steps to avoid Noise and Softness

[Martin Ubilluz]

What do you do in the case that your shooting a video that has many many different FPS and shutter speed changes through out the day... am I supposed to stop production every time there is a change? The video will never get done.

 

[Björn Benckert]

OK - let's start with Back Focus; I am happy to have someone correct me if I am wrong here but unless you are using the distance marks on your lens to determine your focus, back focus very likely has little or nothing to do with whether your image is in focus (unless it is WAAAAAAY off). Given there is no really good way to dial in perfect back focus (at least with your Canon mount) and the likelihood of your back focus being way off as it comes from RED is very low), 99% of people probably SHOULDN'T mess with the back focus adjustment.

If you are using all the other tools on your Scarlet to determine focus, this is most likely a total non-issue.

Cameras that comes form red is usually far from perfect backfocus calibrated, and so are the lenses.

But I agree, as long as you go by image and not focus marks the back focus is quite irrelevant.. most of us finger tip zoom in find focus and shoot. when having a focus puller around the above list is usually all taken care off.


All sensor has noise. Epic delivers compressed noise It might not be visible to your eye, but your keyer will definitely see it. so will your selective color corrector. So to use things like neatvideo or even better FLAME Denoise, unsharp mask and then regrain nodes will let you remove the quite ugly patterns hiding in your image. Also giving it sharpen and a slight regrain on top will make your image hold up better when keying and applying forced grading.

 

[Scott Crawley]

The info we have regarding frame rate changes is not terribly clear. This is one area that Red plans to improve according to Biederman. The impression I get is that the higher your frame rate is, the more beneficial it is to to a corresponding black shade. So @ 50fps I kind of expect no issue, but at 120+, sure. It will be interesting to learn more.

As to stopping production every time you change the obvious answer is no. Know what rates you will be shooting during the day and black shade for each during AM prep. You can save multiple calibrations you know.

What do you do in the case that your shooting a video that has many many different FPS and shutter speed changes through out the day... am I supposed to stop production every time there is a change? The video will never get done.

 

[Brett Clements]

Thanks for the effort. Pro cameras are like F1 race cars.

 

[Antony M]

As to stopping production every time you change the obvious answer is no. Know what rates you will be shooting during the day and black shade for each during AM prep. You can save multiple calibrations you know.

Thanks Scott. I did not know that you can now save multiple calibrations. I will read the new manual
Which Firmware build introduced this?
Can you name the calibrations so you know which is which?

 

[Scott Crawley]

Thanks Scott. I did not know that you can now save multiple calibrations. I will read the new manual
Which Firmware build introduced this?
Can you name the calibrations so you know which is which?

I think I gave you some bad info.

I went to my menu to check (because I hadn't yet done it myself) and only found buttons for two calibration settings; one user setting and one for current fps.

Maybe I had read that saving multiple shadings was a new feature that was forthcoming? My memory is fuzzy on this. Sorry.

 

[Uli Plank]

You don't need to get too concerned about re-calibration. It's more important to get the camera up to working temp before calibration.

You should be fine with one value for regular fps/shutter speeds and one for very long exposure times (below 1/10th).

 

[Will Keir]

ADD to list:

During Shooting #7 needs more detail. 7. White Balance (easiest do do while shooting. Graeme prefers to white balance off a neutral gray card to not pick up reflected light)

I didn't find out there was an advanced window to Calc White Balance till now. If I touch and hold the Kelvin menu with the touch screen I can get into the advanced menu. Another way of doing this with the side handle is to scroll to the kelvin window and press the "up" arrow key. This will bring you to an advance menu where you can set kelvin manually, set Tint and do a "Calc White Balance."

What I am wondering is if this "calc white balance" option is the same as the user key defined side handle button "WB Calc"

 

[stu aull]

What I am wondering is if this "calc white balance" option is the same as the user key defined side handle button "WB Calc"

+1
Good Q...

Stu Aull
S-X 00199
Alaska

 

[Les C.]

