Curiously Nikon specifically advertises the 24-70 VR for video use, while Tamron specifically forbids to use its corresponding lens for anything but stills shooting. It's likely mostly a matter of AF wear and tear on the lens, though, so video shooters who focus manually theoretically shouldn't worry. Still it's a policy written down in the Tamron documents and the tech in Tamron's service center was very quick to point to it when I mentioned that I specifically bought the VR Tamron for video.
For manual video focus the advantage of the Tamron over the Nikon is its focus ring being close to the body whereas on the Nikon it is at the tips of the long lens.
The drawback of the Tamron is its lose tolerances and bad quality control. Especially its mechanical aperture's precision leaves a lot to be desired.
Whoa, what is it?
a3dtot: I've been using Canon for years now and it seems that the faults that are found technically have little or nothing to do with actual results. I understand that we have a bias as to which camera brand we prefer and in my own opinion there are reasons. My old 5D mkII gives me everything I want. I have never looked at one of my prints and thought I wish I had more dynamic range like Nikon or whomever. Canon gives me exactly what I want and except for the resolution, which I would like to have 50mp, there has never been anything wrong. I get the incredible color that I feel only canon provides (bias) and I also get great black and white. The detail is as good or better than film, which for me is saying something. It seems to me that many Nikon people find the technical pluses and announce them but I don't see them discussing photo quality as the main plus for their camera. Just saying. (stirring it up a little)
Thanks for the explanation Iliah, that's understood. It was just that the article underlined the advantages of more well capacity and corresponding possibility for lower ISO, while I wondered how well this can be used in practice.
I am no landscape shooter myself and most of the time need unhealthy high ISO, so it's more of a curiosity/academic question for me, maybe more practically relevant for others. Not to mention that I am using a Nikon D750 myself. ;)
What about those cameras that had the mirror box (simple labelled "Front") replaced already? Is the shutter part of that box or a separate part behind that?
It's at least a valid point of critique. When I saw the DR curves and explanation about larger well capacities and lower ISO my first thought was: But can photographers make use of these in practice when exposure times are dictated by movement and light availability?
Not every landscape is bathed in high contrast summer sunlight without wind to allow long exposure times at ISO 64. In order to fill all those photon wells you need to capture enough photons. Leaves, birds and whatnot don't exactly hold still during that time. Of course the lower noise floor still allows to get away better with less photons, too. But obviously only up to a certain point.
So the proof is in the pudding, aka, in practice things don't necessarily line up as on graphs. As always it depends on the usage scenario. Right tool for the job.
Abhijith Kannankavil: they should have done something about the sensor manufacturing process. But, i guess for the intended purpose and customers, it may sometimes be a perfect match.
And, there's one thing that have always been in my mind when i read comparos between cameras. when you compare camera 'a' with camera 'b' which has lower pixel count, by downsizing the images from camera 'a' to the size of camera 'b',.... aren't you giving an advantage to the images from camera with lower pixel count? , because it is not edited at all, whereas the other one has gone through the Photoshop once. ? ? ?
One could even argue the other way around. More information (mp) captured and then downsized with a quality algorithm will provide more information in the final image.
So a 50 mp image downsized to 36 mp may provide more information than an original 36 mp image. That's the whole idea of providing more mp.
That being said, downsizing algorithms would benefit from applying low-pass filters to cut out high frequency information of the original image that wouldn't make it into the final downsize anyway. I would appreciate the results much more compared to the aliasing and "false detail" we usually get with most downsizing of today.
Danke für die gute Arbeit, Julia Reda!
I can follow all of this. But the enforced multi-point tracking with third party lenses still do deliver somewhat rough results in that you will get *some* part of the face in focus. And this often means the most protruding part, especially the nose. This is reminiscent of what my Olympus E-M5 did regularly when it recognized a face, but not eyes.
Frankly I don't understand why the algorithms that detect faces without eyes don't try to focus on the upper third of the face box instead of the usually most protruding center nose.
I would also like to know whether the A7R2 combines the PDAF information with CDAF or does it solely use the non PDAF sensor readout for pattern recognition?
Timur Born: The examples with third party lenses don't exactly leave me with awe, I do appreciate the effort, though. There are too many shots with the nose in focus and focus jumps around too much (ears, nose, eyes + cheeks).
If people are lured to rely on this instead of their own focusing skills (mine are woeful) then it might lead to more disappointment than what the high entry price suggests (body, adapter, lenses that you might or might not own yet).
