SL2-S

[bilbrown]

Advertisement (gone after registration)

9 hours ago, Slender said:

4K is hard enough on most machine without a prores workflow... and who can really enjoy 6K content anyway? Perhaps they keep this as a suprise hattrick for later with raw output... or for a more expensive SL2 SP (or C ine) version 😜

This would get rid of all thing allialing and moire, by the way. I really would like to see some footage of fine detailed object shot in FF vs APS-s (pixel mix vs full pixel readout0 on this camera

6K can edit pretty well on the modern OS systems and has been optimized. You can actually edit 8K or 12K (BM Ursa Mini) very easily... in RAW. The main reason for 6K+ is downresing with finer noise and crop-zoom. It's useful. I have had a RED Scarlet W and BMPCC 6K and its not horrible to edit. The SL2-(whatever) will need to have ProRes, ProRes RAW or R3D/BRAW to really work well. CDNG is an option, and there are programs that cut the file size of CDNG for storage, and of course DaVinci Resolve imports CDNG as a clip – FCP X and Adobe have a work around. We have been using the compressed CDNG software with the Sigma FP and it works well. Here it is.

 

https://www.slimraw.com

 

 

[Slender]

Yep, thanks for the link. I was on sets using 8K RAW (sometimes in 3D...) back in 2013 already -extensively with FS65 back then-. It has come a long way since, but most content distribution (I am talking 24 meters wide screens) still happens mostly in 2K or so. The more from the get go the better, sure... but sometimes the most efficient productions I have seen skip raw entirely and rely on exposing correctly in the first place with log and ProRes in camera... (Arri boasts a lot about this).

The Leica technician rep from New Jersey (that is also where ARRI is based in the US...) from the Leica America presentation of the SL2s seemed quite blase when Atomos and raw recording were mentioned by a question about an upcoming update. "we are big on no compromises... its all about the essential, not adding stuff etc..."

I believe their idea of the average Leica video enthusiast (didnt dare say professional user...) will mostly want to keep using the camera as it is: a great still platform with a quick switch towards very commendable, no add-on needed, video capabilities. It is somehow bitter their many "stories" about the SL2 almost never mentioned this. Now they have this very good infomercial with Dominic Nahr at the Bernina Hill Climb Race (sorry imentioned it many times elsewhere already...) raving all bout internal recording on SD cards. And from the look of it in 4K in vimeo, this level of quality is already very impressive and certainly good enough for 97% of potential SL2s buyers... (I do not mention anything AF side, cause I dont care and we all know the competition does it better for video).

So sadly, unless there is a huge surge of market demand for it, I dont see the Atomos integration getting better for Leica cameras soon... and neither do I hope for LockCircle to do more reasonably priced video accessories for Leica SL... because they have no competition against them b/c not a popular platform... I secretly want it all tho 🤪 (Lockcircle cage, atomos and Ronin RS2 compatibility... and ProRes in camera)

Edited December 28, 2020 by Slender

[bilbrown]

2 hours ago, Slender said:

The Leica technician rep from New Jersey (that is also where ARRI is based in the US...) from the Leica America presentation of the SL2s seemed quite blase when Atomos and raw recording were mentioned by a question about an upcoming update. "we are big on no compromises... its all about the essential, not adding stuff etc..."

Yeah, I'm the "Bil Brown" friend of Leica that asked that question...

 

I'm just looking at the Sigma FP and Panny S1/S1H with the same sensor and thinking if Leica could fix heat sink without fans they could really work with this - maybe in camera like the FP.

[Radost]

Dont need prores raw. I need prores internal recording . Better than mp4.

[Photoworks]

So much extra work Editing ProResRAW , noise reduction, lens distortion, data size.

Most production are still made in ProRes ARRI Alexa 3.2k and up-res them after post.

Just use a Video dedicated camera for that. 

[bilbrown]

8 hours ago, Photoworks said:

So much extra work Editing ProResRAW , noise reduction, lens distortion, data size.

Most production are still made in ProRes ARRI Alexa 3.2k and up-res them after post.