ADD to list:

During Shooting #7 needs more detail. 7. White Balance (easiest do do while shooting. Graeme prefers to white balance off a neutral gray card to not pick up reflected light)

I didn't find out there was an advanced window to Calc White Balance till now. If I touch and hold the Kelvin menu with the touch screen I can get into the advanced menu. Another way of doing this with the side handle is to scroll to the kelvin window and press the "up" arrow key. This will bring you to an advance menu where you can set kelvin manually, set Tint and do a "Calc White Balance."

What I am wondering is if this "calc white balance" option is the same as the user key defined side handle button "WB Calc"

I modified #7 with your suggestions. Of course, this brings up the question of how to properly set the Tint and Kelvin manually?

What do you do in the case that your shooting a video that has many many different FPS and shutter speed changes through out the day... am I supposed to stop production every time there is a change? The video will never get done.

It's not really about frames per second... All about the exposure time. (I added this to the list)

Hopefully a future build will allow saving and loading more presets, possibly from the Flash card? Future builds should improve the time it takes to black shade.

But, for the lowest noise ... black shading is the current reality with the Epic/Scarlet.

I site no reference, these are just my recent thoughts as I get more familiar with Scarlet and the craft. I am not formally trained, or even particularly well read. One day soon I will go to the ASC Manual for some definitive answers.

As we have mentioned, exposing to the right and bringing levels down to where you want them is best. It yields the most flexibility with the widest range of grays, but the larger point was don't be paralyzed by the possibility that some of your blacks are clipping. If you do it, know exactly what you are doing and why. Have a good reason.

Think of it this way; There is no noise in clipped blacks. There is only blackness, nothing, no data. If you grade the footage properly your image will bear that out. If you don't you will introduce (or reveal) noise. You can not create data where it doesn't exist, just noise. (Which I guess counts as data, but I digress.)

More important than the presence of black clipping itself is where you place the middle, as this will determine what falls off and how rapidly it falls off.

This evening I decided to plumb the depths a bit in search of my Scarlet's noise threshold. This was not a "test" or anything formal or scientific, just a guy and his new girl getting to first base. Here is what I did.

I hung up a black chipboard, a tungsten light with CTB and warmed Scarlet To normal operating temp six feet from the chipboard. It was the same temp at which black shading was done today; 29/45c. ISO 320, 1/48 24-70L @70mm f3.5 (Sharpness was not being measured. In retrospect f4 would have been more useful just the same.) I turned on false color and used it to measure three points: no false color at upper threshold cyan, cyan at upper threshold of purple and full on purple (noise). I varied lamp focus to achieve the desired intensity levels.

My meters are not ideal for this exercise, but I used the Sekonic 758R for incident readings and the C-500R for color temp and footcandle readings, though it is not an exposure meter. The 758 is not a cine meter, so it doesnt read fc or lux, and I had to read in 1/30th shutter measurements.

Results were as follows:
no false color at upper threshold cyan- 18.1fc/1/30 f8.1 @320 5240k
cyan at upper threshold of purple- 13.2fc/1/30 f7.7 @320 5160k
full on purple (noise)- 9fc-1/30 f7.2 @320 5070k

So my Scarlet's noise threshold (Black clipping point) appears to be between 9-13fc @f3.5 and f7.2-7.7 at 320 1/30. Your mileage may vary.

I will repeat this exercise tomorrow night with straight tungsten. It should make for an interesting comparison.

Thanks a lot for your work with testing - and the graphs on post http://www.reduser.net/forum/showth...e-and-Softness&p=990426&viewfull=1#post990426

I agree that there's no noise in a perfectly clipped black.

However, the MX sensor, while not low noise at low light levels, is still sensitive enough to operate as ISO 2000 and above. So, getting perfectly clipped blacks (a zero level in R,G,B) may be very difficult to do?

I'll be interested to see your results with 3.0.6 - and I can add them to the list on noise.

I'd also be very interested to see what the rgb values for the 'clipped blacks' show in the R3D files - even more than what the levels on the LCD show.

