FS7: Cine EI Slog 3 vs. Custom Hypergamma 7

joobny

Member
Hey all,

I have a shoot coming up at the weekend and I'd like some advice.

I love shooting Cine EI sgammut3.cine/slog3 on my FS7, but I'll be shooting in a situation where I need to capture events during the day and also at night (with very minimal lighting - probably car headlights and street lights only). I want to shoot Cine EI Slog3 during the day, but obviously this isn't going to work at night time.

I always used to shoot predominantly with Slog2, and if I found myself in that situation I would just shoot on fast lenses in custom mode and as the light dropped I'd bump the ISO up to 5000 or so.

I've read a lot about log footage being no good for low light situations though, so for my shoot at the weekend I was thinking about using Hypergamma 7 when the light drops. My questions is... how well will this footage match up with the other Slog3 footage that I'm shooting during the day? And at what point will I make the decision to switch to Hypergamma 7? Presumably when the light gets to a point that I'm no longer to overexpose the Slog3 footage by one stop?

Would like to hear everyone's thoughts on this. Thanks in advance!
 
In my experience you can shoot in low light using Cine EI / S-Log3 if you expose to the right. Set your ISO to 800 and use a fast lens. Shoot some tests and give it a try. I've never used custom mode so I can't comment on HG7. Interested to see what other write on this subject.
 
Hi Jack,

Thanks for your reply. The problem is that in the lighting situation that I'll find myself in at the end of the day (street lights and car headlights at the very best), I will not be able to expose to the right, even on my t1.5 lenses. But I'd like to shoot Cine EI during the day when I am able to overexpose, so was wondering what the best alternative settings for the night scenes will be. I've read that log footage is no good for low light situations and that Hypergammas are they way to go (with 7 being the best for low light, and 8 for daylight), but I'm unsure of how this will match up with the daytime Slog3 footage when it comes to editing. I'm also a little unsure at what point I should switch from Cine EI to custom mode - presumably as soon as the light starts to drop below the necessary amount to expose Slog3 footage correctly?

Thanks again.
 
Funny you would come up with this right when I've had similar dilemma - whether to use Slog3 in a dark room with just one bulb of a desktop little lamp. I run a test, comparing HG7 (the brightest of all FS7 HGs, with lots of DR too) with Sgamut3/Slog3 - and I must tell you I had to run NR in Resolve in both cases, but at the end I preferred the Slog3 picture!

Piotr
 
Hi Jack,

Thanks for your reply. The problem is that in the lighting situation that I'll find myself in at the end of the day (street lights and car headlights at the very best), I will not be able to expose to the right, even on my t1.5 lenses. But I'd like to shoot Cine EI during the day when I am able to overexpose, so was wondering what the best alternative settings for the night scenes will be. I've read that log footage is no good for low light situations and that Hypergammas are they way to go (with 7 being the best for low light, and 8 for daylight), but I'm unsure of how this will match up with the daytime Slog3 footage when it comes to editing. I'm also a little unsure at what point I should switch from Cine EI to custom mode - presumably as soon as the light starts to drop below the necessary amount to expose Slog3 footage correctly?

Thanks again.

Hi,

Try to film some quick tests before your shoot, perhaps on your street on an evening with the street lights. Even at the native 2000 ISO in CineEI I think it should work. You may have to use Neat Video to remove some noise. If you really feel the need to use HG7 then I'm sure the low light footage can be graded to match in Resolve. I would test both to see which you prefer. You could also try burn the REC709 Type A LUT in to your footage in EI mode and see how that works for you. I really can't speak about custom mode because I've never felt the need to use it but perhaps some other on here may comment. Cheers
 
I was in the same boat http://www.dvxuser.com/V6/showthread.php?349684-need-advice-(custom)-settings-for-event-docu
and went for HG7 matrix off. It went all fine. The low light perfomance was good but the FS7 isn't a low light monster in general.

I shot everything in a hurry with HG7. After the event I wish I had switched back to CineEI for some interviews with LED panels.
It's okay in HG7 too but if you are used CineEI you know easy it would have looked nicer. But nevertheless HG7 is not toxic and
the XAVC-I is very flexible for adjustments
 
Last edited:
Okay, so I decided to do some tests and would be interested to hear your thoughts.