Just use a Video dedicated camera for that. 

Well... that's not completely true:

Yes, if the camera is an Arri Alexa you will see 3.4k or 2.8k ArriRaw or ProRes 4444 2.8k, or if it is a RED you will likely see 2:1 to 5:1 RED Monstro, Gemini or Dragon 6k or 8k cropped/downresed, or if 35mm Kodak Vision3 250D or 500T (same film that is used by Cinestill for stills), but there is a lot of Open Gate Arri 65 and RED Vista Vision work coming down the pipe in features as well as broadcast because of the production standard folks like Disney and HBOMAX are pushing for "cinematic" work and have even begun projecting UnReal Engine footage onto screens instead of green screen or locations. The deliverables are still currently 4K, but with all the SPFX in post, 5k-12k is becoming more or less expected if not standard.

The Netflix camera standard for what they pay to produce is a good general guide (although if you make it then sell it to them it doesn't matter, only for their own productions. And yes, I know we aren't likely going to use our mirrorless Leica to shoot a major film or streaming TV show. There are a lot of other considerations, like bitrates etc, that we only get into if we do it.

My point is: if the camera CAN have it, it SHOULD have it – whether we want to use it or not. It's not like it is an inelegant choice when other L mount cameras do RAW internally (like the Sigma FP) or have 6K available.

If Leica can work-out the heat sink without a fan they can figure out how to have a few bells and whistles that will make the camera an even stronger contender for the choice for L Mount video. The Lumix S1H, S1 or S5 just shouldn't be the end all for pro features, and the Sigma FP shouldn't be the ONLY one of the bunch to do much of its (ALL-I CDNG) processing IN camera.

 

Edited December 29, 2020 by bilbrown
clarify

[lx1713]

Advertisement (gone after registration)

 

1 hour ago, bilbrown said:

 

My point is: if the camera CAN have it, it SHOULD have it – whether we want to use it or not. It's not like it is an inelegant choice when other L mount cameras do RAW internally (like the Sigma FP) or have 6K available.

 

Generally, the market decides this.

Features that are fundamental to the market it serves get priority. Commoditised to mass production but having too many features are a feature of consumer tools.

I don't think Leica is in a hurry to define the next generation when they don't really need to. The current SL bodies are general tools serving many.

The moment I see the SL series having special limited editions that are not serving the pro market, things like croc skin, gold finishes, etc. I know they have lost the plot.

So specialised pro items to check off like 8K, global shutters, wireless follow focus, etc are, of course, a bridge too far with the SL system. But it may well be an engineering bridge too far to add internal RAW and 6K. That's Leica's internal decisions. We all benefit when nice, unexpected stuff happens with the SL ecosystem 😉

The SL2-S is, to me, a very logical step. No surprises. The Sigma FP, Lumix S1H are niche players in their own ways. Very lovely idea when you compare with Canon and Nikon and possibly Sony. More choices in the L-mount alliance.

Comparatively speaking, "but with all the SPFX in post, 5k-12k is becoming more or less expected if not standard", that's not the reality in mirrorless market where the SL competes. That's neither consumer expectations & standard nor the average professional stills photographer's standards or even the vast majority of small team videographers.

[BernardC]

17 hours ago, Radost said:

Dont need prores raw. I need prores internal recording . Better than mp4.

MP4 is a container, not a CODEC. H.264 is Leica's chosen CODEC for MP4, for now, although they promise an HEVC CODEC this Spring. I understand what you mean by "ProRes is better than mp4," but it's not a logical statement. They aren't instances of the same thing. One's a container, one's a CODEC. ProRes comes in many quality flavours, as does mp4. ProRes is better for editing, but mp4 is better for distribution. In our case, the SL variants use mp4 exclusively for 8-bit 4:2:0 h.264 files, so ProRes is "better" (unless you want something that you can show directly on a big-screen TV). That's a bit misleading though, since Leica uses MOV (not mp4) for their higher-quality video files.

