j_photo
Veteran Member
I thought phase detect was the primary focus method?Contrast detect, the main AF method in Nikon ML,
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
I thought phase detect was the primary focus method?Contrast detect, the main AF method in Nikon ML,
I don't t know that that would be a true apples-to-apples comparison. You now have more variables to consider. The adapter and it's impact on AF speed (if any), the differences in lenses (some lenses will focus faster or slower on different bodies, especially on bodies that are not native to the lens, so putting an E-mount on a Z body) because I believe the impact there might be translating the AF algorithms and things like that.It would be interesting to see how the TZE-01 performs compared to the FTZ adapter. Would it rack focus an E-mount lens as fast as say an A7/A9 camera? If it is slower there then the issue isn't the FTZ adapter but the Z body itself.
Yeah, you're probably right. Just a thought.I don't t know that that would be a true apples-to-apples comparison. You now have more variables to consider. The adapter and it's impact on AF speed (if any), the differences in lenses (some lenses will focus faster or slower on different bodies, especially on bodies that are not native to the lens, so putting an E-mount on a Z body) because I believe the impact there might be translating the AF algorithms and things like that.It would be interesting to see how the TZE-01 performs compared to the FTZ adapter. Would it rack focus an E-mount lens as fast as say an A7/A9 camera? If it is slower there then the issue isn't the FTZ adapter but the Z body itself.
Not trying to nitpick, and it would be interesting but I'm not sure how much it would really tell us. That's sort of what I was getting at.Yeah, you're probably right. Just a thought.I don't t know that that would be a true apples-to-apples comparison. You now have more variables to consider. The adapter and it's impact on AF speed (if any), the differences in lenses (some lenses will focus faster or slower on different bodies, especially on bodies that are not native to the lens, so putting an E-mount on a Z body) because I believe the impact there might be translating the AF algorithms and things like that.It would be interesting to see how the TZE-01 performs compared to the FTZ adapter. Would it rack focus an E-mount lens as fast as say an A7/A9 camera? If it is slower there then the issue isn't the FTZ adapter but the Z body itself.
If the Z was slower than a Sony body, it would be telling us that Nikon has for whatever reason limited the power or racking speed for some reason. I understand the F-mount glass is completely different, but it does mean that Nikon could improve this with either a firmware update or possibly a FTZ VII (with a AF motor ?? LOL)Not trying to nitpick, and it would be interesting but I'm not sure how much it would really tell us. That's sort of what I was getting at.Yeah, you're probably right. Just a thought.I don't t know that that would be a true apples-to-apples comparison. You now have more variables to consider. The adapter and it's impact on AF speed (if any), the differences in lenses (some lenses will focus faster or slower on different bodies, especially on bodies that are not native to the lens, so putting an E-mount on a Z body) because I believe the impact there might be translating the AF algorithms and things like that.It would be interesting to see how the TZE-01 performs compared to the FTZ adapter. Would it rack focus an E-mount lens as fast as say an A7/A9 camera? If it is slower there then the issue isn't the FTZ adapter but the Z body itself.
In some respects, I think Nikon did hamper the Z cameras, and sort of "had a chance" (and still do) with the Z II's, but the marginal upgrade in terms of AF (mostly for eye AF and animal AF from what i can see, aside from AF-F in video) there wasn't much improvement. I hope that Nikon is reading the forums (not just here but everywhere) and will come up with some sort of update that can really make use of the second processor. Maybe this is Nikon's new approach... who knows. But from the additions I've seen in the Z II's I don't think it necessarily requires a second processor so I'm thinking Nikon has an updating coming at some point to really show off the Z (but they probably wanted to address some of the existing complaints ASAP like lack of a battery grip and second memory card slot first).If the Z was slower than a Sony body, it would be telling us that Nikon has for whatever reason limited the power or racking speed for some reason. I understand the F-mount glass is completely different, but it does mean that Nikon could improve this with either a firmware update or possibly a FTZ VII (with a AF motor ?? LOL)Not trying to nitpick, and it would be interesting but I'm not sure how much it would really tell us. That's sort of what I was getting at.Yeah, you're probably right. Just a thought.I don't t know that that would be a true apples-to-apples comparison. You now have more variables to consider. The adapter and it's impact on AF speed (if any), the differences in lenses (some lenses will focus faster or slower on different bodies, especially on bodies that are not native to the lens, so putting an E-mount on a Z body) because I believe the impact there might be translating the AF algorithms and things like that.It would be interesting to see how the TZE-01 performs compared to the FTZ adapter. Would it rack focus an E-mount lens as fast as say an A7/A9 camera? If it is slower there then the issue isn't the FTZ adapter but the Z body itself.