Rob: My Olympus E-M1 has eye-detection, but on the LCD screen there's only the option to draw a box around the face (when a face is detected), with no indication of which eye is being focused on. You do get to choose which eye to automatically focus on in the menu though, as well as set it to automatic so the closest eye is focused on.
My 5D Mark III has face detection in live view using Contrast Detection, but there's no face detection with Phase Detection, nor is there eye detection. I wonder if Mark IV will have comparable features as the latest Sony cameras.
The E-M1 *does* display which eye it focuses on, but this only works with AF-S, not AF-C. So one superior feature of the Sony is to couple eye detection with AF-C.
@Birdkai: That is non-sense. I am coming from Olympus and went to Nikon D750, one reason was improved AF-C tracking. Especially Nikon's 3D tracking works reasonably well for tracking eyes of my ever moving kids in indoor light. So for the time being my request is met by the Nikon.
The video with the native Sony lens looks very promising. Those very large windows look like they let in lots of light, though. And the AF starts failing specifically when the camera get to close proximity, where camera movement really starts to matter. Still very welcome to finally get this for video. I do video with manual focus on the Nikon, because video AF is less than "meh", and then there isn't even focus peaking (I abuse zebras for that). One request not met by Nikon, but met well by Sony.
As far as I understood, single point AF is not available with third party lenses. But this is not the point. The vibe is that the AF system works fantastic with third party lenses, but my personal conclusion from the example video is that it is not reliable enough to depend on with shallow depth of field settings.
The video with the native lens is more along the lines of expectations fueled by the article: "the high-speed readout of the image sensor also means fast and accurate tracking of your subject no matter where it moves to within the frame. In fact, the system can be so accurate as to find and track not just a face, but the eyes within a face."
The examples with third party lenses don't exactly leave me with awe, I do appreciate the effort, though. There are too many shots with the nose in focus and focus jumps around too much (ears, nose, eyes + cheeks).
Thank you for the article and example videos.
Please repeat the D810 style close proximity head/eye AF test. The closer the target the more every inch of proximity change counts.
I also wonder how the D750 does in comparison, as its AF points are squeezed closer together for less total coverage, but also smaller gaps in between.
Is the sensor read at full resolution for Live View? Aka does the increased resolution of the EVF even get used outside of playback?
The Olympus E-M1 is one example where the higher EVF resolution is mostly only useful for playback compared to the lower resolution EVF of the E-M5 (mk1).
The faster processing of the stacked sensor may or may not allow for this. But it may also still be too slow to provide 60-120 fps at full resolution.
That's generally something I would to see tested, as EVF resolution numbers on their own don't tell much about the detail of the EVF Live View image. Furthermore it gives some clues about potential AF improvements.
Thanks for the video!
Not meant too seriously, but the advantage of the vibration control was specifically mentioned in one scene where the camera was put firmly on a bridge handrail and another scene where the camera was on a tripod. Just two cases of the video image not entirely fitting to the spoken words. ;)
Eleson: If optical steady shot means 'in-lens' then I don't understand 5-axis stabilization?
Electronic stabilization (aka cropping and rotating the final image) is possible, of course. But then either a larger area of the sensor needs to be utilized (possible, as the sensor offer 22 mp total) or a part of the image is cropped (as with many video implementations) and maybe even stretched back afterwards.
In any case this is inferior to sensor stabilization, because it is in essence a software solution. The drawback of sensor stabilization is that it needs more space inside the camera (see Olympus E-M5/1 sensor module size).
"making it the first truly 'pocketable' compact high-zoom camera with a built-in electronic viewfinder"
Sorry, but no. The TZ70 may come with 5 mm more height, but in return it comes with 1.6 mm less depth. And in this general height x width size category it's depth that defines how pocketable a camera is. I guess you can put both in a jeans pocket, but the LF1 still trumps them in size (only up to 200mm in return for a bigger sensor).
For comparison, the HX90 is 36 mm deep, the LF1 is 28 mm. And while you can squeeze a RX100 into a jeans pocket the LF1 does that far more comfortably because of its lesser depth.
Still a nice offering, if only Sony would include a raw file format, which the Panasonics do.
How can in-lens stabilization compensate for rotational movement? When you rotate a piece of glass the image does not rotate. When you rotate a sensor, on the other hand... ;)
Has AF-C performance been improved *considerably* since the pre-production firmware demonstrated at Photokina?