For anyone sensitive to how fast your camera displays each photo taken (automatic image review), here's an observation I made and measured today between Expeed 6 cameras (Z6/Z7/Z5/Z50/Z30) and Expeed 7 (Z6 III/Z5 IIZf/Z8/Z9/Z50 II):
On a ~20/24 MP sensor, Expeed 7 takes only 266 milliseconds (1/4 second) to enter image review vs 1,183 milliseconds (1.2 seconds) on Expeed 6.
Here's a video showing the Z30 (Expeed 6) vs Z50 II (Expeed 7), exposure triggered on timer. Same settings and SD media card type (Samsung 128GB EVO Plus). Image format is raw lossy compressed.
.
I reproduced this same Z50 II differential vs my original Z6. Note it's not sensitive to raw file type.
I think the most eye opening thing in the video is how much bigger the Z50II is compared to the Z30.
But yeah, the Expeed 7 is a pretty large step up, and image review is likely the least important, though certainly easier to test than AF algorithms and such. Cool stuff!
Jman13 wrote:
But yeah, the Expeed 7 is a pretty large step up, and image review is likely the least important, though certainly easier to test than AF algorithms and such. Cool stuff!
I agree image review is pretty low on the totem pole but it's so laggy on my Z6/Z7 that I intentionally avoid using it even in situations where it's useful.
Pretty snappy. I never use auto image review. I have the button under my right hand middle finger configured to review and use it when I need it. The fact that a mirrorless Z camera can display image review in the EVF is one of my favorite mirrorless features.
fjablo wrote:
Do we actually know whether all Expeed 6 or 7s are created equal?
Nikon could very well use different clock speeds & voltages to adjust the processor to the needs and thermal limitations of different bodies.
I suspect they do use different Expeed clock speeds based on the processing demands and media card interfaces of each body. I measured the Z9 and got roughly the same result as the Z50 II, at least within the margin of measurement error based on my crude measurement of counting video frames of the camera's exposure and playback. That's with 2x the pixels vs the Z50 II.
DWOfPaul wrote:
Interesting test. Do we happen to know if image review is faster on a dual Expeed 6 camera, such as a Z6II or Z7II?
It's not faster on the dual Expeed 6 cameras. Getting to image review requires applying the fully image processing pipeline to the raw data, meaning every step needed to produce the raw-embedded jpg (shooting raw) or standalone JPG (shooting jpg). Most of those steps are performed sequentially on the same set of data and thus can't be parallelized, which means they can only be done efficiently on one of the two Expeeds. What the dual Expeed provide is the ability to process multiple images in parallel, which is why they improve the effective buffer capacity of the cameras. I dive into all this here:
EB-1 wrote:
How does that particular review operation equate to anything else like AF?
EBH
It translates into much deeper buffer capacities because the camera can generate the finalized image data much faster for writing, which is almost always the bottleneck in buffer clearing rather than the speed of the media card. Here again is my post here which dives into the detail of Expeed operation.
I guess that we can expect the Expeed 8 to be around as much faster compared to the Exceed 7 as the 7 was vs the 6. And considering what Canon and Sony have done, I wouldn't be surprised if we had a 2 processors set up in the Z9II.
fjablo wrote:
Do we actually know whether all Expeed 6 or 7s are created equal?
Nikon could very well use different clock speeds & voltages to adjust the processor to the needs and thermal limitations of different bodies.
I’m thinking that the clock speed is faster, but there are also physical differences in busses, memory, etc. that make the two different. The Exceed 7 was built as more of a general purpose platform so Nikon could simply add new features via firmware and reduce the time/cost to market of new models.
bernardl wrote:
I guess that we can expect the Expeed 8 to be around as much faster compared to the Exceed 7 as the 7 was vs the 6. And considering what Canon and Sony have done, I wouldn't be surprised if we had a 2 processors set up in the Z9II.
Exciting times ahead.
Cheers,
Bernard
The Z8 is already so incredibly reactive that it's hard to imagine any meaningful improvements.
Aside from the automatic image review, which I deactivate immediately.
So, in terms of waking up from standby or turning on, it's basically all instant.
It might be because I only switched from Expeed6 to the Z8 a few days ago, but this thing is already incredibly fast in comparison and feels like switching from an underpowered van to a Ferrari.
story_teller wrote:
I’m thinking that the clock speed is faster, but there are also physical differences in busses, memory, etc. that make the two different. The Exceed 7 was built as more of a general purpose platform so Nikon could simply add new features via firmware and reduce the time/cost to market of new models.
What I meant was differences between different chips that Nikon calls "Expeed 7". Almost certain that there are differences from camera to camera (chip in the Z9 is not the same one used in the Z50II, despite both being called Expeed 7)
fjablo wrote:
What I meant was differences between different chips that Nikon calls "Expeed 7". Almost certain that there are differences from camera to camera (chip in the Z9 is not the same one used in the Z50II, despite both being called Expeed 7)
I would guess they're the same chip but as you indicated, perhaps running at different clock rates. I base this guess on the economics of ASIC design+manufacturing cost vs the volumes Nikon deals with.
Vento wrote:
The Z8 is already so incredibly reactive that it's hard to imagine any meaningful improvements.
Aside from the automatic image review, which I deactivate immediately.
So, in terms of waking up from standby or turning on, it's basically all instant.
It might be because I only switched from Expeed6 to the Z8 a few days ago, but this thing is already incredibly fast in comparison and feels like switching from an underpowered van to a Ferrari.
Indeed yes. The Nikons are already insanely reactive.
Most additional power would probably be directed at AF algos so as to run more performant convolution neural networks.
The more power you have the more parameters your model can have and the better it can become at recognizing features such as eyes. Especially when you consider the time series aspect. Meaning more sampling meaning more reactivity.
Other potential areas that would probably benefit are:
- higher frame rates
- higher EVF refresh rate
- …
Finally since the key metric is power per watt both for battery life and hear generation Nikon could also have decided to not only increase speed but also to decrease the power envelope of « their » chip.