I've uploaded two back to back clips to YouTube. The first clip is HG7 at ISO 5000, and the second is Cine EI Slog3 with the exposure lifted using levels (RGB Gamma to 120, because you obviously can't crank the ISO higher than 2000 in Cine EI). Both of them have a Slog3 Rec 709 LUT applied (I know that's not ideal for HG7, but to quickly match the overall look it should do the job for the purpose of this test).

I'd be interested to hear which you prefer.

https://youtu.be/2BxYOhezjUA

I think for the sake of sticking with one gamma for the entire shoot, I'd be inclined to stick with Cine EI. I think looks ever so slightly noisier, but not enough to make me want to swap to HG7.

Thanks!
 
I'm actually more keen on the HG7 - I think it's got a lot of YouTube compression going on, might be better to include some .jpg frame grabs of both out of resolve or Premiere. I completely get what you're saying about the trade offs, better to have the majority of your project in SLog3 and have a few obviously under exposed (and slightly noisy) night exteriors.

Failing that, make it a feature, crunch the blacks loads and rely on the outline of the cars against lights, when they catch specular highlights or using the headlights and their effect on the ground. Good luck man!
 
Last edited:
Hypergammas have no advantage over SLOG2/3, they are just a baked 1D LUT. I would never use hypergammas if it's avoidable. Just shoot Slog2 or Slog3 in Custom and use gain if you need in low light. Hypergammas can be OKish in low contrast scenes but in high contrast (such as at night) highlights get all messed up due to oversaturation and color shifts that can't be controlled with a 1D LUT or simple gamma curve.
 
Hypergammas have no advantage over SLOG2/3, they are just a baked 1D LUT. I would never use hypergammas if it's avoidable. Just shoot Slog2 or Slog3 in Custom and use gain if you need in low light. Hypergammas can be OKish in low contrast scenes but in high contrast (such as at night) highlights get all messed up due to oversaturation and color shifts that can't be controlled with a 1D LUT or simple gamma curve.

@nyvz So many misconceptions in your answer. You may want to read these articles: http://www.xdcam-user.com/2015/10/if-shooting-in-low-light-dont-use-log/ http://www.xdcam-user.com/2016/01/log-is-a-poor-choice-for-low-light/
 
Hypergammas have no advantage over SLOG2/3, they are just a baked 1D LUT. I would never use hypergammas if it's avoidable. Just shoot Slog2 or Slog3 in Custom and use gain if you need in low light. Hypergammas can be OKish in low contrast scenes but in high contrast (such as at night) highlights get all messed up due to oversaturation and color shifts that can't be controlled with a 1D LUT or simple gamma curve.

????
 
Okay, so I decided to do some tests and would be interested to hear your thoughts.

I've uploaded two back to back clips to YouTube. The first clip is HG7 at ISO 5000, and the second is Cine EI Slog3 with the exposure lifted using levels (RGB Gamma to 120, because you obviously can't crank the ISO higher than 2000 in Cine EI). Both of them have a Slog3 Rec 709 LUT applied (I know that's not ideal for HG7, but to quickly match the overall look it should do the job for the purpose of this test).

you applied a sLog3/REC709 LUT to HG7 footage?
 
Okay, so I decided to do some tests and would be interested to hear your thoughts .... Both of them have a Slog3 Rec 709 LUT applied (I know that's not ideal for HG7, but to quickly match the overall look it should do the job for the purpose of this test).

First thought is WHHHAAAATTTTT? Second thought is that the bit about 'quickly match' by applying an S-Log3 LUT to HG7 material is fundamentally wrong, and when used to produce a qualitative comparison test is, I'm afraid, at best pointless and potentially misleading to those who see it. I am honestly sorry to be a downer, but I personally think that it is important to understand this, as I've seen plenty of inappropriate tests on YouTube which have resulted in unhelpful conclusions on forums.

In gamma terms HG7 - indeed all the hypergammas - are NOT log responses. Bung HG7 material into an S-Log3 LUT and unsurprisingly it will look rather different to S-Log3 material fed in. Dark stuff will still be dark, bright stuff will still be bright, but that's about it in terms of 'matching'.

