Epic HDR

[Eirik Tyrihjel]

 

 

Well actually, again, no. Interesting choice of words, because 4K processing is still an enormous pain in the neck, just as much as it was when Dalsa tried it. At least Arri put a debayer board in the Alexa, so you aren't forced to do half the work that (in my view) the camera should be doing as a very expensive part of your post process. I suspect that whenever you cut Red stuff on your macbook, you're working in considerably less than 4K resolution and with a considerably poorer debayer than the one about which Red like to make their most grandiose claims.

 

Processing can mean all kinds of... processing. It does not necessarily refer to a 4K full debayer, and neither does it have to. When I offline on my Macbook Pro using proxies, It still technically processes these from the full 4K data, and it´s still part of the process of getting to a full debayer in online later (even on an overnight render on the laptop). Anyway it´s all nitpicking.

Over the past 30+ months I have personally shot and ehm.. processed over 120 hrs of 4K R3D files using a laptop and a Mac tower, for all kind of end results.

 

My POINT was/is: RED made working with 4K files managable, something the critics back then claimed they couldn´t do - they did it, so maybe they can deliver some smart solution in HDR as well.

[Frank Glencairn]

I suspect that whenever you cut Red stuff on your macbook, you're working in considerably less than 4K resolution ..

 

It would not make much sense to edit full 4k in a little window of a screen that is barley a quarter of that resolution.

 

Unless you got - at least - a 4k screen there is no point to edit in 4k.

With a good cuda card or two and a card that outputs 4k to a 4k screen we we are getting there - but probably not on a Mac Book today.

 

Frank

[Saul Rodgar]

Oh yeah, I edit my 16K video on my iPod, so there . . .

 

So sick of this small-penis-syndrome video resolution war. For 99.00% of individuals out there who use a RED or similar video camera, 2K workflow would be perfectly adequate for their purposes.

 

I mean, unless you are going to a film print (tiniest percentage of RED users do so, and much to everyone else's chagrin) just how many resolution Ks are necessary for frigging youtube distribution?

 

Instead we got some consumer marketing genius (Jannard) shinning a 20K up everyone's butt with his "my-sensor-is-bigger-and-with-more-Ks-than-yours" crap and most everyone just gets suckered in. :ph34r:

Edited September 14, 2010 by Saul Rodgar

[Eirik Tyrihjel]

 

So sick of this small-penis-syndrome video resolution war. For 99.00% of individuals out there who use a RED or similar video camera, 2K workflow would be perfectly adequate for their purposes.

 

I mean, unless you are going to a film print (tiniest percentage of RED users do so, and much to everyone else's chagrin) just how many resolution Ks are necessary for frigging youtube distribution?

 

Instead we got some consumer marketing genius (Jannard) shinning a 20K up everyone's butt with his "my-sensor-is-bigger-and-with-more-Ks-than-yours" crap and most everyone just gets suckered in. :ph34r:

 

I think you are way off Saul, your knowledge of the RED ONE seems to be inaccurate at best.

 

With the RED ONE you can shoot natively at 4K, 3K and 2K, even if you choose to shoot 4K there are now some beatiful workflows out there that will allow you to work at lower resolutions, and render out at any resolution you want.

 

The added benefits of shooting 4K among others are: You can go back in the future and make a 4K version if there is ever a need, you can extract a still image for whatever purpose in 4K (I do this a lot), you can reframe easily in post and not compromise on resolution, you can add a zoom in post and still stay way above 2K resolution (handy when shooting live shows), if you need to shoot at low light - downressing from 4K will help you get rid of noise.

 

Saul, I think you are making very unfair, uncorrect generalisations about who is using the RED by your assumptions about most shooting for youtube, if you bother to check you will find no lack of feature films shot on the R1 by some very respected cinematographers.

So please check out what the R1 does and who uses it before you make unfair general statements about 5000+ owners whom you never met! (And not to mention all the users who rent.)

