This was an interesting and helpful discussion. I’m looking forward to receiving the Z9 and have no doubt that Nikon elevated the read-out speed sufficiently and that they would not otherwise release their first mirrorless flagship camera. The many videos posted by photographers who have used and tested the Z9 illustrate that the electronic-only shutter works fine, and the ability to accurately focus so quickly is amazing. (See, for instance, the roughly 400 images of a bird flying toward the photographer that was shown during The Camera Store’s livestream several days ago.)
Agree, does the Z9 utilize LIDAR? Or laser autofocus assist like newer smartphones? On the front top left side, just to the tight of the main shutter release, there is a round black or clear sensor or light, normally where there is an AF assist lamp on past DSLR's, etc.
I'm a Nikon D850 user, and I'm taken by the specs of the new Z9 .. your video gave me huge knowledge about what I was searching for. I guess i can buy it with peace of mind ... Thank you ... Please share more. you are well appreciated ...
Glad found your video. Got my z9 yesterday. I took a still image but when hit play or preview, I got a black screen, no image. I imported to computer, also shown a black screen instead of the image I shot. Any suggestion?
I had the chance at a z9 seminar to test it, unfortunately it's a big problem at Higher (faster) shutterspeeds in mixed lighting.. if you turn on the 120fps mode it's completely useless, the stripe banding is causing around 20 stripes or more in your frame.
Very good explanation of rolling shutter concepts and how both mechanical and electronic shutters realise this. I just wanted to point out that your conclusion on the Z9 RS speed is not quite right. RS speed should not necessarily be bound to the flash sync speed in the way you've articulated - it must only be faster than it. If you check out Matt Granger's Z9 review "The Nikon Z9 is EPIC!", he does an actual flash sync test of the Z9 and it works perfectly at 1/250 - 1/320 had a banding at the top. Therefore the RS speed of the Z9 is likely faster than 4ms as it would have to be at least that speed to achieve that flash sync. Seems like a conservative flash sync specification by Nikon.
What do you mean by at least that speed? Flash sync speed is the time it takes a mechanical shutter to move the full height of the frame/sensor. Shutter speed is how long after the first curtain starts opening that the second curtain starts to follow it (such that each pixel of the sensor receives light for the shutter speed time amount of time). These are exact times.
@@cooloox I think you've missed the conclusion of the video where there is an attempt to work out the possible rolling shutter (RS) speed of the Z9 by considering the Flash Sync (FS) speed. My point is that this approach is not correct. Firstly, the Nikon specified FS speed is well inside what is actually possible (1/250, Matt Granger), and secondly the relationship between RS and FS is such that RS must be faster than FS to make it all happen - as you say, the whole sensor has to be exposed to flash. Therefore RS must be faster than 4ms on the Z9.
@@cooloox Flashes aren't instant. A normal compact flash at full power will burn for about one millisecond. So X-sync timing in all cameras is designed such that there is at least one millisecond of the entire sensor exposing. And the numbers work out just fine - Nikon says 1/270 readout speed, specifies 1/200 X-sync, difference is 1.2 ms. People go ahead and test it and find it still works fine at 1/250, because they're not using full power. If you wanna stop fast motion with a flash, use a low power setting. 1/128 on one of mine takes me down to about 1/30000s (30 µs) burn duration.
Flash sync speed is only about coordination (or the lack thereof) of external device (a very quick flash) and camera internals... it says it in the word sync, like synchronisation?
thanks for this video, it clarify a lot for me. one thing doesn't make sense for me, using the electronic shutter or normal shutter the read out of the sensor is still line by line and not global. Why is it that we see banding in electronic shutter and not in the normal shutter. also if shoot faster than flash shutter speed will the read out of the sensor be faster than the 4-5MS?
Global shutter CMOS sensors are clearly the future but if you do some web search you will find that there are some obstacles with global shutter CMOS sensors which technology and development will need to solve.
I may have to watch this again to be sure I get it, but please help me out here. Is "banding" - the effect of having dark stripes in the photo - a result of rolling shutter or are they two separate items? As a concert photographer, I work in situations with stages illuminated by LED lights all the time. On two occasions, I used a "silent shutter" mode. My shots from those occasions were horribly unusable due to the "venetian blinds" (banding) running through the photos. I'm told this has to do with the frequency at which the LED would "flicker", and that if I were to use a variable shutter speed that could match the LED frequency (I think the Sony A1 has this, but the Z9 does not), I could get around the problem. So ... from thus video, it seems like the Z9's system is fast enough to avoid the "distortion" caused by rolling shutter, but I'm not really sure it's telling me that it should also prevent "banding" in LED-lit scenes. (I remain hopeful. :) )
Hi Ian, on the mechanical shutter banding in flash photography means horizontal darker areas where the flash light is (partially) blocked due to a shutter speed faster than the flash sync speed. The rolling shutter effect applies to electronic shutter on corresponding CMOS sensors.