MOV is also a container. It can contain h.264 or ProRes (not at the same time), or a number of other CODECS. Given that MOV (QuickTime) is an Apple container, it's the default container for ProRes.

 

ALL-I h.264 with a high bit rate should be indistinguishable from ProRes, in terms of image quality. They are both lightly-compressed 10-bit 4:2:2 formats. If you want, you can convert your Leica h.264 MOV to ProRes, using any number of tools. You can also tell your NLE to create ProRes-like* proxies automatically.

Others here have reported that the SL2 MOVs are very NLE-friendly, causing no slowdown or hesitation on the timeline. Perhaps you will find the same thing, it really depends on your hardware and software.

 

* They will be ProRes in FCP, but probably not in Avid or Premiere or Resolve.

 

[bilbrown]

5 hours ago, BernardC said:

ALL-I h.264 with a high bit rate should be indistinguishable from ProRes, in terms of image quality. They are both lightly-compressed 10-bit 4:2:2 formats. If you want, you can convert your Leica h.264 MOV to ProRes, using any number of tools. You can also tell your NLE to create ProRes-like* proxies automatically.

Others here have reported that the SL2 MOVs are very NLE-friendly, causing no slowdown or hesitation on the timeline. Perhaps you will find the same thing, it really depends on your hardware and software.

 

* They will be ProRes in FCP, but probably not in Avid or Premiere or Resolve.

 

I have yet to really dig into my video features, and comparison with my SL2-S, for the L Mount variants – for ease I will put their basic specs here:

 

1. Sigma FP does CinemaDNG（8bit / 10bit/ 12bit） / MOV：H.264 (ALL-I/ GOP) 12bit is specifically CDNG and is large but pretty glorious
2. Lumix S1 has 4K HLG (Hybrid Log Gamma) internal video recording, HEVC codec in 10-bit 72Mbps. 4K 50/60p, 4:2:0 8-bit, cropped to APS-C, internal recording limited to 29:59 min. 4K 24/25/30p Full Frame, 4:2:0 8-bit, unlimited recording.
3. I don't have the S1H, so I won't comment much the article here says most of it: https://www.dpreview.com/reviews/panasonic-lumix-dc-s1h-review/5

 

As the original SL typ 601, the technical info of the SL2-S given by Leica is an EXTENSIVE list (maybe too extensive). In practice, as with the SL 1 I only used a few of the containers but it seems consistent with what everyone is saying here. https://us.leica-camera.com/Photography/Leica-SL/Leica-SL2-S/Technical-Insights

 

I dunno, guess I should just shoot something and share it against the cinema cameras at my disposal: The RED Scarlet W 5k, and BMPCC 6k. That is if anyone is interested...

 

One last thing, Steve Yedlin the DP on Knives Out (Rian Johnson Dir.) said this that I thought was interesting on his choice of shooting 2.5K (ARRIRAW 2880×1620) for the film

"Prevalent superstitions notwithstanding, we are truly post-format these days, meaning our primary leverage point for designing a photographic look (the rendering of tones and colors) is the color pipeline, not the selection of camera. So, with the artistry of visual authorship decoupled from the camera format, selection of that format becomes technical/operational rather than artistic: so I chose the Alexa Mini for its excellent colorimetric reliability, its low noise, its sensor size that could be covered by the lenses I wanted to use, its ability to go up to 200 frames per second without having to change our framing area (i.e. “window in”), and the fact that I have a color pipeline already designed for it and didn’t need to start from scratch. Since I know I’ll sculpt the color/tonal rendition to my vision no matter which camera I select, I’m free to choose the camera type that is the pragmatic best choice for getting our shots."

 

 

So, in that, the thought I have had all along is true: whatever works... 

