SL3 Rumors

[BernardC]

Advertisement (gone after registration)

57 minutes ago, padam said:

If those rumors are actually true (I am a little sceptical as to why Sony would provide a sensor that is much better than what's in the A7IV) then it has substantial improvements on the video front (faster readout, uncropped oversampled 4k60p, 4k120p).
If those are not true though, then it may only be a variation of the existing sensor with PDAF like S5II, which would be disappointing like the SL3.

Sony Semiconductor and Sony Consumer are different companies. There have been several cases where new sensors came-out on "competing" cameras (and smartphones, their primary market). That could be the case here, especially since Sony's own 60MP camera was updated a year ago. They wouldn't sit on a new sensor for 3 or 4 years.

It could also be that the sensor comes from a completely different fab, like the previous 47MP sensor.

It's encouraging that the rumoured video frame rates are different, that's a sign that it's not the same sensor.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Dr. G]

25 minutes ago, hoolyproductions said:

I hadn't spotted that extra dial, does look quite out of place. I do hope it is not a PASM (or as I like to call it "SPAM") dial - I use M mode 100% of the time and have developed a strong dislike for them (bit too strong sense of aesthetics 😁). Given that most of the other dials and buttons are (until now) unlabelled and configurable, hopefully it is neither a dedicated WB or SPAM dial.

I'm thinking it will probably be assignable.  I shoot aperture mode most of the time and use my back dial for both exposure compensation and adjusting ISO after pressing one of my customizable buttons.  It would be great to just have one dial ready to go.  If I had a separate dial for shutter speed, aperture and ISO always at the ready I may start using manual mode more.  

[LBJ2]

1 hour ago, Dr. G said:

There is one thing I'm concerned about.  With the M11 I'm assuming there is no PDAF matrix over the sensor.  With the Q3 it is there, but the wider angle of the lens may not make what I'm concerned about show.  On my Sony and Canon systems when some lenses were used there was a very strange flare pattern which showed up as a grid of square shaped flares on images.  After inquiring about it I was told that it was from the PDAF matrix over the sensor.  Hopefully this isn't the case on the SL3.

Yes. This is known potential of PDAF. Maybe one of the reasons Panasonic stayed away from including PDAF for so long. For all the years I've been shooting with PDAF I've only really seen the pattern if I intentionally induce or do a ISO/Noise push test. But never really run across it much on my normal shooting ( I do typically tend to avoid heavy backlighting for multiple reasons if possible) 

[ZHNL]

5 hours ago, Jon Warwick said:

I personally wasn't always a fan of the rendering off the SL2 (for my taste, the images with the APO Summicron would have too high acutance, or maybe more sharpening is baked into the DNGs even with the sharpening slider at zero, so it looked often too digitally sharp for what I prefer)

...whereas the M11 files has a better mix of recording a ton of fine detail, but does so with a greater gentleness. The images off the M11 look so "natural" and alive as a result.

..unlike the SL2, I find the SL2-S images remind me a lot more of what I like about the M11 rendering.

Having both, now back to SL2. I personally prefer SL2 color and file better by some margin. Mainly because when I care about IQ, it is at low ISO. 
 

the new SL3 so far didn’t bring anything significant to me. Triple resolution is a good feature I default use for m11 and m11m. A useful feature if you shoot a lot. Flip screen is nice to have feature. But e shutter read out speed (1/10s compare to 1/40 SL2) will be a huge downgrade to me to the point I won’t  consider this camera. 
 

I hope leica/panasonic did some magic to the Sony sensor on read speed though the chance is very slim. 

[BernardC]

12 minutes ago, ZHNL said:

But e shutter read out speed (1/10s compare to 1/40 SL2) will be a huge downgrade to me to the point I won’t  consider this camera. 

I haven't seen that number anywhere. Is that part of a rumour, or is it what the a7r does? I've read CineD's "rolling shutter" tests, and they are closer to 20ms (1/50th), but I suspect CineD are measuring something different.

[SrMi]

1 hour ago, ZHNL said:

Having both, now back to SL2. I personally prefer SL2 color and file better by some margin. Mainly because when I care about IQ, it is at low ISO. 
 

the new SL3 so far didn’t bring anything significant to me. Triple resolution is a good feature I default use for m11 and m11m. A useful feature if you shoot a lot. Flip screen is nice to have feature. But e shutter read out speed (1/10s compare to 1/40 SL2) will be a huge downgrade to me to the point I won’t  consider this camera. 
 

I hope leica/panasonic did some magic to the Sony sensor on read speed though the chance is very slim. 

I assume SL2 has the same readout speed as Q2. Q2's readout speed is 1/26 sec (DPR link).

The 60MP sensor used in Sony cameras needs 1/10 sec (Jim Kasson link).

 

[ZHNL]

Advertisement (gone after registration)

I have a test done myself using US AC freq. I measured (roughly) about 1/30 for SL2S and 1/36-1/40 SL2. Not sure about Q2 and why it is even slower than SL2S. 

[ZHNL]

1 hour ago, BernardC said:

I haven't seen that number anywhere. Is that part of a rumour, or is it what the a7r does? I've read CineD's "rolling shutter" tests, and they are closer to 20ms (1/50th), but I suspect CineD are measuring something different.

I am not following video spec but if it is not using full pixel by skip row, in theory, read out speed can be fast depend on output resolution. For photography, M11 and m11m I have it is around 1/10, I still use it a lot during landscape shot but just not to the point that I can use it without care. SL2 is way better in this regard. 