@@mathphotographer I probably don't have a proper handle on the terminology. :). My concern is that when I used the electronic/silent shutter mode of my Z6ii at a concern that was lit with LED stage lights (I never use a flash at a concert), 99% of the pictures came out looking like I was viewing the scene through venetian blinds - dark bars appearing throughout the image, running the length of the long side side of the photo. Never had that issue when I used the mechanical shutter. Call it whatever it is, I just want assurances that the Z9 wont have such "venetian blinds" when I photograph concerts. What do you think? Has the speed of the sensor eliminated this concern?
@@igibbs2000utube I believe that is also banding. I know they have a flicker reduction setting in the z9. Z6ii I can’t use anti flicker with silent shutter. I’m interested to see how flicker reduction will pan out. I’m assuming good otherwise I don’t think it would be released without a mechanical shutter. Could be wrong though. 🤷🏻♂️
@@FlyinRyanProductions406 I've sent a question in to NIkon. The first person replied "shouldn't have any additional banding in the Z9 because it uses the same mechanical shutter as other cameras". Uh. No. So i've called in to a second person and hope to get an answer within 1-2 days.
I agree. The most recent sports cameras, Sony A1 and Nikon Z9 provide reasonably fast data processing and readout - so we will see more of that. Lets get rid of mechanical parts in cameras :) Better durability for sure!
Wait, so the readout of the sensor is 12X faster? Not the processor? Everyone is saying the processor is 12X faster than the Z6/7ii, but it's actually just the readout speed?
You can not consider readout and processor speed in disconnected way: if one of the two is slower, then that's constituting the binding constraint and will slow down the whole data process.
This explains the difference in electronic vs mechanical and rolling VS global shutter very nicely by physically showing us.... which I was already fully aware of. The flash sync speed is the time a curtain takes to travel fully across the frame and your shutter speed is actually how long after the first curtain leaves the second curtain starts to follow it. It also, obviously, equals the amount of time any part of the sensor is exposed to light. What I don't understand is why even using mechanical shutter doesn't give the 'same' rolling shutter effect when the sensor is still scanning line by line, regardless of whether you shoot electronic or mechanical shutter? In fact, if the rolling shutter of the sensor isn't perfectly keeping up with the mechanical shutters, then wouldn't the mechanical shutters get in the way and block part of the image, similar to shooting above the flash sync speed while using flash?
You don't understand why the mechanical shutter does not create rolling shutter, because that's not the case - mechanical FPL shutters produce rolling shutter because they ARE a rolling shutter. You're completely right!
Nice video - the question remains to be answered though, if the sensor is "scanned" only once. It might be scanned 5 times in that 1/200 sec (or 5 ms) and each scan may end up in separate registers that might be compared in order to compensate for motion blur. Note the Z 9 has way more rows and columns than a current classical sensor where the "few" additional rows and columns served raw processing. As "raw processing" bluntly put is the mathematically precise and repeatable wild-assed guessing for missing colors, a sensor without the additional rows and columns would require a separate "edge algorithm" in that wild-assed guessing. Now with way more additional rows and columns in the Z 9, these must be part of its IBIS and potentially serve a purpose like hypothesized above. In short, we don't know. And this applies too, to very long exposures where motion blur is all over the place. A slow blended scan will not give as uniform a motion blur of the night sky as we see in such shots. As 1/200=5ms=0.005 sec and processors work in GigaHertz territory, a 1 GHz equals 1/1,000,000,000 sec which means something happens 5,000,000 times in the 5 ms. If that's a sinus of 1GHz and we use the up and down half of one period we can double the number. Likely the limitation will be in the C-MOS design and maybe someone can shed a light on that aspect. If we go back to "flash" then also note that cheap flash kit at maximum flash output have a flash duration of 1/200 sec. A shorter exposure with whatever shutter means you cut the flash off early and waste energy when you cannot fully utilize the power you thought you had bought. In High-Speed Sync you get more flash/energy efficiency but loose a lot of power/intensity. With Auto-FP you might prefer HyperSync as the sync moment between camera and flash is much earlier (per Auto FP) and so you can use the 1/200 flash time to properly shoot much faster "shutter" times. This would be a thing with the Hasselblad with its leaf shutter at 1/2000 sec that generates a big question mark with all "sensor scan" analyses.