For the record, I sold my Sony stuff and went back to Nikon. I just haven't updated my gear list. As great as the AF was on the Sony I couldn't stand the ergonomics. Also tried the R6 but again, the Z bodies just felt better to use. I am also happier with Nikon's colour, much nicer than Canon/Sony.
But does it matter at all if some of the AF processing is done in the adapter versus in the camera body at all? I mean, you can view the Z mount camera body with the FTZ adapter attached as one whole unit, same thing as the Z camera body without the adapter attached.I don't think that is all the FTZ is. I think it has its own processor chip built in that does some protocol conversion work. The Z mount is certainly designed to make use with F mount easy, and Nikon has said that the F mount operational decisions are taken by the camera, not the adapter, but that's not the same as saying that it's a simple pass-through. That's for no other reason than that if it was a simple pass-through, we'd be seeing cheap knock-off products all over Ebay, and we don't, unlike, for instance, Four Thirds to Micro Four Thirds. I would hazard a guess that when they designed it Nikon very much had in mind not making it easy for the knock-off merchants.I'm going to say that it's not the FTZ or the lens for the most part that is the "problem". It's the implementation and algorithms that Nikon used in the Z that's the biggest factor to poor AF tracking. The FTZ is nothing more than an electronic extension tube if you will It has no optics and it just sends the info to and from the lens through the electrical contacts. It jus spaces the lens at the appropriate distance from the sensor needed for DSLR glass to work on the Z. That's all.
No, it doesn't matter at all. I was just commenting on the false idea that the FTZ is just a pass-though, simply because it isn't true.But does it matter at all if some of the AF processing is done in the adapter versus in the camera body at all? I mean, you can view the Z mount camera body with the FTZ adapter attached as one whole unit, same thing as the Z camera body without the adapter attached.I don't think that is all the FTZ is. I think it has its own processor chip built in that does some protocol conversion work. The Z mount is certainly designed to make use with F mount easy, and Nikon has said that the F mount operational decisions are taken by the camera, not the adapter, but that's not the same as saying that it's a simple pass-through. That's for no other reason than that if it was a simple pass-through, we'd be seeing cheap knock-off products all over Ebay, and we don't, unlike, for instance, Four Thirds to Micro Four Thirds. I would hazard a guess that when they designed it Nikon very much had in mind not making it easy for the knock-off merchants.I'm going to say that it's not the FTZ or the lens for the most part that is the "problem". It's the implementation and algorithms that Nikon used in the Z that's the biggest factor to poor AF tracking. The FTZ is nothing more than an electronic extension tube if you will It has no optics and it just sends the info to and from the lens through the electrical contacts. It jus spaces the lens at the appropriate distance from the sensor needed for DSLR glass to work on the Z. That's all.
All AF lenses have a processor built-in. The point is what the processor does. One of the main jobs that has to be done is matching the AF commands with the motor system in the lens. The F mount system was developed directly from the screwdrive system, and all that matching work is done by the camera. Essentially, the AF-S protocol just provides commands to turn the motor on and off and change it's direction. The lens sends back information on how it is actually moving in the same way as it did for the screwdrive system, and the camera's AFD processor uses this information to know when to turn the motor on and off. In more modern AF systems all this is done by the lens' processor. The camera sends commands saying how much it wants the focus moved, and the lens' processor arranges for the lens to be moved that much. This can be fine tuned to the actual motor and mechanics of the lens, whether it be a DC motor, stepper or ultrasonic. So each lens includes routines specifically for that kind of motor, which is not possible with the AF-S protocol. The native Z protocol is undoubtedly a modern AF system, based on theCanon EF model, not another development of the screwdrive system. It is built with a secondary protocol specifically for AF-S lenses, and we are told that it is the camera processor, not the FTZ one that is handling that. But it appears that the FTZ is not just a wire-though adapter, as were the Four Thirds to micro Four Thirds adapters. This is likely because the F and Z protocols work with different communication PHYs.The only difference is that apparently, some additional processing is needed to control a F mount lens when it is mounted on a Z camera opposed to a Z mount lens when it is mounted on a Z camera.
