Build 15 & 16 news...

[Georg Misch]

Hi!

After doing some tests with two DP's over the weekend I have to say that there needs to be sme kind of metering for when the sensor clips. I believe this is absolutely necessary in order to achieve the highest dynamic range without the danger of clipping.

Georg

 

[Taylor Morrison]

Putting aside waveforms, histograms, etc. for a moment...is there a rough under and over exposure idea we can take into the field/set with us. For instance I new absolutely that when shooting good kodak 16mm stock I had 2 1/2 to 3 stops over my base exposure (using spot meter) before blowing out into white. As well I had 3 stops below t stop for shadows. Of course we could then fool all we wanted in film to tape but at least it was all there to begin with. For me... forget the histogram and let me know what is the under and over for the sensor. I'm getting new batteries for the spot meter and leaving the monitors at home (assuming I get my EVF!)

 

[Greg M]

Hi!

After doing some tests with two DP's over the weekend I have to say that there needs to be sme kind of metering for when the sensor clips. I believe this is absolutely necessary in order to achieve the highest dynamic range without the danger of clipping.

Georg

the zebras will tell you this info...I have no problems keeping the highlights from clipping.

 

[jbeale]

I'm curious- pretty much all DSLRs can record in both Raw and JPEG formats, and they do show histograms. Are they showing the RAW or JPEG histogram? I think they all show the JPEG histogram (that is, after white balance). You do a white balance and fill the screen with a white card, you see just one spike. If it was RAW you'd see three for the R,G,B channels at some other white point.

But as everyone is saying- it's obviously useful to have an option for your traffic light to show you the "true" (raw) clipping point. Or both! Say a half-circle light for REC709 clip, and color the full circle for RAW clip.

I don't know the internals, but I'm guessing this must be the easiest of all the requested features to actually implement. I don't think it requires much additional processing over what the camera must be doing already.

 

[Barry Green]

the zebras will tell you this info...I have no problems keeping the highlights from clipping.

No, the zebras will tell you when the processed REC 709 signal is clipping. Not when the sensor is clipping.

And there's a problem there anyway -- the zebras don't let you go to full 109. You can set the high zebra to 108, but not to 109, so you can have blown-out areas in your image with no indication that it's clipping! Set your zebras to the highest settings (105 and 108) and point a light into a white wall, and overexpose it, and look at what you get: you'll get a ring of zebras around the hot spot, with no zebras in the hot spot.

 

[Barry Green]

A similar comment about focus, the tools are working, so don't expect to see major changes in these areas.

This is profoundly disappointing. This is, perhaps, the most disappointing news I've heard come from the Red team yet. Many of us do not feel that "the tools are working" to the degree that we need.

Having said that, feedback and suggestions are always welcome.

We want raw monitoring for focus and exposure. We want pixel-accurate monitoring for focus, and we want accurate peaking that kicks in only when something is very crisply in focus at the 4k level. And we want to monitor what the sensor's doing, not some unrelated and practically arbitrary 709 gamma interpretation of what the sensor's doing.

I'm tempted to replace the words "want" with "need." Please reconsider.

 

[Chris Gearhart]

But don't you then have the clipping indication in that hot spot? You can also use the "traffic light" for RGB, or the luminance histogram, etc. etc. Am I being stupid here?

EDIT--oh, just reread to see the "sensor" and "raw" comments. .

 

[Barry Green]

But don't you then have the clipping indication in that hot spot?

There is no indication in the hot spot.

You can also use the "traffic light" for RGB, or the luminance histogram, etc. etc. Am I being stupid here?

Can't use any traffic lights, as I'm still waiting for my EVF to show up.

My HD-SDI monitor doesn't show any of that info.

 

[Chris Gearhart]

We want pixel-accurate monitoring for focus, and we want accurate peaking that kicks in only when something is very crisply in focus at the 4k level.

I dunno, I thought that way too, but I'm kinda liking the "analogue" effect of the peaking--similar to what one does by eye. It feels . . . more organic to me.

 

[Finner]

This is profoundly disappointing. This is, perhaps, the most disappointing news I've heard come from the Red team yet. Many of us do not feel that "the tools are working" to the degree that we need.

I have not seen hardly any people asking for the focus features you say "many" feel they need. It is only you and a small handful of others Barry that want this. For the large majority the current focus tools are working great and exceed those of most other high end cameras on the market.

 

[Barry Green]

I have not seen hardly any people asking for the focus features you say "many" feel they need. It is only you and a small handful of others Barry that want this. For the large majority the current focus tools are working great and exceed those of most other high end cameras on the market.

I'm not clamoring anymore, I've been shouted down enough. But whenever it gets raised, others do voice their wish for it too.

What I don't get is: if it's possible, why not do it? Who does it hurt, to give us this option?

Furthermore, if a 720p display is adequate to show all the detail you need, then why do we need 4K at all? Why not just go with 720? And, if 4K is necessary (specifically because it has more detail), why wouldn't you want to be able to see that to know that you've focused it adequately?