Edited February 15, 2024 by ZHNL

[BernardC]

OK, so it's a different test from rolling shutter. I wonder if there's a correlation though. Has Kasson tested any real cine cameras like the Alexa or Venice? Those end-up in the 5 millisecond range, with 15 stops of DR, compared to mirrorless cameras in the 20ms range with 12 stops of DR. Those cameras use a lot more power, and have lots of active cooling, but it shows how far we could progress.

[SrMi]

1 hour ago, ZHNL said:

I have a test done myself using US AC freq. I measured (roughly) about 1/30 for SL2S and 1/36-1/40 SL2. Not sure about Q2 and why it is even slower than SL2S. 

Higher-resolution sensors typically have slower readouts than lower-resolution sensors (47MP vs 24MP).

P.S.: I also measured about 1/40 sec for SL2 using the quick-and-dirty method.

Edited February 15, 2024 by SrMi

[SrMi]

35 minutes ago, BernardC said:

OK, so it's a different test from rolling shutter. I wonder if there's a correlation though. Has Kasson tested any real cine cameras like the Alexa or Venice? Those end-up in the 5 millisecond range, with 15 stops of DR, compared to mirrorless cameras in the 20ms range with 12 stops of DR. Those cameras use a lot more power, and have lots of active cooling, but it shows how far we could progress.

Jim has tested only readout time for still photography. Some mirrorless cameras with stacked sensors are now in the less than 5ms range.

[BernardC]

27 minutes ago, SrMi said:

Some mirrorless cameras with stacked sensors are now in the less than 5ms range.

That's weird, since they are twice or three times that value in cine mode. I assume you mean the R3, Z9, etc? The A7s III is better (8.7ms), but it uses a 12MP non-stacked sensor. At least I think it's a non-stacked sensor, or else they would mention it.

[matted]

1 hour ago, BernardC said:

OK, so it's a different test from rolling shutter. I wonder if there's a correlation though. Has Kasson tested any real cine cameras like the Alexa or Venice? Those end-up in the 5 millisecond range, with 15 stops of DR, compared to mirrorless cameras in the 20ms range with 12 stops of DR. Those cameras use a lot more power, and have lots of active cooling, but it shows how far we could progress.

The Alexa LF sensor is also only ~13MP… I would certainly be ok with that for stills if it had a global shutter but many would not be!

[Hendo]

8 hours ago, BernardC said:

Sony started-out with Minolta lenses. Their new generation of "GM" lenses is much more competitive, as they upgrade legacy designs. My point was that Leica started with a clean sheet of paper (clean CAD?), they've used a more modern approach from the start with the 24-90 and 90-280.

Could you explain what the more modern appraoch involves? I do not understand..

Edited February 15, 2024 by Hendo

[jaapv]

1 hour ago, Hendo said:

Could you explain what the more modern appraoch involves? I do not understand..

Especially as one of Leica's strengths is their unmatched archive of lens design.

[tritentrue]

When I questioned a local Leica Store employee about the SL3 earlier this week, she disclosed only that it's  "smaller" and "a different form than the other SL cameras."

[SrMi]

4 hours ago, BernardC said:

That's weird, since they are twice or three times that value in cine mode. I assume you mean the R3, Z9, etc? The A7s III is better (8.7ms), but it uses a 12MP non-stacked sensor. At least I think it's a non-stacked sensor, or else they would mention it.

We should not mix video and still modes, and in this thread, we are talking about still photography. 

Still photography, full sensor readout: Z9 has a 3.7ms readout speed. a7sII should have a bit more than 10ms.

[BernardC]

2 hours ago, Hendo said:

Could you explain what the more modern appraoch involves? I do not understand..

Very light focusing groups that can be moved very quickly. 

Simpler lenses focus as a whole, meaning that all of the elements are on a helicoid and move back and forth. The next evolution is internal focus where one group of lenses moves. Note that this is different from close focus compensation where one group moves a different amount from the others, as used in several M lenses. 

In order to get really fast AF, you need to minimize the mass of the ficusing group (or groups), as well as the distance that they need to be moved. 

It's an evolution. The SL lenses started with optimized focusing groups, but many legacy designs have much more mass to move around. 

As with everything in optics, it's all a compromise. M lenses are at one end of the scale with fewer elements and more moving mass, and SL lenses have more elements but less moving mass. 

[Advertisement]

Hallo,

Look here: Leica L-Mount

[la1402]

11 hours ago, BernardC said:

DxO rates the M11 sensor higher than any other full-frame 35mm sensor. The a7R5 ranks below the S1r (same sensor as the SL2). They test in a calibrated lab environment, unlike PtP.

Both get a score of 100, with a very tiny advantage of the M11 only in ISO performance. The lack of IBIS will remove that micro advantage in real life. 

[Planetwide]

9 hours ago, LBJ2 said:

Yes. This is known potential of PDAF. Maybe one of the reasons Panasonic stayed away from including PDAF for so long. For all the years I've been shooting with PDAF I've only really seen the pattern if I intentionally induce or do a ISO/Noise push test. But never really run across it much on my normal shooting ( I do typically tend to avoid heavy backlighting for multiple reasons if possible) 

I have run into this grid pattern on my R5, when doing extreme shadow pulls and aggressive noise reduction. My R3 does not display this issue FYI. There is definitely more pull available from a Leica SL2/S1r. This makes sense because the SL2/S1r and R3 are all BSI sensors, whereas the R5 is not. 

Edited February 16, 2024 by Planetwide