SCARLET / EPIC swiss army knife camera details announced

[ZazaCast]

New article:
http://www.studiodaily.com/main/news/headlines/10215.html

 

[Denis Haineault]

Sony knows very well how to make CMOS. But in the paraphrased words of Sony's own Juan Martinez, "If you want the best quality for digital cinema applications, you want a CCD. But if you're making a small handheld camcorder, that's where CMOS is more useful."

But is this because CMOS technology cannot reach/exceed CCDs?

I have read about the differences between the two technologies (from you Barry) and understand the strengths/shortcomings for both, but I'm under the impression that this comment you are quoting is becoming less and less valid, and will eventually (soon?) be unwarranted.

Is my understanding incorrect?

 

[Otis Grapsas]

I recently tried a very big global shutter CMOS sensor. No roling shutter issues but even though pixels were huge, the shadow saturation, the dynamic range and the overall spectrum was not very good. I didn't have much time to optimise it but I don't think it could be improved. They are still working on it becuase the bayer version was just introduced and the board was a very early engineering sample with lots of bugs, so who knows!

Btw, its monochrome brother was very good, but I doubt it can be used in a camera that costs less than 40,000 euro. The sensitivity was also comparable to CCDs 4-6 time smaller in area.

Anyway, don't expect global shutter cmos in videocameras in the next 2-3 years at least.

 

[Barry_Green]

But is this because CMOS technology cannot reach/exceed CCDs?

We will see. Dalsa and Sony are two of the biggest chip manufacturers, both have digital cinema divisions, and both chose CCD for their imagers.

CMOS is not new, it's been around for 40 years. CCD has basically always been more suitable. For big chips, it still is. You just can't cram the same # of pixels onto a small CCD chip; if it's maximum pixels you need on a small chip, that's where CMOS's primary advantage comes into play.

Sony likes to position it like CMOS is inevitable; that remains to be seen. The vast majority of 2/3" professional cameras out there are CCD and there certainly doesn't seem to be any big rush to CMOS there.

Red loves CMOS, clearly, and CMOS does allow certain things that CCD doesn't (like windowing the sensor for a faster readout, which is how they get 4K @ 30fps and 2K @ 120fps). For what Red needs to do, CMOS is the only practical choice. But for ultimate performance in digital cinema, Sony and Dalsa are still choosing CCD.

 

[Postmaster]

For the record: Dalsa has not sold enough cameras to make any profit.
Now they are try to sell the Dalsa Digital Cinema Assets to Arri.

Frank

 

[Dvader]

To be honest, I am surprised about the lack of talk and mention of the 28k camera.... hello? Is this thing on? Am I the only one who has seen this?

28k is ridiculous. I bet RED will not produce this camera or will bankrupt in pursuit of building the model. It's just not needed. Just a waste of RD, time and money.

I see mad reasons for 28k, the number 28 isnt as much what I mean, 16k/24k/28k they'd all be gravy.

I see the giant resolution really changing things in filmmaking, allowing more creativity and freedom, which also could be used to be more lazy and less skilled. Kind of the same trade off we've always had with new and better technologies. It reminds me of NLE editing, it has helped improve cinema tremendously while it has also allowed for many very bad projects.

Storage and CPU may seem like a problem now, and perhaps for a while the only people capable of handling it will be studios, but with the rate of chip advancement and the always dropping prices of storage I don't see this problem lasting for too many years.

It gets me thinking about 10-15 years from now and how changed things could be. Its great to know exactly what you want or to find exactly what you want while you are shooting while you're shooting in the field, but the crop ability of something like 28k really makes my head spin when I think about the new possibilities of editing, which I think as do many others, is the essence of cinema.

 

[ThereIsNoLimit]

But you have to remember that if you crop a wide shot to a close-up the audio is going to be messed up. The mic will be far away even on close-ups and the perspective is not what it should be. Audio guys are going to hate that 28k.

 

[Dvader]

You're right in certain situations the audio will be tricky for dialogue, but if you're aware of the possibilities you can then start working around them, lavs would easily do the trick in most cases I bet there are other ways too. You could crop to a hand smoking a cigarette and lay the audio over it if you decide hey I want to isolate that cig and never had thought about it on set, just a primitive example.


I see a lot of things being possible in the visual realm of the editing, dialogue scenes could be changed too but I'm even more excited about pure visuals, one that comes to mind would be an epic battle scene between giant armies.

