I go over that in blog post about the shot w/ tom. I talk about exactly where the elements in the shot are with relation to the mid grey (when I rate midgrey) at 30-35.
Slog must not be as good as they claim
- Thread starter Douglas Villalba
- Start date
Douglas Villalba
Well-known member
I'll check it out.
Hey Timur let me ask you something since you actually have seen the mount for the KiPromini. Does it have a single plate between the battery and the KPM or does it have a plate at both sides of the KPM?
Hey Timur let me ask you something since you actually have seen the mount for the KiPromini. Does it have a single plate between the battery and the KPM or does it have a plate at both sides of the KPM?
Both sides. It sandwiches the kipm
Rick Burnett
Section Moderator
The chart is there to take all the guesswork out of the exposure and curve of exposure for people curious about it. The best exposure tool I have used is a waveform monitor. I exposure for what the waveform monitor shows and that's it. I can see where EVERYTHING is exposing, and make adjustments to the camera accordingly. Once I've been shooting with a camera for a few months, I honestly don't have to think about it too much. I also use zebras as well if I am not using a waveform (or histogram, as I have an FS100 now, but I sure miss the WFM of the AF100 for exposure).
I don't use an F3, but I've looked and worked with enough footage from the Alexa and Red to understand what is going on. One thing that is true, and I think has been elaborated on in a few places here is that S-Log or high dynamic range is not going to fix an incorrectly lit scene. Just like the image of the blown out sky on the 7D, the work to fix that on video would take quite a bit (of course, at least having that option could be argued a good thing). But, I think that is more for a lack of tools available right now then possibility. Have you played with HDR photography? HDR can be abused, or used amazingly. What it shows more than anything is this unbelievable control of more DR and how that DR is distributed in a smaller DR. You get to pick how that works. And, as stated, I think the HL handling can be done a lot better in post with more control over it.
In fact, I listened to many of the same criticisms on other 'milky' footage from the Epic and the Alexa because what most people start showing is all these videos of poor lighting that just contain a HUGE range of light that you'd never be able to catch before. I downloaded a few to play with and realized, wow, this lighting is so bad, nothing could be done to fix the image. HOWEVER, being able to show that much dynamic range was what was important in that particular test. As is aggravating you, they didn't show a good use of the feature. But, seeing a beach scene with no blow-outs and knowing my own work on beach scenes with the 7D, I was instantly amazed at how my lighting decisions (or shading) would be different and have more freedom. I didn't have to fit everything into a smaller window, I could have more detail in the shadows AND the highlights. However, I still needed to get my important subjects, like the people, at the right exposure!
The understanding I have of S-Log is that you want to not crush the top or bottom end, but allow more detail to spread out over a larger bit depth. The purpose is not to let values in the compressed ends at the top and bottom get quantized away. If you do let something quantize as you are crushing the image in camera, as you stretch that area back out, say to open up and get the detail out, it's gone. Or, you can get banding. Or, you can crush right down into the noise. From the graphs I have seen, it's just more of a linear recording, which isn't that pretty
With S-Log, you are giving yourself freedom at the top and bottom end by not recording the image for looks, but for information saving, so that you can have it there to use in post. It just gives you more freedom to push your image up or down, and generally means people are going to shoot more of a neutral image and NOT try to get the final result in camera (which is what most DLSR people are used to doing). It's a whole different mindset. Shooting neutral like this is a lot harder I think because you want the image to look closer to the end product, if that is what you are used to shooting. I know I'd have a hard time with it.
With the Red I believe you can change your monitor curves so you can view it any way you'd like and still have it recorded flat.
I don't use an F3, but I've looked and worked with enough footage from the Alexa and Red to understand what is going on. One thing that is true, and I think has been elaborated on in a few places here is that S-Log or high dynamic range is not going to fix an incorrectly lit scene. Just like the image of the blown out sky on the 7D, the work to fix that on video would take quite a bit (of course, at least having that option could be argued a good thing). But, I think that is more for a lack of tools available right now then possibility. Have you played with HDR photography? HDR can be abused, or used amazingly. What it shows more than anything is this unbelievable control of more DR and how that DR is distributed in a smaller DR. You get to pick how that works. And, as stated, I think the HL handling can be done a lot better in post with more control over it.