 

Edited December 29, 2020 by bilbrown
added info

[Svetozar]

10 minutes ago, bilbrown said:

I have yet to really dig into my video features, and comparison with my SL2-S, for the L Mount variants – for ease I will put their basic specs here:

 

1. Sigma FP does CinemaDNG（8bit / 10bit/ 12bit） / MOV：H.264 (ALL-I/ GOP) 12bit is specifically CDNG and is large but pretty glorious
2. Lumix S1 has 4K HLG (Hybrid Log Gamma) internal video recording, HEVC codec in 10-bit 72Mbps. 4K 50/60p, 4:2:0 8-bit, cropped to APS-C, internal recording limited to 29:59 min. 4K 24/25/30p Full Frame, 4:2:0 8-bit, unlimited recording.
3. I don't have the S1H, so I won't comment much the article here says most of it: https://www.dpreview.com/reviews/panasonic-lumix-dc-s1h-review/5

 

As the original SL typ 601, the technical info of the SL2-S given by Leica is an EXTENSIVE list (maybe too extensive). In practice, as with the SL 1 I only used a few of the containers but it seems consistent with what everyone is saying here. https://us.leica-camera.com/Photography/Leica-SL/Leica-SL2-S/Technical-Insights

 

I dunno, guess I should just shoot something and share it against the cinema cameras at my disposal: The RED Scarlet W 5k, and BMPCC 6k. That is if anyone is interested...

 

That would be very interesting to see. You have my vote ! 

[bilbrown]

BTW, pretty likely although the Lumix S1 isnt the same sensor (it isnt BSI) the Sigma FP is... maybe we should be looking at how Sigma is making the sensor work for them.

 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Chaemono]

vor 6 Stunden schrieb bilbrown:

BTW, pretty likely although the Lumix S1 isnt the same sensor (it isnt BSI)

The S1 sensor is BSI. DxO:

https://www.dxomark.com/panasonic-lumix-s1-sensor-review/

For its first foray into full-frame cameras, Panasonic has selected a CMOS sensor with the same 24MP resolution as that is used by both the Sony A7 III and the Nikon Z 6. With just one point between them in the DxOMark rankings and similar Portrait, Landscape, and Sports scores, the new Lumix DC-S1 sensor performs practically identically to its most direct competitors.

...

With the new 24MP Lumix DC-S1, Panasonic has adopted a class-leading sensor for its venture into full-frame. The test results are practically identical to those of the sensor used in the highly-acclaimed Sony A7 III, and in some areas, such as dynamic range, the S1 is able to offer an advantage in the highly desirable ISO 800 to ISO 1600 range, not just at base ISO 100.

It achieves all this while maintaining low noise levels (as we’ve also seen with the Nikon Z6 and the Sony), which demonstrates that 24MP full-frame CMOS BSI-type sensors offer a desirable balance between pixel count and sensor dynamics.

[scott kirkpatrick]

Yes, according to Bill Claff, there is a strong similarity between the S1 and the SL2-S's sensors.  However, he uses open data points for the highest ISOs on the S1, suggesting that things are getting a little forced.  I'd like to see a comparison between the two cameras when they stretch above ISO 25000.  The Sl2-S does surprisingly well there.

I don't have an S1, but Sean Reid is likely to get one and maybe also an S-5 for an upcoming article, comparing them to the SL2-S.

[Radost]

16 hours ago, BernardC said:

MP4 is a container, not a CODEC. H.264 is Leica's chosen CODEC for MP4, for now, although they promise an HEVC CODEC this Spring. I understand what you mean by "ProRes is better than mp4," but it's not a logical statement. They aren't instances of the same thing. One's a container, one's a CODEC. ProRes comes in many quality flavours, as does mp4. ProRes is better for editing, but mp4 is better for distribution. In our case, the SL variants use mp4 exclusively for 8-bit 4:2:0 h.264 files, so ProRes is "better" (unless you want something that you can show directly on a big-screen TV). That's a bit misleading though, since Leica uses MOV (not mp4) for their higher-quality video files.

MOV is also a container. It can contain h.264 or ProRes (not at the same time), or a number of other CODECS. Given that MOV (QuickTime) is an Apple container, it's the default container for ProRes.

 

ALL-I h.264 with a high bit rate should be indistinguishable from ProRes, in terms of image quality. They are both lightly-compressed 10-bit 4:2:2 formats. If you want, you can convert your Leica h.264 MOV to ProRes, using any number of tools. You can also tell your NLE to create ProRes-like* proxies automatically.