Anyway, I've said my take, I've said it's disappointing, I've said why I think it's necessary, but I'm not going to perpetuate some sort of campaign, so this is the last of it. Stuart says they're not gonna do it, then they're not gonna do it.

 

[Stuart English]

Zebras and False Color

Zebras and False Color

No, the zebras will tell you when the processed REC 709 signal is clipping. Not when the sensor is clipping.

And there's a problem there anyway -- the zebras don't let you go to full 109. You can set the high zebra to 108, but not to 109, so you can have blown-out areas in your image with no indication that it's clipping! Set your zebras to the highest settings (105 and 108) and point a light into a white wall, and overexpose it, and look at what you get: you'll get a ring of zebras around the hot spot, with no zebras in the hot spot.

As long as the REC709 clips first, then you have avoided sensor clip. As Graeme has explained before, that is a deliberate implementation on our part.

There may be a bug with regard to 109 not showing up, (noted, thanks for the feedback) but the likelyhood that a value of 109 is there and is not accompanied by values such a 108, 107, 106 etc is extremely low.

 

[Thom Steinhoff]

1:1 focusing is a big one. If it's a lot of CPU to ask for, I would be happy if it was even only available when the camera was not recording and when you hit record it switched back to 2x.

This would handle the vast majority of narrative users as we always confirm focus with the camera idle anyway.

ENG types wouldn't be happy but at least you get the rest of us off your backs!

My eyes do suck so I struggle with the LCD for critical work. The VF should be better but I would really like 1:1.

 

[Jeff Kilgroe]

We want raw monitoring for focus and exposure. We want pixel-accurate monitoring for focus, and we want accurate peaking that kicks in only when something is very crisply in focus at the 4k level. And we want to monitor what the sensor's doing, not some unrelated and practically arbitrary 709 gamma interpretation of what the sensor's doing.

I agree on the 1:1 zoom (and a way to pan the zoomed area or specify that area) is a must. This is going to be especially critical for users who intend to use still lens mounts / SLR lenses with their cameras. There is just no other way to hit accurate focus in a consistent manner when many such lenses only have a 5-degree barrel turn to focus between 5ft and 30ft.

However, other monitoring isn't that simple to do with RAW data, not as far as I'm aware. Like I said above, there would probably be a need for some sort of bayer filtering to see a meaningful 1:1 monochrome (luminance) image on a 1:1 pixel zoom. To see clipping and other such image characteristics, the RAW data needs to be processed into some color space and some sort of curve applied. In this case RED is using REC 709. Will we have more options once the camera allows us to upload our own LUTs? By just monitoring linear RAW data, I think the only thing that can be checked for is individual photosites that are entirely over-loaded. And that could potentially be displayed via histogram. For over-exposed photosites, would it even be possible to overlay zebras on a processed image and have them based off of RAW values?

I'm hoping someone can tell me I'm completely wrong about this because I would like all these features too. But there are limits to what can be processed in-camera and how that data can be presented to those operating the camera.

 

[Joel Kaye]

I have not seen hardly any people asking for the focus features you say "many" feel they need. It is only you and a small handful of others Barry that want this. For the large majority the current focus tools are working great and exceed those of most other high end cameras on the market.

Then you haven't been watching. There are at least as many people on these boards concerned about focus assist as the Infrared Problem for instance. Face it, a few people here do the talking for a lot of lurkers when it comes to all this stuff... whether it be crummy parts in the production pack or whatever.

Just about EVERYONE wants 1:1 focusing.

Even Gibby would like to see it. And anyone who's ever used the EX1 focusing system wants it. The fact is RED is getting it's ass handed to it by lower end cameras when it comes to focusing tools. This is a digital camera, not a film camera. It needs digital focusing aids.

You're the ONLY person other than Stuart I've ever seen make this assertion the focus tools are OK in all my time traveling the RED boards. But I don't read everything... feel free to point to anyone else that thinks having a 1:1 focus preview would be a waste of time.

Additionally, I don't think I've seen anyone say they actually LOVE the current focus overlay/assist. I've seen dozens of people on other boards say they love the JVC and Sony focus assists. That's the cold, hard truth. And somebody over at RED needs to smell the coffee.

I finally got to use the RED focus overlay this weekend on many types of scenes. I left it on 100% of the time and adjusted it heavily to see if I could get it to work for me. I constantly found myself just focusing off the LCD instead. The focus overlay simply doesn't work very well. I can think of a few things that would make it better, but the whole idea is so wrong I don't want RED wasting time working on it. Don't throw good money after bad. It's a bad idea. Replace it with the EX1 style focus assist.

They also need 1:1... you could use that to set up your lenses perfectly if nothing else. I know you're a fan of perfectly set up lenses.

Operators focus more than anything else. Period. Those tools have to be the best.

 

[Joel Kaye]

As long as the REC709 clips first, then you have avoided sensor clip. As Graeme has explained before, that is a deliberate implementation on our part.

OK, that makes sense to me. But I would like a way to brighten a dark scene to see it better on a monitor (in the way that raising the ISO does currently) but WITHOUT changing the ISO that the exposure monitors are reading.