Would it be right to assume that Z mount lenses have their own processor built in, that enables fluent operation on a Z camera body, or is it right to assume that F mount lenses need additional processing (done in the FTZ adapter), because the AF drive works in a fundamentally different way?
Thank you, yes I understand the issue with Nikkor AF-S lenses lacking their own dedicated processor in a similar fashion to Canon EF lenses.No, it doesn't matter at all. I was just commenting on the false idea that the FTZ is just a pass-though, simply because it isn't true.But does it matter at all if some of the AF processing is done in the adapter versus in the camera body at all? I mean, you can view the Z mount camera body with the FTZ adapter attached as one whole unit, same thing as the Z camera body without the adapter attached.I don't think that is all the FTZ is. I think it has its own processor chip built in that does some protocol conversion work. The Z mount is certainly designed to make use with F mount easy, and Nikon has said that the F mount operational decisions are taken by the camera, not the adapter, but that's not the same as saying that it's a simple pass-through. That's for no other reason than that if it was a simple pass-through, we'd be seeing cheap knock-off products all over Ebay, and we don't, unlike, for instance, Four Thirds to Micro Four Thirds. I would hazard a guess that when they designed it Nikon very much had in mind not making it easy for the knock-off merchants.I'm going to say that it's not the FTZ or the lens for the most part that is the "problem". It's the implementation and algorithms that Nikon used in the Z that's the biggest factor to poor AF tracking. The FTZ is nothing more than an electronic extension tube if you will It has no optics and it just sends the info to and from the lens through the electrical contacts. It jus spaces the lens at the appropriate distance from the sensor needed for DSLR glass to work on the Z. That's all.
All AF lenses have a processor built-in. The point is what the processor does. One of the main jobs that has to be done is matching the AF commands with the motor system in the lens. The F mount system was developed directly from the screwdrive system, and all that matching work is done by the camera. Essentially, the AF-S protocol just provides commands to turn the motor on and off and change it's direction. The lens sends back information on how it is actually moving in the same way as it did for the screwdrive system, and the camera's AFD processor uses this information to know when to turn the motor on and off. In more modern AF systems all this is done by the lens' processor. The camera sends commands saying how much it wants the focus moved, and the lens' processor arranges for the lens to be moved that much. This can be fine tuned to the actual motor and mechanics of the lens, whether it be a DC motor, stepper or ultrasonic. So each lens includes routines specifically for that kind of motor, which is not possible with the AF-S protocol. The native Z protocol is undoubtedly a modern AF system, based on theCanon EF model, not another development of the screwdrive system. It is built with a secondary protocol specifically for AF-S lenses, and we are told that it is the camera processor, not the FTZ one that is handling that. But it appears that the FTZ is not just a wire-though adapter, as were the Four Thirds to micro Four Thirds adapters. This is likely because the F and Z protocols work with different communication PHYs.The only difference is that apparently, some additional processing is needed to control a F mount lens when it is mounted on a Z camera opposed to a Z mount lens when it is mounted on a Z camera.
Would it be right to assume that Z mount lenses have their own processor built in, that enables fluent operation on a Z camera body, or is it right to assume that F mount lenses need additional processing (done in the FTZ adapter), because the AF drive works in a fundamentally different way?
I think that is a misinterpretation of what has been said in this and other threads.Still, Nikon AF-S lenses are very fast and very accurate on dslr cameras, Nikon has perhaps the best dslr AF system in the market. If I understand correctly, Nikon cannot succesfully drive AF-S lenses on Z-cameras (combined with the FTZ-adapter), because the neccessary real time continuous feedback of the level of OOF needed for this to be done fast and accurately, cannot (at this time) be provided by the OSPDAF system.
I may have used an awkward phrasing. By "successfully" I meant: as intended by design in an optimal way. Because OSPDAF is now the new standard, I can imagine a lot of R&D is presently going into state of the art implementation of OSPDAF in imaging sensors. Nikon adopts a latest generation modern stacked sensor for the Z9, so it is still unknown how capable its OSPDAF will be. At least, that is what I assume.I think that is a misinterpretation of what has been said in this and other threads.Still, Nikon AF-S lenses are very fast and very accurate on dslr cameras, Nikon has perhaps the best dslr AF system in the market. If I understand correctly, Nikon cannot succesfully drive AF-S lenses on Z-cameras (combined with the FTZ-adapter), because the neccessary real time continuous feedback of the level of OOF needed for this to be done fast and accurately, cannot (at this time) be provided by the OSPDAF system.