I'm not sure what the last few minutes were explaining. Some birds dive at over 300 Km/h. A tennis ball can move at 200. It's probably me but something didn't sound right!
Read-out time is ~ 1/270s measured. A mechanical shutter has a scan speed of around 1/300s yet sync speed is still often only 1/250s. X-sync is always slower than scan speed. Sony A1 read speed is 1/240s so Nikon is about 10% faster.
Thanks - can you provide a reference/source for the measurement? I saw the 1/270s in blogs but no reference, the only reference (as explained) is the flash sync speed. I think some TH-camrs tested flash for 1/250sec which still works, as I said in the video, but so far I did not find hard coded evidence for the sometimes claimed 1/270sec.
Although I do not use flash at all, some sports photographers use it for indoor sports like hockey and basketball. In these cases they want a very short flash sync speed. At this point in time, the Sony A1 can deliver a full frame sync speed of 1/400 and a half frame sync speed of I think 1/500 with the mechanical shutter. So, if you are shooting sports indoors with flash, and you want the shortest sync speed, you would probably choose the A1.
High speed sync (HSS or FP) will still work with the Z9 up to 1/8000s. At 1/400 you'd loose about one stop of the flash power. Faster than that, and A1 has the same problem (e-shutter sync speed 1/200s). But if shooting solely with flash lights, the sync speed doesn't matter so much, unless the flash duration is really slow. Flash will freeze any motion.
So I can understand better, 5ms (.005 seconds) is 1/200 of a second. How are the motion freezing effects of faster shutter speeds achieved by the camera? I have a Fuji GFX 50R which has the choice of electronic or mechanical shutter. The electronic shutter is only useful if nothing is moving very much and you don't want the sound of the mechanical shutter for some reason. I understand that the Z9 can choose to only use a portion of the readout for exposure purposes, but selecting 1/1000 would seem to have the same effect as shooting at 1/200 with an ND6 filter. In other words, no advantage in freezing action. What am I missing here? Is there an elephant in the room only I can see or is it all going completely over my head?
@@mathphotographer My pleasure. By the way, you could re-use much of this content in a more generalised video about flash sync speed and HSS. This is probably a more thorough explanation on that subject, and better illustrated, than I have seen elsewhere.
If the readout time is 4 ms. (1/250 of a second) and the sync speed of the Z9 is 1/250 of a second. AND the readout of lines of pixels is done sequentially, immediatly after exposure. Then there is no time for the flash to go off. By the time the last line of pixels get exposed the first line of pixels is read out. There is virtually no time between the last line 'openening' and the first line 'closing'. So the readout time must be faster or it is not done sequentially. Modern flashes have a short flash duration, but it is not zero. (SB 900 flash duration is 1/880 sec). Older flashes have often longer flash durations. So or the readout time is faster than 4 ms or the readout of the sensor is not done sequentially immediatly after exposure. The 4 ms and 5 ms are probably deferred from the synctimes and the is reproduces and repeated over the internet. (If there is another source please say so). Travel time of a shutterblade over the film area was about 1/400 of a second. The second blade would start for example 1/200 of a second later, so the complete shutter is open for the duration of 1/400 of a second. Enough of the modern flashes, not enough for the older powerfull flashes. For the flash there needs to be a duration where all pixels are sensitive for light. (CF flash is a different technique, there the flash keeps flashing during the complete shutter operation, most imaging-light going lost on the shutterblades).
Would be interesting with a sensor where you could control the integration process rather than be limited to the readout speed. Hopefully they manage this somehow in the future, or just manages so high readouts that it does not matter for a camera with 1/32000. But than again, we will want higher shutter speed then so it will still not work.
Well one way would be to fix the issue of "one line at a time" and read the whole sensor during one time unit, which should eliminate the rolling shutter. Either way, a new type of sensor.
Thanks. I have a further question. In flash photography, say in a dark studio, the light comes only from the strobe, and this flash duration can be as fast as 1/10000s. I don’t understand how it can illuminate the entire frame in that short period if it takes 5ms to completely scan the sensor? Traditionally we say electronic shutters are not compatible with strobes. But how does the z9 do it?
