Photokina 2018 Part 1: Panasonic S1 and S1R full frame 4K/60p and Fujifilm GFX 100MP 4K Medium Format. WOW!

[Aussie Ash]

2 hours ago, kaylee said:

No Louis Vuitton Edition?  PASS!

Lets hope they don't make an" Edward VII edition" the mind boggles

[wobba]

1 hour ago, Mokara said:

Depends on the processing power and the responsiveness of lenses. There is no inherent reason why PDAF is better than CDAF, actually if you have enough processing power the reverse would be true since you are dealing with what actually appears on the image rather than a parallax difference. Consequently it is more accurate, provided you are gathering enough data.

CDAF works better with a narrow depth of field, so putting it in a FF camera is not a disadvantage. PDAF works better with deeper depth of field, since it is essentially a crude rangefinder independent of what is in the image. This is why you will find both types in most MILCs, PDAF is used to get a rough range so the lens focus point is in about the right place, and then CDAF is used to get it set accurately. If your lens is responsive and fast enough however, you don't really need PDAF as much.

Yet in every example/comparison I have ever seen, PDAF has proven to be vastly superior to CDAF, particularly with respect to continuous AF video performance.  

[sanveer]

There seems to be something terribly curious about the Panasonic S Range.

Like someone pointed here, or somewhere else, Panasonic disclosed very few feature of the GH5, when it first announced its development. This appears to be the case with the S Series too. Except the Megapixel count and the other usual Panasonic features (IBIS, 4k, DFD etc), everything else seems to be suspect (including dynamic range, fps etc).

I guess we will have to wait about 6 months to get the entire feature set. On video alone, I am getting it will lead the FF wolf pack. Though the guys incharge of continuous autofocus and its feature set (DFD), should be thrown out. Right away, before he and his team do more harm. 

[PannySVHS]

Hmm, an announcement of an announcement. If it does not top GH5 video specs with 10bit plus higher bitrate in VFR, it would be a GH5s with a FF sensor and IBIS for 3000- 4000 USD. Guess, gotta wait a couple more weeks for the real announcement.

[kye]

4 hours ago, Mokara said:

Depends on the processing power and the responsiveness of lenses. There is no inherent reason why PDAF is better than CDAF, actually if you have enough processing power the reverse would be true since you are dealing with what actually appears on the image rather than a parallax difference. Consequently it is more accurate, provided you are gathering enough data.

CDAF works better with a narrow depth of field, so putting it in a FF camera is not a disadvantage. PDAF works better with deeper depth of field, since it is essentially a crude rangefinder independent of what is in the image. This is why you will find both types in most MILCs, PDAF is used to get a rough range so the lens focus point is in about the right place, and then CDAF is used to get it set accurately. If your lens is responsive and fast enough however, you don't really need PDAF as much.

You might be right about CDAF being best for fine-tuning the focus, but PDAF not only knows something is out of focus but which way the focus is to be found.

After watching dozens or hundreds of beautiful moments pass while the focus mechanism has charged off in the wrong direction and the moment concludes and the shot is lost while the camera is still wondering why the entire frame is a complete blur, I will always be deeply deeply skeptical of CDAF-only cameras.

[Mokara]

3 hours ago, wobba said:

Yet in every example/comparison I have ever seen, PDAF has proven to be vastly superior to CDAF, particularly with respect to continuous AF video performance.  

It is just a computational problem. Once cameras have sufficient power and lenses are fast and reactive enough, CDAF will be better. Ultimately it is more accurate and does not require calibration.

1 hour ago, kye said:

You might be right about CDAF being best for fine-tuning the focus, but PDAF not only knows something is out of focus but which way the focus is to be found.

After watching dozens or hundreds of beautiful moments pass while the focus mechanism has charged off in the wrong direction and the moment concludes and the shot is lost while the camera is still wondering why the entire frame is a complete blur, I will always be deeply deeply skeptical of CDAF-only cameras.