 

[Denis Haineault]

Thanks for the response Otis and Barry

 

[Otis Grapsas]

In most production situations, options and capabilities are sometimes keeping the project behind. The purpose of filmmaking is to decide what your closeup will be, light for it, get the best quality audio you can get with it and get a performance that is finetuned to that closeup if your talent is capable. If you start asking for infinite post options you are reducing prouctivity and creating problems for the rest of the crew and the actual product.

Also, most theaters would be glad to have a guaranteed resolution of 500lines per pixel height. Reference film projection tests estimated an 700-750 lines maximum. You can bet a half hd 960x540 uncompressed source will rock at your local chain theater and be better than the projection system. On the other hand the 720x480 compressed and heavily chroma sampled source (400 lines luma pph and down to 100lines horizontal chroma pph) coming from a dv camera will look soft in technical exaimination. The audience will not mind though, that was proven many times. How much resolution do we actually need really?

 

[Postmaster]

Here´s some nice reading on that subject:

http://www.etconsult.com/papers/Technical Issues in Cinema Resolution.pdf

Frank

 

[combatentropy]

Thanks. Sounds like 2K is as good as any release print. 4K will be a little nicer.

 

[Otis Grapsas]

Here´s some nice reading on that subject:

http://www.etconsult.com/papers/Technical%20Issues%20in%20Cinema%20Resolution.pdf

Frank

Thanks Frank. I wasn't aware of this test. A sharp high mtf 720p source appears to be capable of an excellent film release. I have done a series of tests using a digital filter that drops mtf to emulate the resolution of a 1280x720p sensor sent through the film release process. These used HD material, film transfers and actual commercial video releases of hd contect. I can't see a loss of resolution when the mtf limiting filter is engaged in most of the material. The interesting thing is that after working in uncompressed HD for a while, films look a little soft in cinemas With HD we are certainly gathering a lot more than what can be seen in a release print projection. This is also true for 35mm film shooting.

 

[arrestthisman]

Oh, I'm not worried about them being "safe." Maybe I didn't articulate this fully, but I'm simply worried that the bigger companies will copy the RED strategy. It's not the monopoly worry. It's the lack of technological diversity. Good post tho Otis. I agree with everything there.

Speaking of different technology. I just saw some footage from the new Ikonoskop dII. I was SO smitten with this camera before... NOT ANYMORE. I thought it looked very mediocre.

I think this is part of the single sensor problem. (Note, I do not know nearly enough about the tech involved to make a "call," this is simply what I think/guess.) It seems the single sensor cameras need to compensate for the bayer/stripe/etc process by oversampling, and recording at the highest possible quality format.

I don't think we've reached a point where we have a tech that is superior; only Different. Seems like three-chip cams with CCDs, can get away with more compression, etc. because they begin with a more complete image. I also don't buy the prism argument with Nattress.

The proof is in the pudding. And the pudding all looks far more similar than you'd think. All of this amazing RED specs amount to a better image. Not necessarily a new league. I think they have an excellent cost effective method for obtaining malleable, quality images. So far it's not the Holy Grail.

And the 28k camera is just ridiculous. I'm sure there will be couple of people who may benefit from this; but jesus christ 291MP. WTF?

My cent.

 

[Otis Grapsas]

Arrestithisman,

We don't expect any miracles soon in imaging. Some sensors are aproaching the limits of physics for their pixel size. I think we have enough technology to produce good results these days. We know the limits of existing technology such as smear, rolling shutter artifacts, I-frame and GOP compression so we can chose the compromise we actually prefer or uncompressed with its own workflow issues.

We know 3ccd has its advantages and flaws and that bayer results can be good or bad depending on what you do with it, but bayer is very mature these days with all the DSLRs out there and these don't appear to have any image quality problems, so bayer is not to blame. It samples chroma 2:1 (just like distribution) but the luma does not have to be soft. You can get 1000lines from an 1000line sensor today.

The sensor is not a problem with Ikonoskope because it is quite capable. The product is still in development so it will be improved and finetuned. There are details like antialiasing filters, infrared filter quality, color matrix decisions, gamma etc that can change the output a lot. It's also pretty subjective. Check the attached low light image. Is it good or bad? You will get different opinions from different people based on experience, background in video or film etc. It's not an absolute science.

I don't think technical specifications on imaging are the answer. I believe most people involved seriously into video are more concerned about the aesthetics of the recorded image and how well the camera integrates with the way they work and the tools they use.

 

[Dvader]

The purpose of filmmaking is to decide what your closeup will be, light for it, get the best quality audio you can get with it and get a performance that is finetuned to that closeup if your talent is capable. If you start asking for infinite post options you are reducing prouctivity and creating problems for the rest of the crew and the actual product.