I takes 4ms to start accumulation (exposure) and to stop exposure/read out the full sensor. In shutter speeds of 1/250 and below, all sensor rows are accumulating at the same time - thus, the whole sensor will capture the flash. Then read out starts and it takes 4ms to read all lines off the sensor. Above 1/250, sensor accumulation on the first rows exposed stops before all lines on the sensor have started accumulation, thus some lines will not "see" a flash exposed at that point. Just like in a mechanical shutter. At speeds up to its sync speed, the front curtain is able to fully open and expose the entire sensor before the rear curtain begins to close. At speeds faster than sync speed, the rear curtain has to start moving before the front curtain is fully open. Thus you have the "moving slit" of mechanical shutters duplicated in an electronic shutter.
That's why readout time in the Z9 is ~1/270 but X-sync is only ~1/200 - full power flash burns for about one millisecond, which is pretty much exactly the difference between these two. If you use faster strobes, then you should not see any problems with 1/250 X-sync on the Z9.
I don't believe the flash synch speed (1/200s - 1/250s)) is same as the sensor readout speed. The readout speed has to be faster. Nikon has specified 3.7ms or 1/270s as the Z9 readout speed. The remaining 1-1.3ms is for the flash light, when the electronic shutter is "fully open".
Nothing much has changed in 70 years..... my 4x5 'speed' graphic showed rolling shutter when using the rear 'focal plane' shutter and had flash synch at 1/500 sec with the front (in lens) shutter.
Indeed - thx for confirming, applies to the faster cameras of course. My Hasselblad X1D II is basically not capable to do timelapses on elechronic shutter as shown in the timelapse in the video. Hasselblad also points this out in the manual and refer to a slow sensor readout time.
This Chanel consistently provides great reviews, comparisons and in this case explications for otherwise difficult to understand concepts All for free Your not paying enough to make rude critical remarks, next time you feel expressing your trivial concerns, Don’t
The concept of transitivity goes back at least 2,300 years. In the Elements, Greek mathematician Euclid of Alexandra (c. 325-c. 265 BC) includes it as one of his 'common notions. if A=B, B=C then A=C
You should have paid attention in school, The concept of transitivity goes back at least 2,300 years. In the Elements, Greek mathematician Euclid of Alexandra (c. 325-c. 265 BC) includes it as one of his 'common notions. if A=B, B=C, then A=C
@@longrider9551 Theory is one thing, practice is another. Steve Perry also pointed out this problem on A1. Probably so much about you and your fantastic results in school :D
@@prokremelskidezolati1426 transitivity is a proven concept not a theory, are you referring to Mr. Perry's last video? The one in which he said the A-1 is the best wildlife camera he ever used to date? maybe you could oblige and provide me the timestamp wherein he points out a problem with rolling shutter.
This was an interesting and helpful discussion. I’m looking forward to receiving the Z9 and have no doubt that Nikon elevated the read-out speed sufficiently and that they would not otherwise release their first mirrorless flagship camera. The many videos posted by photographers who have used and tested the Z9 illustrate that the electronic-only shutter works fine, and the ability to accurately focus so quickly is amazing. (See, for instance, the roughly 400 images of a bird flying toward the photographer that was shown during The Camera Store’s livestream several days ago.)
Agree - the Z9 will be fine with electronic shutter in 99% of shooting situations
Agree, does the Z9 utilize LIDAR? Or laser autofocus assist like newer smartphones? On the front top left side, just to the tight of the main shutter release, there is a round black or clear sensor or light, normally where there is an AF assist lamp on past DSLR's, etc.
@@patricksmith2553 AF-Assist illumination
I'm a Nikon D850 user, and I'm taken by the specs of the new Z9 .. your video gave me huge knowledge about what I was searching for. I guess i can buy it with peace of mind ... Thank you ... Please share more. you are well appreciated ...
Well done! The only question I have how the rear curtain sync is now done?
this video clear my doubts, thanks Math. Appreciate
Great explainer on shutter and read out workings - many thanks!
I am happy it helped, thx!
Glad found your video. Got my z9 yesterday. I took a still image but when hit play or preview, I got a black screen, no image. I imported to computer, also shown a black screen instead of the image I shot. Any suggestion?
Thank you for taking your time out to explain
Very well explained Math! You certainly know your stuff! I always walk away knowing more about the amazing digital camera, and I thank you for that!
Thank you Lucy!
I have one point that you missed out. Z9 e shutter speed is at 1/270 s, which gives a fastest scan speed of 3.7 ms.
Great. Curious, where did you find this info? DP Review?
Yes , you can check DP review there is a table called Comparison .
Thanks - can you provide a reference/source for the measurement? Web Link?
I try to share , for some reason it’s not supported or getting removed . Please check the dpreview for the details
Nikon communicated the readout time in verbal communication. It is 1/270s or 3.7ms in still. It is about 15ms in 8k or oversampled 4k video modes.