I really like all the exposure tools you guys have put in. I think getting to them is not supereasy at the moment.

Here's my suggestion for really making the interface usable for many different types of operators.

*Make the user preset buttons behave like car stereo buttons. Hold a button in for 4 seconds and have it memorize the current screen display. Allow those settings to be saved to SD card in XML.

I think you could alternatively get this functionality by assigning an on/off status to all the possible screen display items and have users just edit an XML file, stick it on the SD card and have the camera load those settings. Obviously there are things like waveform monitors where only one type could appear at a time.

The key is all 5 user buttons should really be capable of altering the entire display as opposed to toggling one item. Then you wouldn't have to read everyone's mind and guess how they want to operate. Everyone would set up their own style, save it to the SD card and put in into any RED they are operating. Kewl.

I'd like to see False Color and a histogram at the same time. I like that Zebras appear even when False Color is on. I definitely want False Color on a button to I can toggle it quickly. It's a great exposure aid.

 

[Finner]

Operators focus more than anything else. Period. Those tools have to be the best.

Get a focus puller and get him to bring a $5 tape measure. Problem solved.

As much as some people want this camera to be a one man band ENG camera it simply is not. It's "Red digital CINEMA", not "Red digital ENG". Cinema procedure requires at least 1 camera assistant to work most effectively and proficiently, a red needs a camera assistant. Gibby has even stated many times that the red is not an ENG camera and that it fits more with a EFP or Cine work style. EFP requires some sort of crew even if it is just small and minimal. The first person on the list for that crew should be a focus puller with hopefully enough experience in other areas that they can help grip/elec on the projects. It also allows the DP/Op to light while the focus puller takes care of getting the camera all set up and ready saving lots of time. This camera is not a little prosumer cam and requires different procedures and crew to work smoothly.

If a camera assistant does not fit in the budget of work people do then a red camera would be a poor choice for that level of production work.

 

[Brent J. Craig]

As long as the REC709 clips first, then you have avoided sensor clip. As Graeme has explained before, that is a deliberate implementation on our part...

That's great, but many of us don't want a professional camera to hold our hand. Being able to confidently put something 'right on the edge' without going over is very important.

 

[Axel Mertes]

OK, that makes sense to me. But I would like a way to brighten a dark scene to see it better on a monitor (in the way that raising the ISO does currently) but WITHOUT changing the ISO that the exposure monitors are reading.

I really like all the exposure tools you guys have put in. I think getting to them is not supereasy at the moment.

Here's my suggestion for really making the interface usable for many different types of operators.

*Make the user preset buttons behave like car stereo buttons. Hold a button in for 4 seconds and have it memorize the current screen display. Allow those settings to be saved to SD card in XML.

I think you could alternatively get this functionality by assigning an on/off status to all the possible screen display items and have users just edit an XML file, stick it on the SD card and have the camera load those settings. Obviously there are things like waveform monitors where only one type could appear at a time.

The key is all 5 user buttons should really be capable of altering the entire display as opposed to toggling one item. Then you wouldn't have to read everyone's mind and guess how they want to operate. Everyone would set up their own style, save it to the SD card and put in into any RED they are operating. Kewl.

I'd like to see False Color and a histogram at the same time. I like that Zebras appear even when False Color is on. I definitely want False Color on a button to I can toggle it quickly. It's a great exposure aid.

That was the reason I was asking for an autogain B/W image!

How does the DOP & AC work?

First they adjust exposure, create the color & look they like (or the director wants them to like...).

Then they just want to frame the image, during the shoot. For that goal its much more wise to use an pure frame by frame autogained image, best in black & white to maximize contrast.

On a grayscale image you can then add - if wanted - a JVC style color focus assist overlay.

Maybe the curve should be adjustable to increase that contrast during a shoot to help the DOP&AC to get the take how they want it.

I rarely see someone adjusting iris during a shoot, but focus and framing is the story, 99.9% of the time.

A color "prediction" of the results, including a histogram "prediction" isn't really helpful except for the video village with the director or customers.

I also do not agree that its good exposure if the REC709 amples don't switch on, if we have still more headroom left that we can not judge about. As a post production guy I'd say I want 95%+ used exposure range of the sensor capabilities, because there is our sensor lattidue and 11.x+ f-stops we all want to process in post. That all means nothing else than: RAW histogram, RAW amples. We want full sensor control, thats it.

If we wanted a REC709 only camera, there are soooo many other options, honestly.

I think RED should reconsider some recent statements, beside all the loved ones we cite every now and there when meeting up with non-REDish people to convince them

Now convince us, or get convinced yourself, by the users.

Cheers,
Axel

 

[Joel Kaye]

Get a focus puller and get him to bring a $5 tape measure. Problem solved.

I think you're missing my point. This isn't about operating alone.

I still want to verify he's nailing it. That's what 1:1 and focus assist do. Especially if you can view them in playback or on a paused frame. That's a big part of what we're talking about here.

1:1 is faster and more accurate than tape. Raise a glass to tape.