 

RED ONE Mysterium X is an incredibly powerful tool, at a very good price, it´s not about small-penis-syndrome, it´s about choises and options. I know one thing about mr. Jannards marketing, it´s been amazingly efficient and that´s what marketing is all about. And I see very few unhappy costumers, so they must be doing something right.

There was a huge untapped/underestimated market before the R1 appeared, I can´t explain how the "others" missed it, but they did, and it was a ripe market for RED to pick.

 

What happens when Epic and Scarlet enters stage will be very interesting, with or without HDR.

[Adrian Sierkowski]

Erik, I mean no disrespect but I think you're missing what Saul is trying to say. Saul is simply mentioning that pixel count is not the be all end all in camera technology, and that for the vast majority of productions, 2K resolution (real) is adequate, not saying that over-sampling is bad, but rather that there is a predilection these days to emphasize resolution over the other, and often more important, aspects of a camera system. And, sadly, it is true that a very small minority of projects shot on any system, even film I'd say, are destined for a film print! The youtube comment, I read as a tongue and cheek reference to an "HD" distro channel; You could also say how many Ks do you really need for DVD or BluRay.

[Phil Rhodes]

if you need to shoot at low light - downressing from 4K will help you get rid of noise.

...which might not have existed in the first place if they hadn't insisted on cramming such a lot of pixels on a sensor.

Let's relate this to the real world. Under ordinary circumstances, a Red will produce a picture that scales down to HD with reasonable results. The first sensor wouldn't even really do that and it isn't ever really going to be "4K".

[Oliver Christoph Kochs]

If a given sensor has enough DR to pull 2 stops over and 2 stops under out of the same one image inside the camera, then why shouldn't it be possible to merge this (via LUT) into a 12 bit 444 image. I think that's what is done here. On the other hand (that would be a little bit more true HDR) you could write out two data streams or image sequences at once and then merge / process these two in post.

[Eirik Tyrihjel]

]Let's relate this to the real world. Under ordinary circumstances, a Red will produce a picture that scales down to HD with reasonable results. The first sensor wouldn't even really do that and it isn't ever really going to be "4K".

 

Are you saying that the film "Knowing" for instance is not "HD with reasonable results"? I find that fairly unreasonable.

 

As far as the "4K" debate, it´s a dead horse - beaten over and over.. It´s tempting but I am not going there...

 

I think RED has provided a phenomenal tool for filmproduction and people all around the world are responding to that.

[Adrian Sierkowski]

But Oliver, if the chip itself is already resolving +2 and -2 why would you need to even have HDR.. you'd not be getting anything "extra."

[Mathew Rudenberg]

I think of it in terms of RED being a fairly simple camera that can handle a fairly high bandwidth of data.

 

The simplest way to get HDR out of a camera that can shoot highspeed would be to shoot one exposure at 180 degrees, followed immediately by another exposure at a tighter shutter (for example 22.5 would give you a 3 stop difference)

 

I believe my iphone does something similar for still images.

 

This is much like shutter bracketing still images for HDR - and would probably work better for static frames, as naturally the tighter shutter shot would have different motion blur characteristics and could have a slightly different frame, both of which could cause issues when the frames are merged. Maybe these issues wouldn't be noticeable, I couldn't say without seeing a motion demo.

 

Considering they've only posted still frames this seems to be the most likely solution they've found. Also Jim posted on reduser that HDR would reduce the high speed capabilities of the camera.

 

Now excuse me while I go throw up after reading through all that sycophancy...

[Eirik Tyrihjel]

Erik, I mean no disrespect but I think you're missing what Saul is trying to say. Saul is simply mentioning that pixel count is not the be all end all in camera technology, and that for the vast majority of productions, 2K resolution (real) is adequate, not saying that over-sampling is bad, but rather that there is a predilection these days to emphasize resolution over the other, and often more important, aspects of a camera system. And, sadly, it is true that a very small minority of projects shot on any system, even film I'd say, are destined for a film print! The youtube comment, I read as a tongue and cheek reference to an "HD" distro channel; You could also say how many Ks do you really need for DVD or BluRay.

