FS7: New Sony F7 XDCAM 4K Super 35mm Camcorder

[starcentral]

Here's a graphical view of the data levels you get per stop with slog2 and slog3 when recording 10-bit:

Image taken from The Sony Ultimate Exposure Guide - cinetiny.com/exposeguide

With slog3 you do see more data is allocated to the mid and lower end which is taken from the highlights end of the scale.

It seems like a conundrum, but is it? Where would you rather have more data available for your image if its limited to 10-bit bit depth knowing it will need to go through post for de-logging & grading?

Here's something to consider. slog2/3 are 1300% dynamic range curves. It means that it can capture brightness values up to 14.5 times brighter than 90% white. That is pretty good, its about twice the amount of the original slog, and if you expose "properly" you will put about 200 levels of data to your top 2.5 stops above 90% with slog2 versus 275 levels of data for slog3. Is this difference really going to matter?

 

[Gene Sung]

Here's a graphical view of the data levels you get per stop with slog2 and slog3 when recording 10-bit:

Image taken from The Sony Ultimate Exposure Guide - cinetiny.com/exposeguide

With slog3 you do see more data is allocated to the mid and lower end which is taken from the highlights end of the scale.

It seems like a conundrum, but is it? Where would you rather have more data available for your image if its limited to 10-bit bit depth knowing it will need to go through post for de-logging & grading?

Here's something to consider. slog2/3 are 1300% dynamic range curves. It means that it can capture brightness values up to 14.5 times brighter than 90% white. That is pretty good, its about twice the amount of the original slog, and if you expose "properly" you will put about 200 levels of data to your top 2.5 stops above 90% with slog2 versus 275 levels of data for slog3. Is this difference really going to matter?

I have a few questions about S-Log 2 vs S-Log 3

1) Doug Jensen mentioned that Slog 3 clips at 95 IRE while Slog 2 clips at 110 IRE. On paper that seems huge, but is that a huge difference in highlights in a REAL WORLD sense?

2) On the F5 (and presumably the FS7), does Slog 2 and Slog 3 both have 14 stops of DR? It’s just they place the DR in different areas of the curve, correct?

3) Since Slog 3 has more shadow detail, is it better to use it in lowlight? And Slog 2 in bright daytime since it’s focused on the highlights? Do you guys switch between the two?

4) Those of you on an F5 / F55, which Slog do you guys use more? 2 or 3?

Edit: Sorry Doug, I meant the 95 IRE part that you mentioned in another thread. The 110 IRE on S-Log was my own miss-information. I should not have put them in the same sentence.

 

[Alpha Beta Omega]

Well I never miss any shadow details with S-log so I certainly don't need more but to get more stops allocated to the higher end makes a huge difference.

 

[Grug]

I have a few questions about S-Log 2 vs S-Log 3

1) Doug Jensen mentioned that Slog 3 clips at 95 IRE while Slog 2 clips at 110 IRE. On paper that seems huge, but is that a huge difference in highlights in a REAL WORLD sense?

2) On the F5 (and presumably the FS7), does Slog 2 and Slog 3 both have 14 stops of DR? It’s just they place the DR in different areas of the curve, correct?

3) Since Slog 3 has more shadow detail, is it better to use it in lowlight? And Slog 2 in bright daytime since it’s focused on the highlights? Do you guys switch between the two?

4) Those of you on an F5 / F55, which Slog do you guys use more? 2 or 3?

1) On the F5 (and I believe the F55 as well) SLOG3 clips at 100 IRE (not 95 IRE), SLOG2 clips at 106 IRE. Both hold 6 stops of overexposure at the cameras' native ISO, but as you can see from the chart Dennis posted, they have different strengths - SLOG2 contains significantly more detail in the highlights, where SLOG3 contains significantly more detail in the shadows. I have to deal with the exceptionally harsh Australian sun all the time, and much prefer SLOG2 for it's superior highlight range and detail.