Any Z camera (with an FTZ) is quite capable of successfully autofocusing AF-S (as well as AF-I and AF-P) F-mount lenses. The point is that some of these lenses may not autofocus quite as quickly in some situations as they would on a Nikon DSLR.
They do have a dedicated processor, it's just that the protocol isn't designed in a way that allows the dedicated processor to control the motor in the most effective way. With Canon style interfaces the camera teels the lens where it wants it to focus, and the lens' processor takes care of focussing it there, taking into account the electrical, mechanical and optical characteristics of that particular lens. With the Nikon protocol the camera is only telling the lens processor whether it wants the motor turned off or on and in which direction to drive it. The lens is telling the camera where the lens is currently focussed, and the camera processor is having to work out the best way to focus that lens but can't take into account the specific characteristics of the lens.Thank you, yes I understand the issue with Nikkor AF-S lenses lacking their own dedicated processor in a similar fashion to Canon EF lenses.No, it doesn't matter at all. I was just commenting on the false idea that the FTZ is just a pass-though, simply because it isn't true.But does it matter at all if some of the AF processing is done in the adapter versus in the camera body at all? I mean, you can view the Z mount camera body with the FTZ adapter attached as one whole unit, same thing as the Z camera body without the adapter attached.I don't think that is all the FTZ is. I think it has its own processor chip built in that does some protocol conversion work. The Z mount is certainly designed to make use with F mount easy, and Nikon has said that the F mount operational decisions are taken by the camera, not the adapter, but that's not the same as saying that it's a simple pass-through. That's for no other reason than that if it was a simple pass-through, we'd be seeing cheap knock-off products all over Ebay, and we don't, unlike, for instance, Four Thirds to Micro Four Thirds. I would hazard a guess that when they designed it Nikon very much had in mind not making it easy for the knock-off merchants.I'm going to say that it's not the FTZ or the lens for the most part that is the "problem". It's the implementation and algorithms that Nikon used in the Z that's the biggest factor to poor AF tracking. The FTZ is nothing more than an electronic extension tube if you will It has no optics and it just sends the info to and from the lens through the electrical contacts. It jus spaces the lens at the appropriate distance from the sensor needed for DSLR glass to work on the Z. That's all.
All AF lenses have a processor built-in. The point is what the processor does. One of the main jobs that has to be done is matching the AF commands with the motor system in the lens. The F mount system was developed directly from the screwdrive system, and all that matching work is done by the camera. Essentially, the AF-S protocol just provides commands to turn the motor on and off and change it's direction. The lens sends back information on how it is actually moving in the same way as it did for the screwdrive system, and the camera's AFD processor uses this information to know when to turn the motor on and off. In more modern AF systems all this is done by the lens' processor. The camera sends commands saying how much it wants the focus moved, and the lens' processor arranges for the lens to be moved that much. This can be fine tuned to the actual motor and mechanics of the lens, whether it be a DC motor, stepper or ultrasonic. So each lens includes routines specifically for that kind of motor, which is not possible with the AF-S protocol. The native Z protocol is undoubtedly a modern AF system, based on theCanon EF model, not another development of the screwdrive system. It is built with a secondary protocol specifically for AF-S lenses, and we are told that it is the camera processor, not the FTZ one that is handling that. But it appears that the FTZ is not just a wire-though adapter, as were the Four Thirds to micro Four Thirds adapters. This is likely because the F and Z protocols work with different communication PHYs.The only difference is that apparently, some additional processing is needed to control a F mount lens when it is mounted on a Z camera opposed to a Z mount lens when it is mounted on a Z camera.
Would it be right to assume that Z mount lenses have their own processor built in, that enables fluent operation on a Z camera body, or is it right to assume that F mount lenses need additional processing (done in the FTZ adapter), because the AF drive works in a fundamentally different way?