Cool - so even faster, even less issues with rolling shutter, great to hear - thanks for contributing to the discussion.
This video is very informative. Thank you!
Tried hard to explain the concept ☺️
Very interesting 🧐
Thanks Gary 👊
Thank you for your video! What are your thoughts on the Z9's performance when using strobes or when shooting in mixed lighing?
I will test the camera when I have it, will get back when I had the camera in my hands.
@@mathphotographer - thank you!
I had the chance at a z9 seminar to test it, unfortunately it's a big problem at Higher (faster) shutterspeeds in mixed lighting.. if you turn on the 120fps mode it's completely useless, the stripe banding is causing around 20 stripes or more in your frame.
What about studioflashes and the electronic shutter?
High speed sync is still an option
Very good explanation of rolling shutter concepts and how both mechanical and electronic shutters realise this. I just wanted to point out that your conclusion on the Z9 RS speed is not quite right. RS speed should not necessarily be bound to the flash sync speed in the way you've articulated - it must only be faster than it. If you check out Matt Granger's Z9 review "The Nikon Z9 is EPIC!", he does an actual flash sync test of the Z9 and it works perfectly at 1/250 - 1/320 had a banding at the top. Therefore the RS speed of the Z9 is likely faster than 4ms as it would have to be at least that speed to achieve that flash sync. Seems like a conservative flash sync specification by Nikon.
What do you mean by at least that speed? Flash sync speed is the time it takes a mechanical shutter to move the full height of the frame/sensor. Shutter speed is how long after the first curtain starts opening that the second curtain starts to follow it (such that each pixel of the sensor receives light for the shutter speed time amount of time). These are exact times.
@@cooloox I think you've missed the conclusion of the video where there is an attempt to work out the possible rolling shutter (RS) speed of the Z9 by considering the Flash Sync (FS) speed. My point is that this approach is not correct. Firstly, the Nikon specified FS speed is well inside what is actually possible (1/250, Matt Granger), and secondly the relationship between RS and FS is such that RS must be faster than FS to make it all happen - as you say, the whole sensor has to be exposed to flash. Therefore RS must be faster than 4ms on the Z9.
Thanks for the positive feedback and thanks for the comment, clearly Matt Granger is the reference for Nikon gear!
@@cooloox Flashes aren't instant. A normal compact flash at full power will burn for about one millisecond. So X-sync timing in all cameras is designed such that there is at least one millisecond of the entire sensor exposing. And the numbers work out just fine - Nikon says 1/270 readout speed, specifies 1/200 X-sync, difference is 1.2 ms. People go ahead and test it and find it still works fine at 1/250, because they're not using full power.
If you wanna stop fast motion with a flash, use a low power setting. 1/128 on one of mine takes me down to about 1/30000s (30 µs) burn duration.
Nice explanation.
thx
Flash sync speed is only about coordination (or the lack thereof) of external device (a very quick flash) and camera internals... it says it in the word sync, like synchronisation?
Thanks again for this video
Thanks!
what i want to know is how can we use flash with no mechanical shutter?
thanks for this video, it clarify a lot for me.
one thing doesn't make sense for me, using the electronic shutter or normal shutter the read out of the sensor is still line by line and not global.
Why is it that we see banding in electronic shutter and not in the normal shutter.
also if shoot faster than flash shutter speed will the read out of the sensor be faster than the 4-5MS?
Global shutter CMOS sensors are clearly the future but if you do some web search you will find that there are some obstacles with global shutter CMOS sensors which technology and development will need to solve.
I may have to watch this again to be sure I get it, but please help me out here. Is "banding" - the effect of having dark stripes in the photo - a result of rolling shutter or are they two separate items?
As a concert photographer, I work in situations with stages illuminated by LED lights all the time. On two occasions, I used a "silent shutter" mode. My shots from those occasions were horribly unusable due to the "venetian blinds" (banding) running through the photos. I'm told this has to do with the frequency at which the LED would "flicker", and that if I were to use a variable shutter speed that could match the LED frequency (I think the Sony A1 has this, but the Z9 does not), I could get around the problem.
So ... from thus video, it seems like the Z9's system is fast enough to avoid the "distortion" caused by rolling shutter, but I'm not really sure it's telling me that it should also prevent "banding" in LED-lit scenes. (I remain hopeful. :) )
Hi Ian, on the mechanical shutter banding in flash photography means horizontal darker areas where the flash light is (partially) blocked due to a shutter speed faster than the flash sync speed. The rolling shutter effect applies to electronic shutter on corresponding CMOS sensors.