2) Both have 14 stops, distributed exactly the same way (6 stops of overexposure, and 7 stops of underexposure). The only difference is how they distribute detail across those 14 stops.

3) In low-key scenarios SLOG3 is going to put more detail into your shadows, and as a result provide more detail through the darker sections of the frame than SLOG2 will. Conversely SLOG2 will hold more detail across the frame in a high-key scenario. I don't switch between the two gammas.

4) I stick to SLOG2 because I consider highlight rendering far more important to a 'filmic' look than shadow detail. It also offers the considerable convenience of false colour monitoring in the EVF, which I find extremely helpful for exposure most of the time.

 

[Ben Scott]

This is definitely a wait for fixes thing for me (hopefully by the new year). I went through this with the Odyssey 7Q as I was one of the first to buy it a year ago. BUG CITY! It was so frustrating and I just don't want go through that again.

My experience has always been that a Super 35mm sensor camera is never really finished. It just gets more and more useable. Until you need to upgrade

 

[jshbckr]

crossposting some info from the "gotchas and bugs" thread:

Here are some pictures of the current C300 plate on the FS7. I'm talking to Berkey about shortening the front of it so it doesn't protrude as much... not sure how much we can really bring it down, may need to just slide it back, but I didn't want to encroach on the built-in shoulder pad. Personally, I really dig the Berkey shoulder pad design on this one, mainly because it provides back rods for my V mount plate and other accessories. The best part is that it's basically quick-release with the rods, and if you want the shoulder pad but not the back rods, they just thread out.

Although this plate is flat like another poster said they wanted (they didn't like that the Vocas plate contoured with the FS7 body), it's important to remember that the rods aren't height adjustable. So using this plate on a future camera (if you're looking at future-proofing) totally depends on that camera's base-to-lens difference. I will chat with Berkey about his thoughts on height-adjustable rods (I know Shape and Zacuto have universal baseplates with height adjustable rods... but they are bigger, fixed shoulder pads, they use sony v-lock tripod setups which require that type of tripod or a $200-300 tripod adapter... Costs add up quick.) There is a lot to be said for being able to break down your baseplate to smaller essential pieces.

Gene, if your matte box is fixed-height but NOT at 85mm, then I would look into a way to make it height adjustable or at least fixed to 85mm.

In terms a top plate, are there any hole patterns that people would prefer? I'm giving some feedback to Brian Berkey regarding the FS7 top plate. He might make a prototype tonight for me to test-drive. Any input would be appreciated, since I know lots of people mount different things to their plates. Additionally, I think he's looking at a design that will be horseshoe shaped to accommodate the stock handle, as well as an inset "filler" plate for folks like me who take off the stock handle in favor of a broad/flat cheese plate top.

One idea I just had is to have a solution for a rods-on-top setup for people who are into that.

 

[skimmel]

Doug Jensen also noted that Sony assured him that S-Log2 will be incorporated into the FS7 (just not sure when)

 

[starcentral]

I have a few questions about S-Log 2 vs S-Log 3

1) Doug Jensen mentioned that Slog 3 clips at 95 IRE while Slog 2 clips at 110 IRE. On paper that seems huge, but is that a huge difference in highlights in a REAL WORLD sense?

2) On the F5 (and presumably the FS7), does Slog 2 and Slog 3 both have 14 stops of DR? It’s just they place the DR in different areas of the curve, correct?

3) Since Slog 3 has more shadow detail, is it better to use it in lowlight? And Slog 2 in bright daytime since it’s focused on the highlights? Do you guys switch between the two?

4) Those of you on an F5 / F55, which Slog do you guys use more? 2 or 3?

You still have the full dynamic range of the camera, it just doesn't go past 95% IRE. Nothing is lost.