CDAF can determine which way to go to find focus as well, is the processor is powerful enough to process the data. Going forward as processors become faster and more powerful they will be able to handle larger data sets and performance will improve, especially when you have lenses that are designed with this in mind. Panasonic are already leading the field in this regard.

Cameras of the future will not be using the technology of the past. Behavior you saw in old cameras is not where technology is heading. You need to look where things are going, not where they have been. Who cares what old cameras were capable of, what matters is what new cameras are capable of.

[ajay]

I find this whole CDAF vs PDAF debate rather irritating. Tony Northrup yesterday already was trashing S1/S1R because of Panasonic's perseverance of using CDAF in the new FF system. Really? No one has had hands-on with a production model yet (and won't for some time) and already is claiming that it will have autofocus issues.

The problem is that everyone watches these reviewers on Youtube and repeat on the forums what is the perceived downfalls of a camera before the camera has ever been tested.

Obviously Panasonic has faith in their autofocus system. As others have stated, as faster processors come into play it will come to a point in which autofocus is so fast you won't be able to tell who is using what system to obtain focus. Who cares if it's CDAF or PDAF as long as it works?

 

[kye]

40 minutes ago, Mokara said:

It is just a computational problem. Once cameras have sufficient power and lenses are fast and reactive enough, CDAF will be better. Ultimately it is more accurate and does not require calibration.

CDAF can determine which way to go to find focus as well, is the processor is powerful enough to process the data. Going forward as processors become faster and more powerful they will be able to handle larger data sets and performance will improve, especially when you have lenses that are designed with this in mind. Panasonic are already leading the field in this regard.

Cameras of the future will not be using the technology of the past. Behavior you saw in old cameras is not where technology is heading. You need to look where things are going, not where they have been. Who cares what old cameras were capable of, what matters is what new cameras are capable of.

If what you're saying is true then that's a good thing, and I hope you're right.

My experience has been awful with older budget cameras, and the issues I've seen with the GH5 (or those that made the cut anyway) are admittedly less than 2005 point-and-shoot cameras, but unfortunately we're still too far off for it to be acceptable performance for me.

I've now filed it under the same category as 8K video - it will be good but it's not available yet and when it is available it will be costly and take time to trickle down into the camera body with the right feature set for my preferences.

[Robert Collins]

41 minutes ago, Mokara said:

Cameras of the future will not be using the technology of the past. Behavior you saw in old cameras is not where technology is heading. You need to look where things are going, not where they have been. Who cares what old cameras were capable of, what matters is what new cameras are capable of.

As I recall, the newest technology between pdaf sensor arrays, cdaf and on-sensor pdaf is on-sensor pdaf (starting with Nikon 1?) which is where most of the progress has been made over the last 5 years. The problem is that Panasonic has refused to adopt this technology and has disappeared down a long dark rabbit hole of DFD which is essentially marrying cdaf with 'proprietary' lens profiles....

[kye]

3 minutes ago, ajay said:

Who cares if it's CDAF or PDAF as long as it works?

No-one. The problem is that sometimes it doesn't.

[jonpais]

DFD sucks and Panasonic’s got cutting edge processors. Fail.

[DBounce]

What is more plausible? That the new Panasonic cameras will have world class AF, or that Panasonic will produce yet another series of cameras with the same poorly performing AF as all the other cameras they have made up to this point? 

My guess... if you want usable AF do not look to Panasonic.

No doubt the video specs will be great. Colors the same as the GH series. It will shoot 4K @ 24 and 60 FPS. It will record 10 bit 422 internal. Maybe compressed raw? Rolling shutter will not be great. There will be a cropped mode. 

Where is the organic sensor they spoke of a few years ago? I would buy it for that one feature. 

A mixed bag in the making me thinks.

[ajay]

Yes, let's trash the camera before it even is tested. Makes perfect sense.