 

Thanks for the reply Adrian, I just think Saul made his point while making a lot of other not so valid points, it all became fairly hostile and ill informed imho, of course that´s just me ;-)

 

Edit: for mispellings

Edited September 16, 2010 by Eirik Tyrihjel

[Keith Walters]

The simplest way to get HDR out of a camera that can shoot highspeed would be to shoot one exposure at 180 degrees, followed immediately by another exposure at a tighter shutter (for example 22.5 would give you a 3 stop difference)

One poster on RedUser made reference to some thread where Jim Jannard supposedly ruled that out, but needless to say, he didn't provide a link. I can't find any such quote, and I've waded through weeks of JJ's quotes.

Sounds a bit like a recent Snark hunt here for the word "Scam"...

 

Something is fundamentally wrong here; as far as I know, nobody makes even a stills camera with an 18 stop dynamic range, but they're doing it at 24fps and more at about 15 megfapixels per frame???

I'll believe it when I see it, but somehow I'm getting the distinct feeling that "18 stops" is going to be to dynamic range what "4K" is to resolution, at least in the Reddi-verse.

 

All I've seen so far is one seascape photo, and the quality of that is somewhat sub-sensational, in fact it looks like a scan from a magazine, and anyway I've made lots of photos of scenes like that with a $199 digital cmaera without too many problems.

If they said: "by the eway, here's one I shot of the same scene with a $1,000 DSLR", and it was obviously struggling, well that would be different.

 

As it is, it's all about as meaningful as the resolution charts they post.

 

 

 

Now excuse me while I go throw up after reading through all that sycophancy...

You actually READ it?

You have to learn to selectively de-focus your eyes :rolleyes:

[Oliver Christoph Kochs]

Adrian, yes i think there is a difference (but not much of HDR in it) that is simply DR with a little more possibilities in post. Arri's Alexa has a dual 16 bit data handling inside the camera but records "only" 12 bit 444. If it would be able to record all this data into two ProRes movs you'd be at least close to HDR in post.

[Stephen Williams]

Are you saying that the film "Knowing" for instance is not "HD with reasonable results"? I find that fairly unreasonable.

 

 

Hi Eirik,

 

I thought Knowing looked awful, I saw it from a film print at the best screen in Zurich.

Horrible skin tones, clippy highlights, blown out skies, focus often in a strange place & any movement was unwatchable due to strobing. I would normally of walked out of the cinema however was curious as it was the first big budget 'RED' movie that I has seen.

 

Stephen

[Dom Jaeger]

 

I'll believe it when I see it, but somehow I'm getting the distinct feeling that "18 stops" is going to be to dynamic range what "4K" is to resolution, at least in the Reddi-verse..

 

 

Perhaps RED stops are different to normal ones? :P

 

At any rate isn't dynamic range somewhat subjective and dependent, like depth of field? I notice that the always sober people at Arri describe Alexa's DR as "well over 13 stops" rather than quoting an

extreme figure that may be theoretically correct but in practice unusable.

 

 

 

As it is, it's all about as meaningful as the resolution charts they post.

 

 

I agree. As interesting as it is to postulate on the possible technology used, making any kind of judgement at the moment is like trying to review a film after only having watched 5 seconds of the overblown trailer on youtube through your dial-up modem.

 

And I've promised myself never to be unwittingly sucked into visiting reduser again. Reading that thread felt like stumbling into the compound of a cult, with Jannard as some sort of L. Ron Hubbard figure surrounded by kneeling acolytes all gushing and weeping... :blink: creepy indeed

[Eirik Tyrihjel]

Hi Eirik,

 

I thought Knowing looked awful, I saw it from a film print at the best screen in Zurich.

Horrible skin tones, clippy highlights, blown out skies, focus often in a strange place & any movement was unwatchable due to strobing. I would normally of walked out of the cinema however was curious as it was the first big budget 'RED' movie that I has seen.

 

Stephen

Hi Stephen,

 

I don´t want to get into a lengthy argument concerning the movie "Knowing", but since your statements above are so incredibly different than how I experienced it, I feel at least I should make a reply.