The F5/F55 and FS7 are 14 stop cameras. It's not the dynamic range that is placed at different spots, its the data allocated to each stop that is different as seen in my chart. If you were to play back the image exactly as it is recorded, how the data allocation would be irrelevant. But because you need to de-log slog back to a "normal" image and then subsequently color grade it, this is where the data allocation helps and/or makes a difference in effort to reduce banding, pixelation, etc.. BUT as always it will depend on your scene of course and what you are shooting, how or if you are lighting it, etc.. 10-bit is pretty good. 12-bit is much better. Linear RAW is king.

I have switched over from slog2 to slog3 (mostly because of LUTs and the mid-tone data levels). In general I personally don't think there is enough data bits with 10-bit XAVC in slog for an extensive grade so I've also recently added the R5 RAW recorder to my arsenal for the projects that call for it. Lastly I also invested in Neat Video NR reduction and can get more usable stops otherwise too noisy or you would crush down in post and lose. It's night and day to have it.

slog3 "could" be better for pushing (increasing levels) images in post, while slog2 could be better for more highlight detail - but as noted, there is not a huge difference between the two and perhaps still not enough with a 10-bit codec as when compared to 16-bit RAW which allocates nearly 160 times the amount of data levels to the top 2.5 stops that 10-bit XAVC does.

 

[Alpha Beta Omega]

Looking at S-log curves it looks like S-log/S-log2 rolls off on the upper scale while S-log3 is almost linear in the same area.

 

[Kholi]

10-bit is pretty good. 12-bit is much better. Linear RAW is king.

True linear, or closest to it, at that. 16-bit+ Linear RAW is deity if 12 is king. Working with 4K 16-bit's night and day from a 10-bit unless that 10-bit is shot near-to-perfect.

Good posts, Dennis.

 

[starcentral]

I might add there are a lot of misconceptions around this subject as well (RAW, slog, bit-depth) because other cameras like Alexa and F35 which have their color science built in-to the camera can use high bit-depth sensor data for computation BEFORE it is stored into a lower bit container. We often forget that when we look at 8-bit footage from the F35, or 10-bit ProRes from the Alexa. Don't forget what might have happened to it before it was stored in a container. So talking "looks" and "bit-depth" is very deceiving from camera to camera...

Hence if you are going straight off the sensor without any or little further processing leaving it to the post process, then you need some decent color science in post to back that up. RED has spent a lot of time, money and further refinement in this area. Alexa did it in camera. Same as Genesis/F35 - don't forget these were quarter million dollar cameras at one point. If you take the sensor data and store it in slog, the most it will allow you to do is store the full dynamic range of the camera into the "container". You still need enough data bits to manipulate the color science in post to get something that "some" of us are after in terms of getting a specific "look".

Further more its also why I don't support Art Adam's article which says that "Log gives us 99% of the benefit of a raw" - I completely disagree with that.

The maybe good news is that ProRes will come to these F cameras soon and if they support 4444 then it might provide some additional edge or muscle over 10-bit XAVC. I know Grug (Mark) here for example mostly shoots using the SR codec which on the F5/F55 is full RGB 444 and he swears by it. The downside is it can only do HD resolution.

 

[Kholi]

Further more its also why I don't support Art Adam's article which says that "Log gives us 99% of the benefit of a raw" - I completely disagree with that.

Good stuff. I think that Art's comment may be closer to my own generalization of RAW, or what we've known as RAW for a while now in an affordable RAW workflow (RED). I'd get strapped to a chair for this, but I don't think REDCODE is much better than a nice 4K 10-bit ProRes 4444 file, save being able to change metadata here and there. Or, in this case, a well compressed 10-bit XAVC file.

If you were comparing it to ARRIRAW or a very-very-close-to-true-linear, 16-bit, with minimal compression, then no... it's not even close.

Quick Q: where did you learn 8-bit > F35? Is the SDI port only 8? Never knew that.

 

[starcentral]

i've only graded REDcode footage, never tried it in prores 4444 to compare, but I trust your observation.