In fact, I listened to many of the same criticisms on other 'milky' footage from the Epic and the Alexa because what most people start showing is all these videos of poor lighting that just contain a HUGE range of light that you'd never be able to catch before. I downloaded a few to play with and realized, wow, this lighting is so bad, nothing could be done to fix the image. HOWEVER, being able to show that much dynamic range was what was important in that particular test. As is aggravating you, they didn't show a good use of the feature. But, seeing a beach scene with no blow-outs and knowing my own work on beach scenes with the 7D, I was instantly amazed at how my lighting decisions (or shading) would be different and have more freedom. I didn't have to fit everything into a smaller window, I could have more detail in the shadows AND the highlights. However, I still needed to get my important subjects, like the people, at the right exposure!
The understanding I have of S-Log is that you want to not crush the top or bottom end, but allow more detail to spread out over a larger bit depth. The purpose is not to let values in the compressed ends at the top and bottom get quantized away. If you do let something quantize as you are crushing the image in camera, as you stretch that area back out, say to open up and get the detail out, it's gone. Or, you can get banding. Or, you can crush right down into the noise. From the graphs I have seen, it's just more of a linear recording, which isn't that pretty
With S-Log, you are giving yourself freedom at the top and bottom end by not recording the image for looks, but for information saving, so that you can have it there to use in post. It just gives you more freedom to push your image up or down, and generally means people are going to shoot more of a neutral image and NOT try to get the final result in camera (which is what most DLSR people are used to doing). It's a whole different mindset. Shooting neutral like this is a lot harder I think because you want the image to look closer to the end product, if that is what you are used to shooting. I know I'd have a hard time with it.
With the Red I believe you can change your monitor curves so you can view it any way you'd like and still have it recorded flat.
The chart is there to take all the guesswork out of the exposure and curve of exposure for people curious about it. The best exposure tool I have used is a waveform monitor. I exposure for what the waveform monitor shows and that's it. I can see where EVERYTHING is exposing, and make adjustments to the camera accordingly. Once I've been shooting with a camera for a few months, I honestly don't have to think about it too much. I also use zebras as well if I am not using a waveform (or histogram, as I have an FS100 now, but I sure miss the WFM of the AF100 for exposure).
I don't use an F3, but I've looked and worked with enough footage from the Alexa and Red to understand what is going on. One thing that is true, and I think has been elaborated on in a few places here is that S-Log or high dynamic range is not going to fix an incorrectly lit scene. Just like the image of the blown out sky on the 7D, the work to fix that on video would take quite a bit (of course, at least having that option could be argued a good thing). But, I think that is more for a lack of tools available right now then possibility. Have you played with HDR photography? HDR can be abused, or used amazingly. What it shows more than anything is this unbelievable control of more DR and how that DR is distributed in a smaller DR. You get to pick how that works. And, as stated, I think the HL handling can be done a lot better in post with more control over it.
In fact, I listened to many of the same criticisms on other 'milky' footage from the Epic and the Alexa because what most people start showing is all these videos of poor lighting that just contain a HUGE range of light that you'd never be able to catch before. I downloaded a few to play with and realized, wow, this lighting is so bad, nothing could be done to fix the image. HOWEVER, being able to show that much dynamic range was what was important in that particular test. As is aggravating you, they didn't show a good use of the feature. But, seeing a beach scene with no blow-outs and knowing my own work on beach scenes with the 7D, I was instantly amazed at how my lighting decisions (or shading) would be different and have more freedom. I didn't have to fit everything into a smaller window, I could have more detail in the shadows AND the highlights. However, I still needed to get my important subjects, like the people, at the right exposure!
The understanding I have of S-Log is that you want to not crush the top or bottom end, but allow more detail to spread out over a larger bit depth. The purpose is not to let values in the compressed ends at the top and bottom get quantized away. If you do let something quantize as you are crushing the image in camera, as you stretch that area back out, say to open up and get the detail out, it's gone. Or, you can get banding. Or, you can crush right down into the noise. From the graphs I have seen, it's just more of a linear recording, which isn't that pretty
With S-Log, you are giving yourself freedom at the top and bottom end by not recording the image for looks, but for information saving, so that you can have it there to use in post. It just gives you more freedom to push your image up or down, and generally means people are going to shoot more of a neutral image and NOT try to get the final result in camera (which is what most DLSR people are used to doing). It's a whole different mindset. Shooting neutral like this is a lot harder I think because you want the image to look closer to the end product, if that is what you are used to shooting. I know I'd have a hard time with it.
With the Red I believe you can change your monitor curves so you can view it any way you'd like and still have it recorded flat.
Excellent post!