@@mathphotographer I probably don't have a proper handle on the terminology. :). My concern is that when I used the electronic/silent shutter mode of my Z6ii at a concern that was lit with LED stage lights (I never use a flash at a concert), 99% of the pictures came out looking like I was viewing the scene through venetian blinds - dark bars appearing throughout the image, running the length of the long side side of the photo. Never had that issue when I used the mechanical shutter.
Call it whatever it is, I just want assurances that the Z9 wont have such "venetian blinds" when I photograph concerts. What do you think? Has the speed of the sensor eliminated this concern?
@@igibbs2000utube I believe that is also banding. I know they have a flicker reduction setting in the z9. Z6ii I can’t use anti flicker with silent shutter. I’m interested to see how flicker reduction will pan out. I’m assuming good otherwise I don’t think it would be released without a mechanical shutter. Could be wrong though. 🤷🏻♂️
@@FlyinRyanProductions406 I've sent a question in to NIkon. The first person replied "shouldn't have any additional banding in the Z9 because it uses the same mechanical shutter as other cameras". Uh. No. So i've called in to a second person and hope to get an answer within 1-2 days.
@@igibbs2000utube bwhahahahhahah! Wow! Lol. That’s a interesting response. 😂😂😂
I’m not sure how I stumbled onto your channel, but subscribe and thumbs up!
Ha! Thanks :)
This is just the beginning of the shutterless camera. The development of the organic sensor will further accelerate this development.
I agree. The most recent sports cameras, Sony A1 and Nikon Z9 provide reasonably fast data processing and readout - so we will see more of that. Lets get rid of mechanical parts in cameras :) Better durability for sure!
Wait, so the readout of the sensor is 12X faster? Not the processor? Everyone is saying the processor is 12X faster than the Z6/7ii, but it's actually just the readout speed?
You can not consider readout and processor speed in disconnected way: if one of the two is slower, then that's constituting the binding constraint and will slow down the whole data process.
This explains the difference in electronic vs mechanical and rolling VS global shutter very nicely by physically showing us.... which I was already fully aware of. The flash sync speed is the time a curtain takes to travel fully across the frame and your shutter speed is actually how long after the first curtain leaves the second curtain starts to follow it. It also, obviously, equals the amount of time any part of the sensor is exposed to light.
What I don't understand is why even using mechanical shutter doesn't give the 'same' rolling shutter effect when the sensor is still scanning line by line, regardless of whether you shoot electronic or mechanical shutter? In fact, if the rolling shutter of the sensor isn't perfectly keeping up with the mechanical shutters, then wouldn't the mechanical shutters get in the way and block part of the image, similar to shooting above the flash sync speed while using flash?
You don't understand why the mechanical shutter does not create rolling shutter, because that's not the case - mechanical FPL shutters produce rolling shutter because they ARE a rolling shutter. You're completely right!
Nice video - the question remains to be answered though, if the sensor is "scanned" only once. It might be scanned 5 times in that 1/200 sec (or 5 ms) and each scan may end up in separate registers that might be compared in order to compensate for motion blur. Note the Z 9 has way more rows and columns than a current classical sensor where the "few" additional rows and columns served raw processing. As "raw processing" bluntly put is the mathematically precise and repeatable wild-assed guessing for missing colors, a sensor without the additional rows and columns would require a separate "edge algorithm" in that wild-assed guessing. Now with way more additional rows and columns in the Z 9, these must be part of its IBIS and potentially serve a purpose like hypothesized above. In short, we don't know. And this applies too, to very long exposures where motion blur is all over the place. A slow blended scan will not give as uniform a motion blur of the night sky as we see in such shots.
As 1/200=5ms=0.005 sec and processors work in GigaHertz territory, a 1 GHz equals 1/1,000,000,000 sec which means something happens 5,000,000 times in the 5 ms. If that's a sinus of 1GHz and we use the up and down half of one period we can double the number. Likely the limitation will be in the C-MOS design and maybe someone can shed a light on that aspect.
If we go back to "flash" then also note that cheap flash kit at maximum flash output have a flash duration of 1/200 sec. A shorter exposure with whatever shutter means you cut the flash off early and waste energy when you cannot fully utilize the power you thought you had bought.
In High-Speed Sync you get more flash/energy efficiency but loose a lot of power/intensity. With Auto-FP you might prefer HyperSync as the sync moment between camera and flash is much earlier (per Auto FP) and so you can use the 1/200 flash time to properly shoot much faster "shutter" times. This would be a thing with the Hasselblad with its leaf shutter at 1/2000 sec that generates a big question mark with all "sensor scan" analyses.