Are they in fact clipped, and is the noise less than if you added some light (or wider f stop), and then crushed them?

Well, for your debayer settings in the transcode Half Pristine works. Full Pristine gives you a huge amount of more data for very little gain. That’s been covered by many people, not just myself. In RCX doing a first light however, you have to bounce back and forth under the resolution tab at the top to get your desired sharpening. Almost all the time I just sharpen to three at the top in the pre gamma stage and then now I use the default setting for unsharp mask and it suits me just fine. Sometimes I might even dial the unsharp mask back a little if it’s too sharp. RPPs resolve VERY sharp. I love them.

On set I use the 1:1 (image magnify) on both cameras and verify with the focus check feature since that is very close to traditional peaking. I always get focus using that. I always pull again when changing focal length on a zoom because I never assume it’s parfocal, even though mine seems to be. I usually always go for the sweet spot on my lens if I can as well. I pull tape to verify sometimes. You have to know your lens for that though...I would never pull tape with a still lens or a lens you don’t know for sure if the marks are on. Do not assume they are. All that coupled with a second check by the Scripty or someone watching a t least a 22’” monitor and you’ll be fine. Even if you don’t have a monitor you should be fine doing the other two. If your subjects are moving and going out of focus when someone is pulling, the first thing to think about is your aperture. It’s tricky to nail at first but it comes very easy once you have it.

Good info!! I added this to the post production area.

 

[Scott Crawley]

What I am wondering is if this "calc white balance" option is the same as the user key defined side handle button "WB Calc"

They are one and the same.

 

[Scott Crawley]

I'll be interested to see your results with 3.0.6 - and I can add them to the list on noise.

I wasn't really planning to do another series. My numbers are not absolute, so the difference between the resultant values for each firmware version +/- error aren't likely to add up to much of anything. It is enough to understand that the in camera histogram and false color should be more accurate with 3.0.6 than I had with v. 3.0.0.

I'd also be very interested to see what the rgb values for the 'clipped blacks' show in the R3D files - even more than what the levels on the LCD show. Are they in fact clipped, and is the noise less than if you added some light (or wider f stop), and then crushed them?

Either or both of us seem to have a fundamental misunderstanding of how this works.

It makes no difference in the absolute blacks of the end result if you Set a higher ISO and crush in post, allow the recorded signal to clip, or light to higher levels and iris down then crush in post... once you correct black levels to zero those blacks are clipped R=0 G=0 B=0 and contain no noise. The other grey levels and where they fall are of primary concern and your method of lighting and exposure is going to determine that.

Maybe this will help you to visualize what I mean.

On this graph the three points at each f-stop represent only the lower end of the signal from adequate exposure to absolute black in the recorded signal.

On the left where the iris is wide open and the least amount of light is falling on the black card, the distance between the top point and the third point (adequate exposure and absolute black) is very short. There isn't a lot of room in there for many levels of grey.

On the right is where the iris is closed and the most light is falling on the black card. The distance between those three points is rather wide. If you projected the line between the three points further up to where white clips, you would see the fullest extent of the recorded signal and there would be plenty of room in there for a full representation of millions of levels of grey. This shows graphically that only by lighting to higher levels can you get the widest latitude the sensor has to offer.

If you light to higher levels and capture information deep into shadows... if your recorded black levels never actually touch 0 and whites never max out, then you have the fullest range of grey possible in your digital neg. and therefore the most flexibility and the most noise free image to begin with in post. That is the goal of the Expose to the right method; either to set all levels to the right as far as possible without clipping white, or at least set them as far to the right as you can with the light you have.

Red wants us to get the most latitude possible from their system and that is why they advocate "expose to the right" and do it in a way that does not clip on either end.

That is how I understand things.

 

[Les C.]

If you light to higher levels and capture information deep into shadows... if your recorded black levels never actually touch 0 and whites never max out, then you have the fullest range of grey possible in your digital neg. and therefore the most flexibility and the most noise free image to begin with in post. That is the goal of the Expose to the right method; either to set all levels to the right as far as possible without clipping white, or at least set them as far to the right as you can with the light you have.