The chart is there to take all the guesswork out of the exposure and curve of exposure for people curious about it. The best exposure tool I have used is a waveform monitor. I exposure for what the waveform monitor shows and that's it. I can see where EVERYTHING is exposing, and make adjustments to the camera accordingly. Once I've been shooting with a camera for a few months, I honestly don't have to think about it too much. I also use zebras as well if I am not using a waveform (or histogram, as I have an FS100 now, but I sure miss the WFM of the AF100 for exposure).
I don't use an F3, but I've looked and worked with enough footage from the Alexa and Red to understand what is going on. One thing that is true, and I think has been elaborated on in a few places here is that S-Log or high dynamic range is not going to fix an incorrectly lit scene. Just like the image of the blown out sky on the 7D, the work to fix that on video would take quite a bit (of course, at least having that option could be argued a good thing). But, I think that is more for a lack of tools available right now then possibility. Have you played with HDR photography? HDR can be abused, or used amazingly. What it shows more than anything is this unbelievable control of more DR and how that DR is distributed in a smaller DR. You get to pick how that works. And, as stated, I think the HL handling can be done a lot better in post with more control over it.
In fact, I listened to many of the same criticisms on other 'milky' footage from the Epic and the Alexa because what most people start showing is all these videos of poor lighting that just contain a HUGE range of light that you'd never be able to catch before. I downloaded a few to play with and realized, wow, this lighting is so bad, nothing could be done to fix the image. HOWEVER, being able to show that much dynamic range was what was important in that particular test. As is aggravating you, they didn't show a good use of the feature. But, seeing a beach scene with no blow-outs and knowing my own work on beach scenes with the 7D, I was instantly amazed at how my lighting decisions (or shading) would be different and have more freedom. I didn't have to fit everything into a smaller window, I could have more detail in the shadows AND the highlights. However, I still needed to get my important subjects, like the people, at the right exposure!
The understanding I have of S-Log is that you want to not crush the top or bottom end, but allow more detail to spread out over a larger bit depth. The purpose is not to let values in the compressed ends at the top and bottom get quantized away. If you do let something quantize as you are crushing the image in camera, as you stretch that area back out, say to open up and get the detail out, it's gone. Or, you can get banding. Or, you can crush right down into the noise. From the graphs I have seen, it's just more of a linear recording, which isn't that pretty
With S-Log, you are giving yourself freedom at the top and bottom end by not recording the image for looks, but for information saving, so that you can have it there to use in post. It just gives you more freedom to push your image up or down, and generally means people are going to shoot more of a neutral image and NOT try to get the final result in camera (which is what most DLSR people are used to doing). It's a whole different mindset. Shooting neutral like this is a lot harder I think because you want the image to look closer to the end product, if that is what you are used to shooting. I know I'd have a hard time with it.
With the Red I believe you can change your monitor curves so you can view it any way you'd like and still have it recorded flat.
Indeed an excellent post. In a sense, yes and no. Yes to everything but the last paragraph. You light how you want it to look, then you shoot lg with a custom lut that massages the log into what you want the signal to look like, on set. So you are still shooting for look, you just have some protection on the top and bottom to help get you there. Its not really that different from shooting 709, the confusion seemt to be steming from placement of values on the wave form. Putting skintone @ 35ire, isn't underexposing. Its proper exposure. It looks wrong and milky if you view it without a lut, you are supposed to use a LUT... that's how the work flow works.....you of course look at log to make sure everything is ok, but you can light based on the log waveform and double check on the lut monitor (that director and or client see) that it looks right. The point being you are still kinda lighting for "709" but now you have all this fantastic protection on the top and bottom.
OK, so this is sLOG with a bit more punch.
This believe it or not was graded by me, on Vegas 10 from Slog captured on XDcam. So it is possible, but the truth is VIMEO compression is hiding alot of the gak that gets brought up. i STRONLGY reccomend you shoot to 10bit for LOG. but in a pinch its possible if you just need one pick up shot. I was however impressed with the color fidelity in XDCAM.
This is what the LOG looked like...
This believe it or not was graded by me, on Vegas 10 from Slog captured on XDcam. So it is possible, but the truth is VIMEO compression is hiding alot of the gak that gets brought up. i STRONLGY reccomend you shoot to 10bit for LOG. but in a pinch its possible if you just need one pick up shot. I was however impressed with the color fidelity in XDCAM.
This is what the LOG looked like...
Rick Burnett
Section Moderator
Harder in the sense if you don't look at it with a custom LUT. I mentioned the Red because I knew specifically you can adjust how you look at the material. I wasn't sure about the F3 specifically, but it only makes sense to me that it has that capability as well. The SLog 'actual' values getting recorder are to maximise the use of the bit space available. However, it's not appealing to most to look at.