 

I watched it on blu-ray, in our grading theatre on a calibrated projector.

So that makes me think: Is it possible that the film print is at fault? (or maybe they fixed the flaws when mastering the blu-ray) Because quite frankly on blu-ray "knowing" looks very very good (picture quality). And a brief search around the net revealed to me that the concensus among people who have reviewed the blu-ray is the picture quality, many of them award it 10 out of 10, none of those I read (about 8) mention any of the defects you mention.

 

So if you have the chance and want to spend more time on it, please check it out on blu-ray and see if that makes any difference.

 

Eirik

 

edit:typo

Edited September 16, 2010 by Eirik Tyrihjel

[Adrian Sierkowski]

I'm almost certain they changed a few things when mastering to BluRay... And, defects on a big screen -v- a smaller screen are pretty remarkable. BluRay almost always involves getting rid of data and a lot of noise reduction just to fit the spec. Now, not having seen Knowing in theaters or on BR, I can't speak to that specific film; but I can say there is an appreciable difference between most films I see on BR and on theaters, even if they were digital to begin with (2012 comes to mind)

[Stephen Williams]

Hi Stephen,

 

I don´t want to get into a lengthy argument concerning the movie "Knowing", but since your statements above are so incredibly different than how I experienced it, I feel at least I should make a reply.

 

I watched it on blu-ray, in our grading theatre on a calibrated projector.

So that makes me think: Is it possible that the film print is at fault? (or maybe they fixed the flaws when mastering the blu-ray) Because quite frankly on blu-ray "knowing" looks very very good (picture quality). And a brief search around the net revealed to me that the concensus among people who have reviewed the blu-ray is the picture quality, many of them award it 10 out of 10, none of those I read (about 8) mention any of the defects you mention.

 

So if you have the chance and want to spend more time on it, please check it out on blu-ray and see if that makes any difference.

 

Eirik

 

edit:typo

 

I think it's quite likely that RED M footage from before 'new color science' does not transfer that well to film, it's a shame. I saw Ben Button a couple of weeks before in the same theatre it was the best looking 'digital film' I ever saw.

I have seen RED footage looking very good on the Sony 4K projector @ IBC, I was expecting to be impressed by 'Knowing' however I was not.

[John Sprung]

It's one of those weak link in the chain things. It's easy to see that there's a weak link. Figuring out which one it is, though, is much harder.

 

 

 

 

-- J.S.

[Keith Walters]

I'm almost certain they changed a few things when mastering to BluRay... And, defects on a big screen -v- a smaller screen are pretty remarkable. BluRay almost always involves getting rid of data and a lot of noise reduction just to fit the spec. Now, not having seen Knowing in theaters or on BR, I can't speak to that specific film; but I can say there is an appreciable difference between most films I see on BR and on theaters, even if they were digital to begin with (2012 comes to mind)

They had more than enough time to massage Knowing frame-by-frame for the Bul-ray release if they wanted to. Did they re-do the "summer-for-autumn" colouring of the Australian forest so that it looked at least vaguely like a real Northern forest in Autumn? On the version I saw, it looked like someone had sprayed the whole forest with Roundup®. Apparently nobody realized that real deciduous trees actually turn all sorts of colours, and some actually stay green.

 

Although apart from the producer, his immediate crew and family, and of course all the self-aggrandising RED owners, I'm not sure who would care, since it was such a crap movie anyway.

 

The question is not just whether a digital camera can produce an acceptable result, it's how much the amount of extra post-production required to get that result erodes the alleged benefits of shooting digitally.

 

It will be interesting to see how Star Wars episodes II and III scrub up for Blu-Ray release. I strongly suspect Episodes IV, V, and VI are going to look embarrassingly better.

 

Like Superman really....