If the LUT applied is a 3D one, then colour also comes into play. In CineEI S-Log3 is generally paired with S-Gamut3.cine (indeed, most LUTs available are geared around this - certainly all the Sony ones). This is a very wide gamut, which means that in the recording colours are extremely muted .Just as gamma curves clip at a white level, colours for a given gamut can clip when values fall outside of the limited range of 8 or 10-bit integers. For example, saturated blue in Rec709 RGB terms might be 100% blue and no red or green. Even more saturated blue could then be 100% blue and negative red and green, but you can't 'legally' have negative values.

By effectively muting the colours recorded, a wide gamut can allow for a greater range before the recording range buffers get hit.

To get a nice picture, a 3D LUT for S-Gamut3.cine will convert to the more limited Rec709 gamut. There will often be tweaks to the basic Rec709 conversion to reduce clipping, but the fundamental, most obvious effect is an increase in saturation. The end result is that a washed-out, muted S-Log3/S-Gamut3.cine picture should appear as something truer to life on the screen.

HG7 in custom mode, on the other hand, is generally tied to a colour gamut closer to (or actually) Rec709. It is a WYSIWYG way of shooting, and nice colours look good to the client and make efficient use of the recording data provided there is nothing much too far outside of gamut. What that does mean is that putting aside the contrast difference, if you apply an S-Log3/S-Gamut3.cine 3D LUT to it, you are effectively doubling up on the saturation lift. You are going to get psychedelically vibrant colours, and a strong chance of chroma clipping even on relatively muted colours in the scene. To me, that's what's happening to the wall in your shot.

Having said all that and going to personal opinion, I am cautious of Sony's mechanism for log at high ISOs (I'd certainly stay well clear of S-Log2). Hypergammas with their baked-in gain can be useful there, and frankly the prime reason for log is to be able to get a decent picture in post from difficult shooting situations. If you can get a nice picture from a Hypergamma there is no reason not to be able to polish it up in post, so why not use it? Mind you, in the hands of a decent colourist, S-Log3 can give decent results across the CineEI range (which is not to say 'better'). I've shot both ways, depending upon the nature of the project and the preferences of post (which, as a lighting cameraman and DoP I am not generally significantly involved in),

Ben
 
Last edited:
First thought is WHHHAAAATTTTT? Second thought is that the bit about 'quickly match' by applying an S-Log3 LUT to HG7 material is fundamentally wrong, and when used to produce a qualitative comparison test is, I'm afraid, at best pointless and potentially misleading to those who see it. I am honestly sorry to be a downer, but I personally think that it is important to understand this, as I've seen plenty of inappropriate tests on YouTube which have resulted in unhelpful conclusions on forums.

In gamma terms HG7 - indeed all the hypergammas - are NOT log responses. Bung HG7 material into an S-Log3 LUT and unsurprisingly it will look rather different to S-Log3 material fed in. Dark stuff will still be dark, bright stuff will still be bright, but that's about it in terms of 'matching'.

If the LUT applied is a 3D one, then colour also comes into play. In CineEI S-Log3 is generally paired with S-Gamut3.cine (indeed, most LUTs available are geared around this - certainly all the Sony ones). This is a very wide gamut, which means that in the recording colours are extremely muted .Just as gamma curves clip at a white level, colours for a given gamut can clip when values fall outside of the limited range of 8 or 10-bit integers. For example, saturated blue in Rec709 RGB terms might be 100% blue and no red or green. Even more saturated blue could then be 100% blue and negative red and green, but you can't 'legally' have negative values.

By effectively muting the colours recorded, a wide gamut can allow for a greater range before the recording range buffers get hit.

To get a nice picture, a 3D LUT for S-Gamut3.cine will convert to the more limited Rec709 gamut. There will often be tweaks to the basic Rec709 conversion to reduce clipping, but the fundamental, most obvious effect is an increase in saturation. The end result is that a washed-out, muted S-Log3/S-Gamut3.cine picture should appear as something truer to life on the screen.