Re: 8-bit F35 I just meant I've seen footage shot to external recorder in 8-bit and it still grades and looks like dream. It's not the container, its the color science inherent to the camera. (well in my opinion anyway)


More relevant side note: when is FS7 section coming to DVX? :huh:

 

[Gene Sung]

4) I stick to SLOG2 because I consider highlight rendering far more important to a 'filmic' look than shadow detail. It also offers the considerable convenience of false colour monitoring in the EVF, which I find extremely helpful for exposure most of the time.

Damn... that's another thing to wait for now on the FS7 (besides the bug fix). I do love the SLOG 2 highlights (as I live in sunny Los Angeles) and wouldn't want to not have it. Please Sony, hurry up with this stuff as my purchase is depending on it.

 

[jshbckr]

Damn... that's another thing to wait for now on the FS7 (besides the bug fix). I do love the SLOG 2 highlights (as I live in sunny Los Angeles) and wouldn't want to not have it. Please Sony, hurry up with this stuff as my purchase is depending on it.

I also didn't realize how big of a difference there was in Slog3 vs Slog2... in my naivety I assumed that it was just a "newer better version", didn't really look into how much the curve was different. I'm also a big fan of Slog2. Hopefully it comes soon for the FS7.

 

[Kholi]

i've only graded REDcode footage, never tried it in prores 4444 to compare, but I trust your observation.

Re: 8-bit F35 I just meant I've seen footage shot to external recorder in 8-bit and it still grades and looks like dream. It's not the container, its the color science inherent to the camera. (well in my opinion anyway)


More relevant side note: when is FS7 section coming to DVX? :huh:

Let me ask about it today. If anything I'll see how soon we can get a sub-header thing.

Good point about the 8-bit to recorder, it's more than ample in most cases. I agree with you, or well it's a mix of both yeah?


RE: SLog-2 vs 3 ... I'm wondering if the purpose of SLog3's a response to the cry for better lowlight response. Just off of the top of my head, shooting SLog3 and protecting highlights should yield pretty deep mids and the ability to simulate/create your own roll off, especially if you're shooting RAW? In that case you get both the benefits of better color retention and lowlight ability in one?

I haven't had breakfast yet, that's my excuse.

 

[Gene Sung]

Hey Josh Becker,

Are you noticing a big difference in the noise between the FS7 and the 7Q? The shadows on SLOG 2 on the 7Q can be quite noisy. I'm wondering if the FS7 implements Noise Reduction in camera to fix this?

I guess it's a trade off as you can just do NR in post when needed, and keep more detail in the image.

 

[Jason Ramsey]

Do you folks want this as a separate sub-forum, or as a prefix in this existing sub-forum?

 

[starcentral]

Regarding slog2 vs slog3

But as explained in my post, if you are still getting your 2.5 stops above 90% white and the only difference are data levels which are mostly important for grading - is it really going to make a difference? Does anyone here really grade highlights that they think the need all that data? You still have 14.5 levels of white above 90% white with slog3 or slog2 - you just have less data per stop but not much less.

I personally have not done any extensive analysis or testing to conclude one way or another but if anyone has anything similar in terms of tests I'd love to see it. I have a hard time believing there would be a big difference. Most tests people have done didn't account for exposing the slogs differently so it caused slog3 to clip earlier than it would with slog2 at the same f-stop. If you clip slog3 its going to look clippy, but that becomes an exposure issue not an slog3 issue.

By the way check out this comparison video on slog2 versus slog3. See if you can find faults with any highlights.

[BIG FAT EDIT HERE]


http://wolfcrow.com/blog/sony-f55-s-log-3-vs-s-log-2-comparison/

 

[Gene Sung]

Do you folks want this as a separate sub-forum, or as a prefix in this existing sub-forum?

Maybe a separate sub-forum as this forum already has 3 prefixes - F3, F5, F55.

If the FS7 does have to share a forum, maybe it's better with the FS700? Less cameras, they are closer in price range and many of the people who own FS700 are moving to the FS7.

Thanks!