[Emanuel A Guedes]

http://reduser.net/forum/showthread.php?t=49940

[Saul Rodgar]

Erik, I mean no disrespect but I think you're missing what Saul is trying to say. Saul is simply mentioning that pixel count is not the be all end all in camera technology, and that for the vast majority of productions, 2K resolution (real) is adequate, not saying that over-sampling is bad, but rather that there is a predilection these days to emphasize resolution over the other, and often more important, aspects of a camera system. And, sadly, it is true that a very small minority of projects shot on any system, even film I'd say, are destined for a film print! The youtube comment, I read as a tongue and cheek reference to an "HD" distro channel; You could also say how many Ks do you really need for DVD or BluRay.

 

 

Yessir!

 

I really do think that online distribution is gonna be the norm, at least in the US. Maybe not necessarily youtube. Netflix uses a slightly higher bit-rate encoding, and it sure looks like the winner of the video distribution wars. Only Red Box seems to be putting up a fight in the states. But, to be sure, even Blue Ray does not absolutely need a 4K master to look very good.

[Mathew Rudenberg]

http://reduser.net/forum/showthread.php?t=49940

 

Thanks for the link, I'd say this strongly supports my theory, although I think I got it the wrong way around - it's the tight shutter first, not second :)

 

Look at this image:

 

 

This is a punch in still from the HDR video they posted. I took it from the fastest part of the pan.

 

Notice how the brightest information (in the lights of the casino) consist of a bright point source followed by a blurred tail. (the pan was right to left for those of you who haven't watched the vid)

 

This, to me, indicates that two exposures were taken - first a fast shutter speed one to capture highlight detail. Due to the tighter shutter there is very little motion blur.

 

The second, (probably) 180 degree exposure captures the shadow detail with more motion blur, hence the trails.

 

This could produce some pretty interesting effects if one could effect the two shutter speeds individually... Imagine shooting a 60 degree shutter exposure followed by a 300 degree and composited together - you could possibly get a sharp image followed by motion blurred trails... interesting...

 

As for the HDR - unimpressive example - Vegas at night simply looks good, even on an hvx, I can't see any increased DR. Similarly that front lit still of a lake doesn't indicate extended dynamic range to me.

 

If they post a moving shot with bright sunlight and deep shadow both holding detail I might change my tune pretty damn quick though. I want to believe, I just can't help being skeptical.

 

Thoughts?

[Adrian Sierkowski]

Good catch on the motion looks Matt; I was looking @ the same clip trying to see if the motion looks bad, and I will say that to me, and perhaps biasedly so, it looks "differant," almost "wrong," in the way the motion of faster objects read, such as that fire truck, and again, doesn't look extreme in terms of DR to me either. I believe this was "easy HDR," but honestly, if the camera can't record that by default... then I'd say they need to work on their basic sensor's DR a lot more before trying to do HDR.

Here's a link to the video too, for those interested:

 

http://reduser.net/forum/showthread.php?t=49940

 

Now, all that being said, this would grade pretty ok; says me but the "point source," and motion trail is a bit, well worrisome to me. On a locked off shot, I think it could do well; but add in motion (too much motion, or even commotion!) and I feel this would fall to shambles pretty quickly.

 

I would also love a pro res to look at, or SOME kind of comparison with NON HDR footage of the same shot! This can't be that hard to do; pick an empty city street, day or night, shoot a few seconds then turn on HDR and shoot a few seconds more.

[John Sprung]

This, to me, indicates that two exposures were taken - first a fast shutter speed one to capture highlight detail. Due to the tighter shutter there is very little motion blur.

 

The second, (probably) 180 degree exposure captures the shadow detail with more motion blur, hence the trails.

 

Hmmm .... We're getting bright points from the short highlight exposure, but dim trails from the longer shadow detail exposure.... Doesn't that sound backwards? Shouldn't we be getting dim from the short exposure and bright from the long one? .... If in fact that's what they're doing?

 

The bright point/dim streak thing looks a lot like an old old time tube TV camera. I'm not sure whether or not it's related to the HDR technology. Could it be that the clearing of photosites between frames doesn't happen completely in severely overloaded areas?

 

 

 

-- J.S.