In the Canon style system the camera has just to monitor whether or not the lens is in focus, and if not make an estimate of the degree of OOF. It has to continuously monitor this and send new commands to the lens and to achieve accurate focus. The lens' processor can deal with these effectively for the specific lens, since it has all the information that it needs for that lens, so can develop a motor drive strategy that best fits the needs of that lens. For the Nikon system, the camera is continuously sending commands to start and stop the lens motor and change drive direction. It has to get the lens to the right position without the benefit of knowing how that lens works (though I suspect Nikon's later systems probably have a list of some of Nikon's lenses with information on the best way to operate them), just using motor on, off and direction commands.Still, Nikon AF-S lenses are very fast and very accurate on dslr cameras, Nikon has perhaps the best dslr AF system in the market. If I understand correctly, Nikon cannot succesfully drive AF-S lenses on Z-cameras (combined with the FTZ-adapter), because the neccessary real time continuous feedback of the level of OOF needed for this to be done fast and accurately, cannot (at this time) be provided by the OSPDAF system.
The problem is that PDAF has to sample from a very small patch of the exit pupil to be sufficiently selective to yield a good estimate of subject distance and a very fine distinction between in and out of focus, sufficiently that the PDAF can be relied on for fine focus. It takes a quite complex (and large) optical system to do that. OSPDAF simply makes the microlenses sample from one half or the other of the exit pupil, rather than a small patch. The advantage over pure CDAF is that it gives an indication of direction of OOF, so the lens doesn't have to hunt back and forth to find which way to turn. It's hard to see how it could be made as selective as DSLR PDAF.My lingering question is about this last statement: is the OSPDAF system inherently incapable of judging the level of OOF as succesfully as the traditional dedicated dslr AF sensor, or can the Nikon engineers evolve the OSPDAF capabilities (built into the sensor so I assume part of sensor design) to a level where AF-S lenses can be made to perform at or near dslr level? Be it through programming with the aid of a fast processor and stacked sensor design, or be it through hardware modifications to the sensor.
I think the Z9 is likely only to operate at its top AF capability with Z lenses.Ultimately, this may be a neccessary condition for the success of the coming Z9.
Thanks, I am interested in this because I was planning to add a 500mm f4E to my 500PF, when the development announcement for the Z9 came through.They do have a dedicated processor, it's just that the protocol isn't designed in a way that allows the dedicated processor to control the motor in the most effective way. With Canon style interfaces the camera teels the lens where it wants it to focus, and the lens' processor takes care of focussing it there, taking into account the electrical, mechanical and optical characteristics of that particular lens. With the Nikon protocol the camera is only telling the lens processor whether it wants the motor turned off or on and in which direction to drive it. The lens is telling the camera where the lens is currently focussed, and the camera processor is having to work out the best way to focus that lens but can't take into account the specific characteristics of the lens.Thank you, yes I understand the issue with Nikkor AF-S lenses lacking their own dedicated processor in a similar fashion to Canon EF lenses.No, it doesn't matter at all. I was just commenting on the false idea that the FTZ is just a pass-though, simply because it isn't true.But does it matter at all if some of the AF processing is done in the adapter versus in the camera body at all? I mean, you can view the Z mount camera body with the FTZ adapter attached as one whole unit, same thing as the Z camera body without the adapter attached.I don't think that is all the FTZ is. I think it has its own processor chip built in that does some protocol conversion work. The Z mount is certainly designed to make use with F mount easy, and Nikon has said that the F mount operational decisions are taken by the camera, not the adapter, but that's not the same as saying that it's a simple pass-through. That's for no other reason than that if it was a simple pass-through, we'd be seeing cheap knock-off products all over Ebay, and we don't, unlike, for instance, Four Thirds to Micro Four Thirds. I would hazard a guess that when they designed it Nikon very much had in mind not making it easy for the knock-off merchants.I'm going to say that it's not the FTZ or the lens for the most part that is the "problem". It's the implementation and algorithms that Nikon used in the Z that's the biggest factor to poor AF tracking. The FTZ is nothing more than an electronic extension tube if you will It has no optics and it just sends the info to and from the lens through the electrical contacts. It jus spaces the lens at the appropriate distance from the sensor needed for DSLR glass to work on the Z. That's all.