Others here have reported that the SL2 MOVs are very NLE-friendly, causing no slowdown or hesitation on the timeline. Perhaps you will find the same thing, it really depends on your hardware and software.

 

* They will be ProRes in FCP, but probably not in Avid or Premiere or Resolve.

 

 

Prores is a codec. Based on Mpeg2 while compensating with higher bitrate. 

Mp4 contains h264.  You know what i mean when I say mp4. 

 

First converting files as an extra steps or working with proxies is not something that i want.

Second h264 being easier for distribution is not a good argument. I am not sending out of camera files and can export to whatever I want. 

 

I have done a quick resiolve  Arri mini prores edit on an I5 surface with 8gb of ram. H264 is a different beast.

ProRes playes like butter. H264/h265 not even close.

 

Edited December 30, 2020 by Radost

[BernardC]

4 hours ago, Radost said:

Mp4 contains h264.  You know what i mean when I say mp4.

Yes, I understood what you meant, from context. You meant low-bit-rate, 4:2:0, long-GOP, h.264 mp4, as used in the SL2S. As I explained, mp4 is just the container, and it can also contain h.265 or MPEG (which is similar to the compression scheme used by ProRes).

What I thought was confusing is that you compared long-GOP 8-bit "mp4" to ProRes, while ignoring the 10-bit ALL-I "MOV" option offered by the SL2-S.

To my mind, that's the more valid comparison. We both know that you shouldn't shoot 8-bit long-GOP if you intend to do extensive post-production (if you can help it). However, that's not the option that Leica provides for this scenario, so why dwell on it?

ProRes isn't the only editor-friendly format. Arguably, Avid's DNxHR is more popular at the high-end, at least in Hollywood and television. Reports here and elsewhere mention that Leica's ALL-I implementation is very smooth in Resolve Studio (the CODEC isn't licensed in the free version). Have you tried it? The bit rates are very close to ProRes 422, so I wouldn't be at all surprised if it performed the similarly, especially if you have a fairly decent GPU.

[Massimo D’Anolfi]

18 hours ago, bilbrown said:

I have yet to really dig into my video features, and comparison with my SL2-S, for the L Mount variants – for ease I will put their basic specs here:

 

1. Sigma FP does CinemaDNG（8bit / 10bit/ 12bit） / MOV：H.264 (ALL-I/ GOP) 12bit is specifically CDNG and is large but pretty glorious
2. Lumix S1 has 4K HLG (Hybrid Log Gamma) internal video recording, HEVC codec in 10-bit 72Mbps. 4K 50/60p, 4:2:0 8-bit, cropped to APS-C, internal recording limited to 29:59 min. 4K 24/25/30p Full Frame, 4:2:0 8-bit, unlimited recording.
3. I don't have the S1H, so I won't comment much the article here says most of it: https://www.dpreview.com/reviews/panasonic-lumix-dc-s1h-review/5

 

As the original SL typ 601, the technical info of the SL2-S given by Leica is an EXTENSIVE list (maybe too extensive). In practice, as with the SL 1 I only used a few of the containers but it seems consistent with what everyone is saying here. https://us.leica-camera.com/Photography/Leica-SL/Leica-SL2-S/Technical-Insights

 

I dunno, guess I should just shoot something and share it against the cinema cameras at my disposal: The RED Scarlet W 5k, and BMPCC 6k. That is if anyone is interested...

 

One last thing, Steve Yedlin the DP on Knives Out (Rian Johnson Dir.) said this that I thought was interesting on his choice of shooting 2.5K (ARRIRAW 2880×1620) for the film

"Prevalent superstitions notwithstanding, we are truly post-format these days, meaning our primary leverage point for designing a photographic look (the rendering of tones and colors) is the color pipeline, not the selection of camera. So, with the artistry of visual authorship decoupled from the camera format, selection of that format becomes technical/operational rather than artistic: so I chose the Alexa Mini for its excellent colorimetric reliability, its low noise, its sensor size that could be covered by the lenses I wanted to use, its ability to go up to 200 frames per second without having to change our framing area (i.e. “window in”), and the fact that I have a color pipeline already designed for it and didn’t need to start from scratch. Since I know I’ll sculpt the color/tonal rendition to my vision no matter which camera I select, I’m free to choose the camera type that is the pragmatic best choice for getting our shots."