Nice educational video indeed.
Thanks!
I'm not sure what the last few minutes were explaining. Some birds dive at over 300 Km/h. A tennis ball can move at 200. It's probably me but something didn't sound right!
The Nikon Z9's Low Light Video Performance Assessment & Examples I Would Like to See.
Very nice!
Thanks!
Read-out time is ~ 1/270s measured. A mechanical shutter has a scan speed of around 1/300s yet sync speed is still often only 1/250s. X-sync is always slower than scan speed. Sony A1 read speed is 1/240s so Nikon is about 10% faster.
Thanks - can you provide a reference/source for the measurement? I saw the 1/270s in blogs but no reference, the only reference (as explained) is the flash sync speed. I think some TH-camrs tested flash for 1/250sec which still works, as I said in the video, but so far I did not find hard coded evidence for the sometimes claimed 1/270sec.
Although I do not use flash at all, some sports photographers use it for indoor sports like hockey and basketball. In these cases they want a very short flash sync speed.
At this point in time, the Sony A1 can deliver a full frame sync speed of 1/400 and a half frame sync speed of I think 1/500 with the mechanical shutter. So, if you are shooting sports indoors with flash, and you want the shortest sync speed, you would probably choose the A1.
yes indeed
High speed sync (HSS or FP) will still work with the Z9 up to 1/8000s. At 1/400 you'd loose about one stop of the flash power. Faster than that, and A1 has the same problem (e-shutter sync speed 1/200s). But if shooting solely with flash lights, the sync speed doesn't matter so much, unless the flash duration is really slow. Flash will freeze any motion.
4 mS sounds good to me and is very close to the 1/270 quoted by DPreview.
So I can understand better, 5ms (.005 seconds) is 1/200 of a second. How are the motion freezing effects of faster shutter speeds achieved by the camera? I have a Fuji GFX 50R which has the choice of electronic or mechanical shutter. The electronic shutter is only useful if nothing is moving very much and you don't want the sound of the mechanical shutter for some reason. I understand that the Z9 can choose to only use a portion of the readout for exposure purposes, but selecting 1/1000 would seem to have the same effect as shooting at 1/200 with an ND6 filter. In other words, no advantage in freezing action. What am I missing here? Is there an elephant in the room only I can see or is it all going completely over my head?
Imaging a mechanical shutter hitting 120 fps. Gonna catch a fire.
😅🤣😂
Ha! YES 🤣
The incisiveness of your logic is a beautiful thing!
Ha! Thanks Steve :)
@@mathphotographer My pleasure. By the way, you could re-use much of this content in a more generalised video about flash sync speed and HSS. This is probably a more thorough explanation on that subject, and better illustrated, than I have seen elsewhere.
If the readout time is 4 ms. (1/250 of a second) and the sync speed of the Z9 is 1/250 of a second. AND the readout of lines of pixels is done sequentially, immediatly after exposure. Then there is no time for the flash to go off. By the time the last line of pixels get exposed the first line of pixels is read out. There is virtually no time between the last line 'openening' and the first line 'closing'. So the readout time must be faster or it is not done sequentially. Modern flashes have a short flash duration, but it is not zero. (SB 900 flash duration is 1/880 sec). Older flashes have often longer flash durations.
So or the readout time is faster than 4 ms or the readout of the sensor is not done sequentially immediatly after exposure.
The 4 ms and 5 ms are probably deferred from the synctimes and the is reproduces and repeated over the internet. (If there is another source please say so).
Travel time of a shutterblade over the film area was about 1/400 of a second. The second blade would start for example 1/200 of a second later, so the complete shutter is open for the duration of 1/400 of a second. Enough of the modern flashes, not enough for the older powerfull flashes. For the flash there needs to be a duration where all pixels are sensitive for light. (CF flash is a different technique, there the flash keeps flashing during the complete shutter operation, most imaging-light going lost on the shutterblades).
Would be interesting with a sensor where you could control the integration process rather than be limited to the readout speed. Hopefully they manage this somehow in the future, or just manages so high readouts that it does not matter for a camera with 1/32000. But than again, we will want higher shutter speed then so it will still not work.
Well one way would be to fix the issue of "one line at a time" and read the whole sensor during one time unit, which should eliminate the rolling shutter. Either way, a new type of sensor.
I knew that I could rely on you for the truth! I now have camera envy though!!
