we are back on the original topic of the Sony A9, I agree with you: it is not the EVF camera that will dethrone the best SLRs and lenses from Canon and Nikon for high-level action photography.

Let's see how it goes after a few more design iterations from Sony, Olympus, Panasonic and Fujifilm.

[Bernard ODonovan]

...
- with mechanical shutter, the EVF can operate at 120fps for a lag of maybe 8.3ms—then adding another 5ms or so to close the first curtain gives a total of about 13.3ms.
These two lags can then be compared to the minimum of 36ms for the fastest SLR shutter release that we know about.

Sounds like the mechanical shutter is not that good in reality

https://m.youtube.com/watch?v=gNaVG7j_g2Y

Seems to have more lag regardless of the refresh, so as I suggested you need to consider how it performs with the rest of the camera using up processing time and the ability to sync all actions at high speed..

[BJL]

...  as I suggested you need to consider how it performs with the rest of the camera using up processing time and the ability to sync all actions at high speed..

And as I said, I am not defending the Sony A9 in particular; I only entered this discussion to dispute your claim that a bit of EVF lag is somehow worse than a even a substantially larger amount of shutter-release lag, so that EVF cameras will always be at a disadvantage.

BTW, you seem to be putting a lot of effort into "oppo research" on a camera that you are not interested in buying or using!

[Bernard ODonovan]

And as I said, I am not defending the Sony A9 in particular; I only entered this discussion to dispute your claim that a bit of EVF lag is somehow worse than a even a substantially larger amount of shutter-release lag, so that EVF cameras will always be at a disadvantage.

BTW, you seem to be putting a lot of effort into "oppo research" on a camera that you are not interested in buying or using!

This thread is to discuss the new SONY A9. The good, the bad and the ugly...

You may wish to post a new thread so you can discuss the merits of EVF's if you wish...

Forgive me if I do not join such a thread to debate as the example I just posted proves reality is often different to assumption...

I find the technology fascinating but SONY have some issues with this camera, no amount of EVF love can fix...

[BJL]

This thread is to discuss the new SONY A9. The good, the bad and the ugly...

You may wish to post a new thread so you can discuss the merits of EVF's if you wish...

It was in this thread that you made your claim about the permanent disadvantage of EVFs; that is why I am replying on that sub-topic here. But I think we have completed the exchange of information and opinions on that, so I will sign off here. (I do like the idea of a separate thread for EVF vs OVF debates though!)