 

[Emanuel]

1 hour ago, jonpais said:

DFD sucks and Panasonic’s got cutting edge processors. Fail.

Coming from an old brand user like you, this has its impact to the discussion. Speaks by itself.

 

5 minutes ago, ajay said:

Yes, let's trash the camera before it even is tested. Makes perfect sense.

 

I believe what people want to see here is increasing tech specs (being more efficient autofocus for video, one of them) to be adopted instead.

 

: -)

[eyesuncloudedphoto]

The most "hands-on" video on the S1/S1R up to this point, I think: 

https://www.youtube.com/watch?v=pyJrh0T5jeY

[Dan Sherman]

1 hour ago, jonpais said:

DFD sucks and Panasonic’s got cutting edge processors. Fail.

get up on the wrong side of the bed today did we.

DFD has the potential to be better than anything on the market given enough processing power. The issue is all the cry babies wanting other new features that eat up the processing power.  Since DFD involves knowing the lenses characteristics, given the processing power, it can actually be used to predict movement, and preempt the lost of focus. everything else can only react to a loss of focus. 

Hence the reason why DFD works amazingly well for stills, but less so for video. The video processing eats a crap ton of processing time.

[Emanuel]

36 minutes ago, Dan Sherman said:

DFD has the potential to be better than anything on the market given enough processing power. The issue is all the cry babies wanting other new features that eat up the processing power.  Since DFD involves knowing the lenses characteristics, given the processing power, it can actually be used to predict movement, and preempt the lost of focus. everything else can only react to a loss of focus. 

Hence the reason why DFD works amazingly well for stills, but less so for video. The video processing eats a crap ton of processing time.

Right. Reason why people here want a hybrid tool to focus on video rather than the opposite. Still hybrid. It's all about that. Seems they are simply neglecting that. Jannard's DSMC, after all, but for affordable range and indies friendly without breaking the bank. No more no less.

[webrunner5]

58 minutes ago, Dan Sherman said:

get up on the wrong side of the bed today did we.

DFD has the potential to be better than anything on the market given enough processing power. The issue is all the cry babies wanting other new features that eat up the processing power.  Since DFD involves knowing the lenses characteristics, given the processing power, it can actually be used to predict movement, and preempt the lost of focus. everything else can only react to a loss of focus. 

Hence the reason why DFD works amazingly well for stills, but less so for video. The video processing eats a crap ton of processing time.

Then why would a company that prides itself on the video side utilize a focusing system that sucks in video then? It is pretty obvious that Panasonic is pretty terrible at making a camera that can AF worth a crap in video mode.

Their stuff is becoming too expensive for the average person to buy, and that is a total change from what they used to be. I think it is going to end up badly for them down the road.

[jonpais]

We’ve only seen Panasonic make incremental steps in video AF, absolutely no reason to expect miracles with the S1. That’d be like hoping Canon would’ve had no crop on their FF mirrorless. It ain’t gonna happen. No cry baby, just being realistic. I’ve bought five Panasonics and still own three of them, and video AF has always been more or less poop. 

[Dan Sherman]

55 minutes ago, Emanuel said:

Right. Reason why people here want a hybrid tool to focus on video rather than the opposite. Still hybrid. It's all about that. Seems they are simply neglecting that. Jannard's DSMC, after all, but for affordable range and indies friendly without breaking the bank. No more no less.

AF in video is more demanding, so it will always lag behind stills. encoding/processing video takes up a lot of cpu power leaving less for the AF engine.

 

32 minutes ago, webrunner5 said:

Then why would a company that prides itself on the video side utilize a focusing system that sucks in video then? It is pretty obvious that Panasonic is pretty terrible at making a camera that can AF worth a crap in video mode.

I would't say its bad, its just not what a lot of people want. I mean some of the tests people do are retarded, jumping completely in and out of frame, or moving towards or away from the  camera quickly. Then in real world use, they just sit in front of it vlog style yet still complain because their extreme tests didn't work.