HG7 in custom mode, on the other hand, is generally tied to a colour gamut closer to (or actually) Rec709. It is a WYSIWYG way of shooting, and nice colours look good to the client and make efficient use of the recording data provided there is nothing much too far outside of gamut. What that does mean is that putting aside the contrast difference, if you apply an S-Log3/S-Gamut3.cine 3D LUT to it, you are effectively doubling up on the saturation lift. You are going to get psychedelically vibrant colours, and a strong chance of chroma clipping even on relatively muted colours in the scene. To me, that's what's happening to the wall in your shot.

Having said all that and going to personal opinion, I am cautious of Sony's mechanism for log at high ISOs (I'd certainly stay well clear of S-Log2). Hypergammas with their baked-in gain can be useful there, and frankly the prime reason for log is to be able to get a decent picture in post from difficult shooting situations. If you can get a nice picture from a Hypergamma there is no reason not to be able to polish it up in post, so why not use it? Mind you, in the hands of a decent colourist, S-Log3 can give decent results across the CineEI range (which is not to say 'better'). I've shot both ways, depending upon the nature of the project and the preferences of post (which, as a lighting cameraman and DoP I am not generally significantly involved in),

Ben

Hey Ben,

Thanks for your response. I did realise at the time that applying a Rec 709 LUT to non log footage didn't make any sense, but thanks for pointing out exactly why. Hopefully it'll help a few people including myself understand the fundamentals of it a bit better.

I've exported another few clips. This time, there are six clips:

1. HG7 ISO 5000, untouched
2. Cine EI Slog 3
3. HG7 ISO 5000, untouched (as above)
4. Cine EI Slog 3 with exposure lifted to closer resemble the HG7 clip's exposure
5. HG7 ISO 5000, untouched (as above)
6. Cine EI Slog 3 with exposure lifted to closer resemble the HG7 clip's exposure and contrast increased slightly to try and mimic the HG7 clip as closely as possible

https://youtu.be/UWy7VN2i-b4

Perhaps I'm still doing something wrong in terms of a fair test, so happy to hear if anything can be changed. It's something that's been bugging me since I've owned the camera so would like to get it cleared up! The question now I guess would be what they look like side-by-side when they've been given a similar grade...? Something I don't have time to test for a few weeks unfortunately.

The reason I have been avoiding HG7 is as per my original post. I'll be shooting during the day time and into the night and wasn't sure at which point during the day it would be logical to make the jump over to HG7 for the sake of a consistent look from day to night.

From shooting the tests and looking at the footage, I've discerned that there's really minimal difference anyway. So in my opinion, I'd be as well just sticking with Slog3 for the whole shoot as this is what I'd like to be using during the day. Interested to see what others would do though.

Thanks again!
 
\Interested to see what others would do though.
Thanks again!

You know, the Hypergammas are not supposed to be stand-alone settings like LOG. They are designed to be the "gamma" component of a Scene File which must encompass many other paint menu settings that are chosen to work in concert with the Hypegamma (or any other gamma that has been chosen) to create a nice finished image. In other words, choosing a gamma without attention to black levels, matrix, detail, saturation, aperture, etc. is rather like judging a batch of cookies that were made only of flour but had no sugar, butter, etc. If you're going to compare HG7 to LOG then you need to compare: A Scene File that has many custom paint menu choices vs. LOG that has been graded properly in post. Then you can find out which one works best for you, but until you actually test them both properly you are really just stumbling around in the dark.
 

So many thing wrong with Alister Chapman's articles on this from what I can tell. It assumes using LOG requires that you expose with the manufacturer's specs for ISO and exposure, and it seems to conflate CineEI and Log...

Only thing correct about that in my mind is that if you don't mind all the downsides of a 1D LUT and the locked in lower DR range and specific DR distribution of HG then yes HG may be better.

His articles completely forget that you can forget about Sony's ridiculous log ISO rating system and get exactly the same performance from log as with HG. Yes if your scene has only 6 stops of range maybe a 14 stop log range is overkill but in many cases you will still want the range in low light since light sources are that much more likely to blow out relative to the rest of your scene, and you can always just use the same gain as you would have used on HG and not waste any highlight range in your LOG, and still have the flexibility of LOG in post.

Please do let me know what you think my specific misconceptions are, because I don't see any brought up by Alister's articles that change my thinking since his system introduces many caveats that I would not accept.
 
Back
Top