While I don't agree with your opinion onthe 'purpose of filmmaking', but even if that were the case, I find it fascinating that things are changing and that wouldn't have to be the purpose anymore.

Reducing productivity is subjective, infinite post options, kind of like color correction have certainly helped the majority of filmmakers and perhaps hurt the productivity of a few, I don't see why this would be any different. And I'm not sure how expanded post options would hinder the crew as their job is already done by the time you're doing post, and the actual product very well may turn out even better.

If resolution like this does ever come to fruition, which I don't see it capable of not, at least in a 10-15 yr span, fundamental ways of fillmmaking will change. The art of cinema is only about a hundred years old so we're only breaking the surface with the possibilities, and I see this as another new and very interesting step forward.

 

[Otis Grapsas]

By post options I don't mean the normal capabilities we get in digital post. If I believe I actually need wider shots on everything just to have reframing options, I will end up spending a lot of expensive production time/money for what could be inferior results. Then I will waste more time in post for decisions that should be made before or during production. A decision in production is worth ten decisions in post.

In production of any work of art or construction a professional knows what he/she wants and when to finalise a part of the work in order to move forward. If you are mixing audio, you have to decide that the mix of individual tracks in a submix if finished and render it at some point, in order to focus on the big picture instead of endlessly finetuning of 100 tracks. In an infinite post option world, even the script itself would never be finalised. Not a rare phenomenon in filmmaking I might say. This road is very dangerous even for the great masters of this art working on huge budgets. It can lead a low budget feature to failure.

 

[brianluce]

In most production situations, options and capabilities are sometimes keeping the project behind. The purpose of filmmaking is to decide what your closeup will be, light for it, get the best quality audio you can get with it and get a performance that is finetuned to that closeup if your talent is capable. If you start asking for infinite post options you are reducing prouctivity and creating problems for the rest of the crew and the actual product.

Also, most theaters would be glad to have a guaranteed resolution of 500lines per pixel height. Reference film projection tests estimated an 700-750 lines maximum. You can bet a half hd 960x540 uncompressed source will rock at your local chain theater and be better than the projection system. On the other hand the 720x480 compressed and heavily chroma sampled source (400 lines luma pph and down to 100lines horizontal chroma pph) coming from a dv camera will look soft in technical exaimination. The audience will not mind though, that was proven many times. How much resolution do we actually need really?

Thank you. 28k feels like a marketing ploy at the moment -- like these small countries scrambling to build the tallest skyscraper.

I'm a RED fanboy too, I love the new line. But 28k? come on. 1080p is good enough for George Lucas.

Btw Mr. Grapsas, I've been following your posts for a while now. Do you have a link to your project? For me, RED's revolution is their pricing -- so much value/dollar. From the sound of it, you've taken the revolution to another level.

 

[Dvader]

By post options I don't mean the normal capabilities we get in digital post. If I believe I actually need wider shots on everything just to have reframing options, I will end up spending a lot of expensive production time/money for what could be inferior results. Then I will waste more time in post for decisions that should be made before or during production. A decision in production is worth ten decisions in post.

In production of any work of art or construction a professional knows what he/she wants and when to finalise a part of the work in order to move forward. If you are mixing audio, you have to decide that the mix of individual tracks in a submix if finished and render it at some point, in order to focus on the big picture instead of endlessly finetuning of 100 tracks. In an infinite post option world, even the script itself would never be finalised. Not a rare phenomenon in filmmaking I might say. This road is very dangerous even for the great masters of this art working on huge budgets. It can lead a low budget feature to failure.

You have valid points for sure. And maybe I should say that I don't think the 28k framing options would apply to every project, but they certainly could to some and to people who want to try new things.

Stanley Kubrick always used to say "I don't know what I want but I do know what I don't want" So I don't really agree with the idea of knowing what you want, but I'm sure others will feel different. Also Kubrick worked tirelessly on his screenplays, researching and pushing other writers to their limits, only to ultimately continue rewriting while on set.

I hope it's not just a marketing ploy, is 1080p good enough, sure i guess, but who wants to settle for good enough?

 

[Otis Grapsas]

Btw Mr. Grapsas, I've been following your posts for a while now. Do you have a link to your project? For me, RED's revolution is their pricing -- so much value/dollar. From the sound of it, you've taken the revolution to another level.

We are working on the site. I don't know if it will be cheaper than a 2/3" Red configuration of the new line, costs are constantly changing due to usd. The sure thing is that 2009 will provide a polyphony of digital cinema products from many manufacturers at very affordable prices.