Jump to content

How is video processed on the SL2-S?


MITDelay

Recommended Posts

Advertisement (gone after registration)

Hello everyone, I just received my SL2-S and I am blown away by this camera in many respect. This is my first Leica so I am new on this forum.

However, I have a question on the video processing aspect of this camera (since it has also been marketed as a strong video camera). The sensor is 24.6MP but the 4K UHD video is 3840 x 2160 = 8.29MP; does it achieve this by cropping, pixel binning, or downscaling? I didn't notice a crop factor so its not that one, and the images look too good to be pixel binned, so is it downscaled? 

  • Like 1
Link to post
Share on other sites

19 hours ago, MITDelay said:

Hello everyone, I just received my SL2-S and I am blown away by this camera in many respect. This is my first Leica so I am new on this forum.

However, I have a question on the video processing aspect of this camera (since it has also been marketed as a strong video camera). The sensor is 24.6MP but the 4K UHD video is 3840 x 2160 = 8.29MP; does it achieve this by cropping, pixel binning, or downscaling? I didn't notice a crop factor so its not that one, and the images look too good to be pixel binned, so is it downscaled? 

Leica told early reviewers it should be oversampled on the SL2s in FF mode. Leica call it "pixel mixing" which is not very clear. One benefit of that is the excellent rolling shutter speed of that camera (14ms FF, below 10ms in s35).

For the SL2s, in s35, the pixel readout/display is nearly exactly the same so you can consider it a pixel-to-pixel mode. There Leica calls it "full scan"; which can mean either pixel-to-pixel or oversampling.

The denser SL2 sensor (47mp=8K) uses downsampling (with a mix of line-skipping, pixel bining? = some aliasing due to processing...) for its FF mode. and traditional oversampling in s35.
 

Both cameras lack an OLPF sensor so moire can be an issue on very fine patterns (fabrics, brick walls at a distance etc...). Both cameras seems to use some sort of dual gain iso circuitery or processing at 400ASA and 3200ASA, especially visible when shooting log. L-Log is largely denoise free, which can be alarming at first but looks really good when processed correctly in your NLE of choice.

The weird discrepancies between the two cameras are:

The SL2 can do 4K60p Full Frame (10bits internally also, soon with new FW) and a 5K 4:3 pixel-to-pixel (only in Log 420, weird but good looking) modes.

SL2s can only do 4K60 in APS-C, for some weird reason... and it performs naturally better in very low light. If leica was ever going to implement RAW (but I don't think they will), only the SL2s should -logically- get it, as the 8K flow from the SL2 sensor would be very hard to manage for the target users.... if at all possible to channel all that data through the camera's hdmi port.

Max data load I have seen come out of my SL2 (not s) was around 1700-1800mbps via the atomos in 4K60 422 10 bits. Curious to know how high it gets with SL2s

Edited by Slender
  • Like 3
Link to post
Share on other sites

Sincere gratitude on the explanations, Slender. If you don't mind, can you share me the links to these specs?

I'm pretty surprised that this is a dual gain (native) iso, because no where does it say that in the everyday marketing materials. The only hint at this feature comes from the fact that it shares the same architecture as the panasonic s1h, which has the dual gain. Before you mentioned this I have always wondered why Leica did not implement the dual native iso like the Panasonic. They should have marketed this more clearly as a feature for video users.  Thanks

Edited by MITDelay
Link to post
Share on other sites

To recap, the SL2-S (and SL before that) has two ways of recording 4K.

In "Super-35" (crop) mode it maps a 4K area in the centre of the sensor, and samples the pixels 1:1. This area is roughly 24mm wide, similar to 35mm motion picture film.

In "full frame" mode it uses the full width of the sensor (a little less for UHD) and over-samples it. This isn't "pixel binning", more like RGGB-to-RGB. Sony uses a similar algorithm in some of their cine cameras, as do others.

The 47MP SL2 uses oversampling all the time.

  • Like 1
Link to post
Share on other sites

Advertisement (gone after registration)

@BernardCHmmm.... I would rather say I am pretty sure the 47mp SL2 downsamples its FF to get the 14ms rolling shutter performance... and that's also what gives the weird processing-induced alliasing more pronounced in this mode... IMHO. Scanning all those Six-Thousand lines vertically and oversampling them would take more time and give us a lot of jello, in that respect the SL2 does better than the S1H, for example (about 22-24ms rolling shutter in FF for the later). For all the rest, obviously the S1H is a (much?) better video camera.

The 5K mode is true pixel-to-pixel, and do not suffer from any over/sub/sampling-induced alliasing at all.

@MITDelay you can find those in the manuals of respective cameras, cf pictures attached.
About marketing, Leica thinks the vast majority of its customer base will respond more to Steve McCurry, "decisive moments"...etc than obscure lingo like dual-gain video processing 🤓and they are quite on the spot for their marketing.

Sadly for video enthusiasts it means: no cook-book for the L-Log and limited interactions withother brands like Atomos, DJI, DaVinci....etc.

See here the Leica lingo for sensor area, resolution, scan, mixing....etc:

SL2

 

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!

 

SL2s

  • Thanks 3
Link to post
Share on other sites

12 hours ago, Slender said:

@BernardCHmmm.... I would rather say I am pretty sure the 47mp SL2 downsamples its FF to get the 14ms rolling shutter performance... and that's also what gives the weird processing-induced alliasing more pronounced in this mode... IMHO. Scanning all those Six-Thousand lines vertically and oversampling them would take more time and give us a lot of jello, in that respect the SL2 does better than the S1H, for example (about 22-24ms rolling shutter in FF for the later). For all the rest, obviously the S1H is a (much?) better video camera.

Thank you both for your replies. Even though the specs and features (i.e. ProresRaw on Ninja) for the S1H is better, but from what I am reading online, the contrast and color science is still much better straight out of the box from the sl2-s. Too bad I don't have a S1H to confirm. I was deciding between S1H and this one (and the A7S3, but 12MP for photos is just too little) becuase I use both video and stills  almost 50-50. Hopefully I made the right choice...

Edited by MITDelay
  • Like 1
  • Thanks 1
Link to post
Share on other sites

@MITDelayIf you shoot still, then yeah absolutely the SL2s takes the cake:). What lens do you have with it?

And yes colors are wonderful, and if you dont want the hassle of raw that is defitively the right choice IMHO :)

Edited by Slender
Link to post
Share on other sites

On 7/24/2021 at 5:24 PM, Slender said:

@MITDelayIf you shoot still, then yeah absolutely the SL2s takes the cake:). What lens do you have with it?

And yes colors are wonderful, and if you dont want the hassle of raw that is defitively the right choice IMHO :)

Slender, I'm currently using the Panasonic 24-70/2.8 and it is really good. I actually got 2 copies on this lens and took the one with better IQ, too bad it rattles ever so slightly when turned on. When I have the time I'm planning to go to a leica store to compare this lens with the summicron-sl 28, but planning to get the 21 or the 24 when it comes out, so I have time. I know it will be better but I have to decide whether it is worth the price better.

Frame-it  Great link. I definitely did not see that article when using google. 

Edited by MITDelay
Link to post
Share on other sites

On 7/26/2021 at 2:38 PM, hansvons said:

Niice one ahah :)

To be fair, Leica cameras can have actual (cool) names, but they are usually internal code names to cover upcoming developments.

The first monochom is "Henri" - for obvious reasons....

First Leica T was "Tyhoon" I believe.... it did invent the L mount after all.

SL2 is, if I recall correctly, " Vader" 😎 so I am sure you can call you blacked out SL2s that way lol

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...