Red wants us to get the most latitude possible from their system and that is why they advocate "expose to the right" and do it in a way that does not clip on either end.

That is how I understand things.

I'm in agreement with what you are saying above, but I'm still unclear on what you were saying here
http://www.reduser.net/forum/showth...e-and-Softness&p=988383&viewfull=1#post988383

"When you attempt to inject light into places where there was none, you get noise ...... It is not incorrect to light a scene and leave parts of it in darkness. To fully illuminate a scene that should be played in darkness is incorrect."

I'm trying to get my head around that, as it seems that if you don't have enough light in the dark areas, you'll end up with noise (which may very well be fixable in post if you want to totally crush them, but un-fixable if you want to actually see into those areas). Our eyes very rarely see total black, so ... But, we may be talking in abstract about different things?

I've seen a few threads about noise in the shadows, such as this recent one:
http://www.reduser.net/forum/showth...-the-shadows-How-do-you-shoot-quot-black-quot

But ... I could be dead wrong, as my 'theory' may not work in practice for a number of reasons, that someone experienced with the reality of shooting scenes with black areas (which I'm not) can point out. It sounds like you've done a number of shoots with darkness, and maybe understand it better?

 

[Scott Crawley]

I'm in agreement with what you are saying above, but I'm still unclear on what you were saying here
http://www.reduser.net/forum/showth...e-and-Softness&p=988383&viewfull=1#post988383

"When you attempt to inject light into places where there was none, you get noise ......

Meaning if you correctly expose a black object and attempt to bring it up to grey in post, it is going to be noisy as hell. You are trying to add light where there wasn't any. The only way to brighten that would be to add some lighter colored pixels to the image amongst the black ones, and those pixels are noise.

It is not incorrect to light a scene and leave parts of it in darkness. To fully illuminate a scene that should be played in darkness is incorrect."

By that I meant if you change the appearance of a dark night scene to the point that it looks like it was shot at high noon just to avoid recording any noise then you are missing the boat.

I'm trying to get my head around that, as it seems that if you don't have enough light in the dark areas, you'll end up with noise (which may very well be fixable in post if you want to totally crush them, but un-fixable if you want to actually see into those areas). Our eyes very rarely see total black, so ... But, we may be talking in abstract about different things?

When you expose to the right so that none of your blacks touch the noise floor, then they will not appear black. They will not be noisy, but they will not be black... at least not yet. You have to then bring them down to black in post. By doing this you will not only be eliminating the noise, but you will be capturing image data all the way down to the darkest portions of the image and you can then make them as dark as you want them in post without adding noise. If you expose the same scene with blacks down to 0 IRE and decide later that you don't want it quite so dark you will have to lift it, but you will have no way to do that without adding noise.

 

[Scott Crawley]

I've seen a few threads about noise in the shadows, such as this recent one:
http://www.reduser.net/forum/showth...-the-shadows-How-do-you-shoot-quot-black-quot

If you look at Justin's before and after in Photoshop, and sample the BG in each, you will see that the net effect of his adjustments was to bring color levels in the BG down from approx. R17 G15 B13 (reddish grey) to R3 G3 B3... in balance and very near to 0, so the BG looks like a black hole. If the intent were to see detail in the BG so you might ID a curtain or whatever behind him, they would have had to expose more open or add light to the BG so that it would read higher than absolute darkness... "to the right". If they wanted a black BG devoid of detail then they exposed fine, it just was not color balanced or graded correctly and their blacks were not close enough to 0 to look black. How close to 0 IRE is black? That depends on several things like display format and display gamma and that gets out of my comfort zone. I'm not a colorist, and maybe just a half decent photographer. ;-) EDIT: Ryan's blacks in the samples he just posted are closer to 9,9,9 and look sweet.

Look at the grey values of the man's suit. In the before they average about 66,44,32, (an obviously red cast) and in the after they are about 43, 34, 34... close to 34,34,34 so they are without color, and around 35 not 0 so they are grey not black.