 

 

So, in that, the thought I have had all along is true: whatever works... 

 

I dunno, guess I should just shoot something and share it against the cinema cameras at my disposal: The RED Scarlet W 5k, and BMPCC 6k. That is if anyone is interested...

thanks a lot

[ruskkyle]

24 minutes ago, Massimo D’Anolfi said:

I dunno, guess I should just shoot something and share it against the cinema cameras at my disposal: The RED Scarlet W 5k, and BMPCC 6k. That is if anyone is interested...

thanks a lot

I’d be very interested 👍

[lx1713]

19 hours ago, bilbrown said:

"Prevalent superstitions notwithstanding, we are truly post-format these days, meaning our primary leverage point for designing a photographic look (the rendering of tones and colors) is the color pipeline, not the selection of camera. So, with the artistry of visual authorship decoupled from the camera format, selection of that format becomes technical/operational rather than artistic: so I chose the Alexa Mini for its excellent colorimetric reliability, its low noise, its sensor size that could be covered by the lenses I wanted to use, its ability to go up to 200 frames per second without having to change our framing area (i.e. “window in”), and the fact that I have a color pipeline already designed for it and didn’t need to start from scratch. Since I know I’ll sculpt the color/tonal rendition to my vision no matter which camera I select, I’m free to choose the camera type that is the pragmatic best choice for getting our shots."

I'm a stills photographer who shoots RAW 100% of the time but I can't perfectly match colours even from cameras that are different camera but same models with same lenses. So I am guessing that what he meant is the artistic vision is very much doable now without the angst over which camera is better. Else the argument seems odd that he had particular reasons for picking an Alexa Mini and not any camera will do.

Thanks, I enjoyed and appreciate the point of view brought up. I'd be interested to see how the Red Scarlet compares.

[Slender]

Can any, happy owner of an SL2s with an interest in video confirm, or test if the major SL2 video drawback is gone or at least better corrected: alliasing?

Whenever the scene gets complex (buildings at a distance, small foilage....etc), regardless of Atomos or internal, APS-C or FF, it seems aliasing with pixel mix is here.

I want to try with some diffusion filter see if I can go around this a bit...

I suspect the issue should be better of absent in the closest thing to 1:1 pixel readout we get in APS-C with SL2s....

I am still torn whereas trading "down" from SL2... the output in daylight with 47mp is quite astonishing and dont feel like going back to 24mp for stills, even with better iso.... but the video thing could be another stone in the balance....

Many thanks to whoever can try this out for the rest of us!
@Steven ?

[Advertisement]

Hallo,

Look here: Leica L-Mount

[Slender]

2 hours ago, Steven said:

would you like me to do a video test and post a link for the aliasing? If yes, give as much details as possible for what you want. Daylight ? Nighttime ? Shutter speed, iso, caméra movement, codec, frame rate, etc ....! 

Thank you @Steven This genre of scenes (even small foiliage... or parrot feathers even with SL2s I am fraid...) is exciting alliasing very easily.... If you could shoot something similar, with a sharp lens, on your SL2s... in FF and APS-C compared, that would be a key factor in my decision...

The mode I love the most is the very capable MOV. in C4K DCI (the 4096 one, not 3840...) at 24P, ALL-I (400MB/S). FF vs APS-C is the key point = Oversampling (6K -> 4K) VS Full Scan. Gammut set to L-Log with -2 sharpening.

In theroy, the APS-C should not exibit alliasing b/c no pixel mixing.... Moire is another matter, even Leica raised it in Film and Digital times.... but that is life without OLPF + sharp lenses in video anyway.

Thank you so very much if you find the time. 😇

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!