All AF lenses have a processor built-in. The point is what the processor does. One of the main jobs that has to be done is matching the AF commands with the motor system in the lens. The F mount system was developed directly from the screwdrive system, and all that matching work is done by the camera. Essentially, the AF-S protocol just provides commands to turn the motor on and off and change it's direction. The lens sends back information on how it is actually moving in the same way as it did for the screwdrive system, and the camera's AFD processor uses this information to know when to turn the motor on and off. In more modern AF systems all this is done by the lens' processor. The camera sends commands saying how much it wants the focus moved, and the lens' processor arranges for the lens to be moved that much. This can be fine tuned to the actual motor and mechanics of the lens, whether it be a DC motor, stepper or ultrasonic. So each lens includes routines specifically for that kind of motor, which is not possible with the AF-S protocol. The native Z protocol is undoubtedly a modern AF system, based on theCanon EF model, not another development of the screwdrive system. It is built with a secondary protocol specifically for AF-S lenses, and we are told that it is the camera processor, not the FTZ one that is handling that. But it appears that the FTZ is not just a wire-though adapter, as were the Four Thirds to micro Four Thirds adapters. This is likely because the F and Z protocols work with different communication PHYs.The only difference is that apparently, some additional processing is needed to control a F mount lens when it is mounted on a Z camera opposed to a Z mount lens when it is mounted on a Z camera.
Would it be right to assume that Z mount lenses have their own processor built in, that enables fluent operation on a Z camera body, or is it right to assume that F mount lenses need additional processing (done in the FTZ adapter), because the AF drive works in a fundamentally different way?
In the Canon style system the camera has just to monitor whether or not the lens is in focus, and if not make an estimate of the degree of OOF. It has to continuously monitor this and send new commands to the lens and to achieve accurate focus. The lens' processor can deal with these effectively for the specific lens, since it has all the information that it needs for that lens, so can develop a motor drive strategy that best fits the needs of that lens. For the Nikon system, the camera is continuously sending commands to start and stop the lens motor and change drive direction. It has to get the lens to the right position without the benefit of knowing how that lens works (though I suspect Nikon's later systems probably have a list of some of Nikon's lenses with information on the best way to operate them), just using motor on, off and direction commands.Still, Nikon AF-S lenses are very fast and very accurate on dslr cameras, Nikon has perhaps the best dslr AF system in the market. If I understand correctly, Nikon cannot succesfully drive AF-S lenses on Z-cameras (combined with the FTZ-adapter), because the neccessary real time continuous feedback of the level of OOF needed for this to be done fast and accurately, cannot (at this time) be provided by the OSPDAF system.
The Nikon system can be made to work very well with DSLR PDAF, because it just has to do one movement, which involves turning on the motor, monitoring how far the lens has moved, then turning it off. By contrast, on sensor PDAF can't give a good estimate of focus distance and is generally hybrid, relaying on a CDAF 'tweak' at the end of the focus movement to achieve fine focus. This tweak is hard to do with the Nikon style of motor control.
The problem is that PDAF has to sample from a very small patch of the exit pupil to be sufficiently selective to yield a good estimate of subject distance and a very fine distinction between in and out of focus, sufficiently that the PDAF can be relied on for fine focus. It takes a quite complex (and large) optical system to do that. OSPDAF simply makes the microlenses sample from one half or the other of the exit pupil, rather than a small patch. The advantage over pure CDAF is that it gives an indication of direction of OOF, so the lens doesn't have to hunt back and forth to find which way to turn. It's hard to see how it could be made as selective as DSLR PDAF.My lingering question is about this last statement: is the OSPDAF system inherently incapable of judging the level of OOF as succesfully as the traditional dedicated dslr AF sensor, or can the Nikon engineers evolve the OSPDAF capabilities (built into the sensor so I assume part of sensor design) to a level where AF-S lenses can be made to perform at or near dslr level? Be it through programming with the aid of a fast processor and stacked sensor design, or be it through hardware modifications to the sensor.
I think the Z9 is likely only to operate at its top AF capability with Z lenses.Ultimately, this may be a neccessary condition for the success of the coming Z9.
Beware of misleading marketing information. Just about every DSLR has a separate processor for AF, since they were derived from film DSLRs where the processor was only doing AF. By and large the AF codes were not migrated to the image processor. The job was relatively simple. PDAF modules do not have a whole load of pixels and the matching algorithm is quite straightforward. The growth in complexity came when they were coupled with metering sensors, which had become small imaging sensors coupled with pattern recognition for face detection, colour linked tracking and so on. That meant that the AF processor would be linked with the image processor to do that kind of image processing. All that happened in the D500 and D850 was that the AF processor was upgraded to have the power to do the image processing function as well.Good discussion.
There is a lot of difference in AF performance for other factors as well as those described. For example, the D500 and D850 have a separate processor for AF.