I hope some of this is helping.

 

[Les C.]

By that I meant if you change the appearance of a dark night scene to the point that it looks like it was shot at high noon just to avoid recording any noise then you are missing the boat.

I saw some very interesting shots, I forget where, from some super high ISO camera.

They were all shot at night in moonlight at the beach, but the shots looked just like daylight, complete with strong shadows from the 'Sun'.

What was most interesting, is that they could be displayed in post both ways, as daylight, or as moonlight, as the data was there, even for the dark shadows.

That scene had no lighting modifications at all.

But, back on Planet RED, I'm not sure of the best way to light a scene at night to minimize noise? Whether you'd want to increase the ambient light, and crush or not? Probably beyond what this FAQ was meant to cover, but ... an interesting topic, nonetheless.

From re-reading your recent clarifications, I think I mis-read what you were saying, and we're both on the same page.

You can't create data where none exists, and to get data, you need light.

As for dark areas without noise, it must be difficult to achieve, as I'm always noticing noise on BluRays in dark areas.

I added a post to your initial graph and subsequent posts to #3 in 'during shooting'.

 

[Will Keir]

Thanks for all the work keeping this thread up, great resource for the young guys at the very least.

 

[Will Keir]

A guy on this forum posted a shoot at the Santa Monica board walk at night. It was a lens tests of some really fast glass and it look good, no perceivable noise.

I saw some very interesting shots, I forget where, from some super high ISO camera.

They were all shot at night in moonlight at the beach, but the shots looked just like daylight, complete with strong shadows from the 'Sun'.

What was most interesting, is that they could be displayed in post both ways, as daylight, or as moonlight, as the data was there, even for the dark shadows.

That scene had no lighting modifications at all.

But, back on Planet RED, I'm not sure of the best way to light a scene at night to minimize noise? Whether you'd want to increase the ambient light, and crush or not? Probably beyond what this FAQ was meant to cover, but ... an interesting topic, nonetheless.

From re-reading your recent clarifications, I think I mis-read what you were saying, and we're both on the same page.

You can't create data where none exists, and to get data, you need light.

As for dark areas without noise, it must be difficult to achieve, as I'm always noticing noise on BluRays in dark areas.

I added a post to your initial graph and subsequent posts to #3 in 'during shooting'.

 

[Scott Crawley]

As for dark areas without noise, it must be difficult to achieve, as I'm always noticing noise on BluRays in dark areas.

That is a pretty good indication that the brightness levels on your TV are too high. Are you beginning to see a trend here?

This was shot entirely by the light of the two candles. The histogram showed the vast majority of the frame was in noise. It was pinned to the left. But that was OK, because I knew that most of the shot was supposed to be black.
There was a quick de-noise done in RCX, but then in Photoshop levels, curves and Noise Ninja were applied to bring it to where it is. If you crank up your monitor you can still see noise, but on something relatively correct it is almost all black. Some of it still has info though. You can see hints of books in the BG.

 

[Les C.]

That is a pretty good indication that the brightness levels on your TV are too high. Are you beginning to see a trend here?

This was shot entirely by the light of the two candles. The histogram showed the vast majority of the frame was in noise. It was pinned to the left. But that was OK, because I knew that most of the shot was supposed to be black.
There was a quick de-noise done in RCX, but then in Photoshop levels, curves and Noise Ninja were applied to bring it to where it is. If you crank up your monitor you can still see noise, but on something relatively correct it is almost all black. Some of it still has info though. You can see hints of books in the BG.

Nice shot. I think you made your point.

A shot like that would have zero added to it by adding light, and the falloff is so abrupt and I bet even in the R3D's the RGB values are very small for the blacks.

And yes, I've really got to take the time to calibrate my BluRay output!! I've calibrated my video server with EyeOne Display, but ... never got around to dealing with the BluRays.

Of course, I may be thinking of the DVDs which have such lousy compression (compared to what they could have, remember the Ultra-bit DVDs that actually used the full DVD bandwidth?) that they actually ARE very noisy.