Thanks. I have a further question. In flash photography, say in a dark studio, the light comes only from the strobe, and this flash duration can be as fast as 1/10000s. I don’t understand how it can illuminate the entire frame in that short period if it takes 5ms to completely scan the sensor? Traditionally we say electronic shutters are not compatible with strobes. But how does the z9 do it?
I takes 4ms to start accumulation (exposure) and to stop exposure/read out the full sensor. In shutter speeds of 1/250 and below, all sensor rows are accumulating at the same time - thus, the whole sensor will capture the flash. Then read out starts and it takes 4ms to read all lines off the sensor. Above 1/250, sensor accumulation on the first rows exposed stops before all lines on the sensor have started accumulation, thus some lines will not "see" a flash exposed at that point.
Just like in a mechanical shutter. At speeds up to its sync speed, the front curtain is able to fully open and expose the entire sensor before the rear curtain begins to close.
At speeds faster than sync speed, the rear curtain has to start moving before the front curtain is fully open.
Thus you have the "moving slit" of mechanical shutters duplicated in an electronic shutter.
That's why readout time in the Z9 is ~1/270 but X-sync is only ~1/200 - full power flash burns for about one millisecond, which is pretty much exactly the difference between these two. If you use faster strobes, then you should not see any problems with 1/250 X-sync on the Z9.
I don't believe the flash synch speed (1/200s - 1/250s)) is same as the sensor readout speed. The readout speed has to be faster. Nikon has specified 3.7ms or 1/270s as the Z9 readout speed. The remaining 1-1.3ms is for the flash light, when the electronic shutter is "fully open".
Thanks - we were looking for hard-coded sources for that information. If it is 3.7ms: even better, and thanks for adding to the discussion.
Try a hummingbird wing in flight
Nothing much has changed in 70 years..... my 4x5 'speed' graphic showed rolling shutter when using the rear 'focal plane' shutter and had flash synch at 1/500 sec with the front (in lens) shutter.
A1 mechanical shutter flash sync of 1/400th in full frame and 1/500th in aps-c is the camera that's gonna get the job done.
Are you trying to make a point?
Are you seriously saying that Nikon have made a flagship camera for 2022 that will not get the job done?
@@freetibet1000 If you do fllash you would know how important it is to have a high flash sync speed.
@@cooloox yes and if you don't get my point, you don't know flash.
You can shoot 1/8000s with HSS with the Z9 no problem.
👍
Absolutely no issues with the electronic shutter in 99% of situations. For all intents and purposes, virtually identical to a mechanical shutter.
Indeed - thx for confirming, applies to the faster cameras of course. My Hasselblad X1D II is basically not capable to do timelapses on elechronic shutter as shown in the timelapse in the video. Hasselblad also points this out in the manual and refer to a slow sensor readout time.
@@mathphotographer I was, as you may have assumed, specifically referring to the Z9.
@@RobertFalconer1967 Not good at all for stopping Hummingbird's wings.
Sync speed is pronounced SINK speed !
:)
Don't really understand why you brought the Sony A1 in the conversation, it ruined your brilliant demonstration.
Because in many blogs people compare the Z9 with the Alpha 1 :) sports camera vs sports camera
This Chanel consistently provides great reviews, comparisons and in this case explications for otherwise difficult to understand concepts
All for free
Your not paying enough to make rude critical remarks, next time you feel expressing your trivial concerns, Don’t
The concept of transitivity goes back at least 2,300 years. In the Elements, Greek mathematician Euclid of Alexandra (c. 325-c. 265 BC) includes it as one of his 'common notions. if A=B, B=C then A=C
@@longrider9551 ?
I would prefer a correct pronounciation Nikon Z!
not Nikon C……
The A1 "has no problems with the rolling shutter"? Maybe you should open your eyes! :))))))
5ms sensor readout will be fine for most situations ...
You should have paid attention in school, The concept of transitivity goes back at least 2,300 years. In the Elements, Greek mathematician Euclid of Alexandra (c. 325-c. 265 BC) includes it as one of his 'common notions. if A=B, B=C, then A=C
@@longrider9551 Theory is one thing, practice is another. Steve Perry also pointed out this problem on A1. Probably so much about you and your fantastic results in school :D
@@prokremelskidezolati1426 Where A=Attitude, B=Bad and C=what is your problem?!
@@prokremelskidezolati1426 transitivity is a proven concept not a theory, are you referring to Mr. Perry's last video? The one in which he said the A-1 is the best wildlife camera he ever used to date? maybe you could oblige and provide me the timestamp wherein he points out a problem with rolling shutter.
Second comment
Congrats! No.2 is cool 😎