That said, I have seen people who record without using a custom LUT to look through. They know the SLog visuals well enough that they use that and just their exposure with that. Just like there are people who will just shoot with B&W viewfinders. Some people just like different methods of viewing material.
I would definitely be using a custom LUT myself!
That said, I have seen people who record without using a custom LUT to look through. They know the SLog visuals well enough that they use that and just their exposure with that. Just like there are people who will just shoot with B&W viewfinders. Some people just like different methods of viewing material.
I would definitely be using a custom LUT myself!
SLog must not be as good as they claim?
Alister Chapman recently wrote an interesting post about “Exposing when shooting S-Log”. And he mentions some interesting things I didn't know, like “there is a +6db increase in NOISE due to the +1 stop increase in sensitivity associated with S-Log.” (emphasis mine). In other words, the S-Log image becomes twice as noisy as the cinegamma material.
So it seems that using S-Log is not always going to give you the best results: “it is quite possible that you would actually be able to push mid ranges and shadows further with Cinegammas than S-log in an underexposed situation due to noise issues.”
He concludes that if he can fit the contrast range of the scene into the 11.5 stops of a cinegamma then he will often use the cinegammas over S-Log because of the noise improvement. “S-Log comes into it’s own where you have an extreme contrast range that needs to be captured.”
And on the use of S-Log with 422 instead of uncompressed, he points out that “one thing to consider is that with S-Log you no longer have the ability to change the white balance in-camera. It is locked into preset mode at either 3200 or 5600.” This means that you will benefit from recording at 4:4:4 over 4:2:2, as you will need to fine tune the color balance in post.
http://www.xdcam-user.com/2011/09/exposing-when-shooting-s-log/#comment-970
Alister Chapman recently wrote an interesting post about “Exposing when shooting S-Log”. And he mentions some interesting things I didn't know, like “there is a +6db increase in NOISE due to the +1 stop increase in sensitivity associated with S-Log.” (emphasis mine). In other words, the S-Log image becomes twice as noisy as the cinegamma material.
So it seems that using S-Log is not always going to give you the best results: “it is quite possible that you would actually be able to push mid ranges and shadows further with Cinegammas than S-log in an underexposed situation due to noise issues.”
He concludes that if he can fit the contrast range of the scene into the 11.5 stops of a cinegamma then he will often use the cinegammas over S-Log because of the noise improvement. “S-Log comes into it’s own where you have an extreme contrast range that needs to be captured.”
And on the use of S-Log with 422 instead of uncompressed, he points out that “one thing to consider is that with S-Log you no longer have the ability to change the white balance in-camera. It is locked into preset mode at either 3200 or 5600.” This means that you will benefit from recording at 4:4:4 over 4:2:2, as you will need to fine tune the color balance in post.
http://www.xdcam-user.com/2011/09/exposing-when-shooting-s-log/#comment-970
On paper that all makes perfect sense, but in real world shooting.... quite the opposite. Its like saying "redocde compresses too much thus it must be no good..." but in reality, its amazing.
The f3 is so grainless that there is hardly a visual difference between 0dB and 6dB. The more accurate tone rendition, color rendition, and increased lattitude FAR outweighs, the benfits of 6dB noise improvement.
It seems alistair and I disagree strongly on this. We have a spirited exchange on dv info. He has super valid points, but he's forgetting that the color accuracy, and tonal accuracy in log is a big part of the better higher end look. This is what I'm after. The color accuracy and high dynamic range just reads beautiful to the human eye.
The f3 is so grainless that there is hardly a visual difference between 0dB and 6dB. The more accurate tone rendition, color rendition, and increased lattitude FAR outweighs, the benfits of 6dB noise improvement.
It seems alistair and I disagree strongly on this. We have a spirited exchange on dv info. He has super valid points, but he's forgetting that the color accuracy, and tonal accuracy in log is a big part of the better higher end look. This is what I'm after. The color accuracy and high dynamic range just reads beautiful to the human eye.
LeavingTheCandy
Well-known member
Amen Timur. I had to get SLog to appreciate it--the sensitivity boost, color rendition FAR outweighs the teeny-tiny grain increase (I don't see any).
I would cowboy-up the money just for the tone and color rendition. It's night and day.
Let the fanboys rejoice!
I would cowboy-up the money just for the tone and color rendition. It's night and day.
Let the fanboys rejoice!