That's not necessarily the case. Modern systems on a chip generally carry a small raft of processors, and as Apple's recent M1 chip shows, integrating functions onto one chip will often increase performance. The real question is not how many processors, but what's on the processor.It was one of the big innovations with the D5/D500 release. The Z5/Z6/Z7 cameras have a single processor and that's going to make AF slower than the top alternatives. But due to the PDAF with a CDAF tweak, AF is more accurate than PDAF only.
I think the more speed is mainly to provide increased image processing power for clever AF modes such as face an eye detection, and subject recognition tracking. And also the doubled up processors allow for a second card port.The additional processor in the Z6II/Z7II is to provide more speed - speed to improve focus speed and improve processing and write speed. I still don't think the camera is fully optimized.
The main difference is that the distance information from the AF sensor can be used for tracking, what Nikon called '3D tracking'. That's not possible with PDAF, the strategy is instead to use subject recognition and motion tracking, which takes a lot more computational power.I can't speak to whether or not the On-sensor AF system is better or worse than the old AF sensor on the bottom of the mirror box. It certainly is more accurate as it eliminates AF sensor alignment issues.
Designers always need to make tradeoffs.As to whether the Canon approach or Nikon approach is better, I can't see a real difference in the AF performance overall, but individual cameras and lenses will have differences. There are too many factors involved to attribute differences to a single factor without also considering other components. For example, there are differences in AF speed related to lens design, AF motor, voltage or power being applied, pre-focusing, lens AF processor, camera AF processor, lens firmware, camera firmware, etc. In most cases you don't want to optimize speed but sacrifice accuracy, but you might make that trade-off in some cases. Add the FTZ to the mix and engineers and designers still have to make tradeoffs.
The point of the Z9 is adding an action/sports/wildlife camera to the Z system. Nikon will have a whole stable of Z lenses out by the time the Z9 is released, but not that many lenses in the action/sports/wildlife category, only two very expensive primes and one or two zooms.I think for now both have advantages and disadvantages, I currently shoot with two D5 bodies and a Z6. Having recently sold my D500 for a Z6, and I have to say I was and am pleased. I do photojournalism and wildlife photography. I had bought a Z6 originally when it came out and returned it within a week or so, and went back to a D500. This time around the Z6 has come a long way. Although I think the Z6 is fine for just about everything, it’s true the DSLR and D5/D6 are definitely better for action. So for now I’ll continue to shoot with my two D5 bodies for action, sports and wildlife photography. Although the Z6 will definitely be used for still subjects in wildlife as well. Being able to put a focus point anywhere and being able to be so quite or even silent, makes the Z cameras have certain advantages. If the Z9 is able to truly rival the D5/D6 in AF performance that will be a game changer. Yet it will have to also be that and do that with F-mount lenses for someone like myself to buy one. There is no way I could afford to replace my all F-mount lenses. I do not own any Z-mount lenses and don’t see myself buying any, anytime soon.
Fortunately I own all recently released AF-S and E-type lenses with the latest VR. I’m hoping Nikon can at least make these lenses work well with the Z9 and maybe even updated Z6 II and Z7 II. If Nikon can’t do this, even with the latest lenses like my 500FL and 70-200FL, they are in trouble. I don’t think there are many people who can afford to sell all of their F-mount lenses to switch to Z-mount. I know that I can’t and I won’t, so hopefully it’s not impossible. I look at all purchases with a mind of “can this object help me get shots I otherwise can’t get”? Or can this piece of gear improve my process and or dramatically improve my workflow, etc. The Z6 definitely allows me to get some images that I could not otherwise get in some situations. However so does the D5, for me I try to use the gear that will be the best tool for the job at hand. More and more often I find myself bringing one D5, and my Z6 as a combo. Both serve their intended purpose and both are great cameras. If the Z9 can combine the D5/D6 and the Z6/Z7 into one awesome camera, wow! I really hope that’s true and I really hope Nikon is capable of at least bringing the most recent E-type AF-S lenses to that point. Otherwise the Z9 will simply be a Z-mount only camera and too expensive to obtain. It will not sell too well and I for one will be disappointed. So I hope it’s possible to make the FTZ and Z9 a winning combo. I have high hopes, but honestly I won’t care if instead of buying a Z9, I buy a D6 or two. I’m going to upgrade my D5 bodies eventually and I honestly don’t care if that’s to the D6 or Z9. But I’m certainly not going to buy a Z9 if I can’t bring my F-mount lenses with me. I’m really hoping it’s not the FTZ or AF-S lenses that are at fault.