Sensory Overload

[Michael Olsen]

What I couldn't add up was "how on earth can you actually do 3k or 5k RAW". How do you manage the data...

With equipment that is both large and extremely expensive at every step in the process, until it is compressed in some manner. Which is why RED is sticking with REDCODE. Trying to deal with 4K or 5K RAW files can really stretch production resources and limit filming ability (a 4K or 5K RAW recorder is not a small or light piece of gear).

Guesswork math: A 16-bit 5K image is going to be about...60MB, conservatively. That comes out to around 1.4GB of footage a second, or 84GB a minute.

Fortunately, for that that "need" the true RAW images, RED will have a 2.5:1 compression ratio that will allow for mathematically lossless recording once media becomes fast enough to record those data rates. This offers a great solution to keep file sizes down while still giving 1:1 sensor information.

 

[Pietro Impagliazzo]

And how are you going to operate this camera without it doing an internal debayering & downconversion in real time to RGB for a monitor or EVF, probably to 720P or 1080P? How are you going to record this 2K uncompressed RAW output on your own if your budget is just $2000 for the camera? Once you start tossing in simple things like that -- real-time debayering / RGB outputs, and recording options, it isn't just a box with a sensor anymore.

Even a simple camera needs a way of framing up the shots and a way of recording the shots.

David previously posted that on another thread.

And from the RED team:

Maybe because there are a few small items missing from "sensor in a box" cameras....

... like monitoring, recording, audio, timecode

 

[Philipe Ratton]

Think about it this way. If you don't need 95% of the features, buy something cheaper with the 5% you need. No disrespect intended, just basic logic.

 

[David Rasberry]

Thanks Michael, appreciate it. Especially with the links and all.

But I played with the footage the day it was released (and I loved it). But since I've never filmed with RED I'm lacking a reference point. There's absolutely nothing wrong with the Chicken footage (apart from being too short =), but I can't say from that alone if I'm getting everything or not (compared to the actual RAW stream).

But I understand the process better now. What I couldn't add up was "how on earth can you actually do 3k or 5k RAW". How do you manage the data...

I wouldn't opt to do it myself, but is it possible on Epic to rip the actual Bayer-stream sans REDcode? I only wonder out of curiosity.

OK if you want to see the closest representation of the raw sensor data set the color box to camera RGB and the gamma box to linear. Hit the "D" buttons to set all other parameters to their defaults. What you see in the view window is a demosaic of the raw sensor values as recorded.

 

[David Mullen ASC]

The underlying theme to a lot of these threads is "I want a cheaper camera, do whatever you have to do to make it cheaper. Except don't lower quality."

The thing is that some of the image control features are just software, it's not adding significantly to the cost of the camera, to make it cheaper would require not just omitting those features, but going with a less powerful processor that makes all those features possible. But less processing power may have an effect on areas that you do care about, not just the ones you don't.

The basic costs of the camera are not really tied up with small extras, it's the cost of the sensor, the processor, the cooling system, the recorder, the basic inputs and outputs, the body, the mount, etc. and the overall sales volume and how it affects manufacturing technique and per unit cost.

If you are talking about a 3K RAW movie camera, you are talking about a lot of processing power, whether or not the data gets compressed, and you are also talking about a lot of data to record (which compression can help with.)

The problem is applying the unit costs for a mass-produced consumer camera and thinking that a smaller volume, hand-assembled pro camera is going to follow the same economics of scale. Even if the basic Scarlet costs $7000, it will be a bargain at twice the price compared to anything else on the market, so at some point, wishing it was $3000 is like wishing a new Porsche or Ferrari was $10,000, it's not very realistic.

Not to say that Red perhaps could find a way to make a very stripped down, dumbed-down camera that costs less, and who knows, they may someday, but it would be a course deviation from their commitment to serve the professional market, not that some professionals might find a use for such a thing, as they do with Canon HDSLR's. But I think that trying to satisfy people with only a few thousand dollars in their pockets to spend is probably a bit lower on their list of priorities right now.

 

[AndreeMarkefors]

The underlying theme to a lot of these threads is "I want a cheaper camera, do whatever you have to do to make it cheaper. Except don't lower quality."

Although what you say seems perfectly reasonable to me (unreasonable would have been: keep all features+don't lower quality), this thread is not like that.

The posts that I can remember, and especially the ones I have written, have clearly made a case for "less product at a cheaper price". There have also been specific examples about what features to omit.

You say RAW is problematic to build cheap, but that was one of the things on the potential "maybe not needed" list.

And the whole idea of making a cheaper version is to turn it into a consumer product. Why else?

Market share would grow exponentially and it could be regarded as branding.

But I agree with your final point: RED doesn't have the interest, or more importantly the resources at the time, to build for the masses.

 

[Charles Angus]

[snip]
I bought the D3 and will upgrade the chip for a small sum, arriving at 230hp and 470Nm. With all the guarantees intact, as Volvo is performing the upgrade.

That's reality, right there.

That's unethical business practices, right there...

 

[David Rasberry]

Although what you say seems perfectly reasonable to me (unreasonable would have been: keep all features+don't lower quality), this thread is not like that.

The posts that I can remember, and especially the ones I have written, have clearly made a case for "less product at a cheaper price". There have also been specific examples about what features to omit.

You say RAW is problematic to build cheap, but that was one of the things on the potential "maybe not needed" list.

And the whole idea of making a cheaper version is to turn it into a consumer product. Why else?

Market share would grow exponentially and it could be regarded as branding.

But I agree with your final point: RED doesn't have the interest, or more importantly the resources at the time, to build for the masses.

RAW and resolution is what defines Red's camera technology. Otherwise it would just be another video camera. Plenty of choices there already.

 

[David Mullen ASC]

What Andree will probably show up someday, maybe soon, a simple 3K+ RAW video camera that costs less than $5000, whether or not it will be Red that builds it is a separate issue. Maybe Canon will build such a thing, who knows. It's not impossible by any means, it's just that the data requirements are quite large - large to record if you don't compress it, large to compress, large in general.

There's a reason why most consumer cameras are compressing HD down to ridiculous levels, at 8-bit 4:2:0. It's hard on the one hand and say "the consumer wants a simple, cheap RAW video camera" and then on the other hand end up with a "pro" RAW data rate output that is so high as to be beyond most consumers' ability to record it or edit it in post. It becomes a real question if, for example, 100+ Mb/sec is a "consumer" type of data rate.

So the problem for Canon, et al., is not so much the camera end of things (though running a sensor at 24 fps or higher and sending full RAW through the pipeline without overheating is a real challenge) but the recording, playback, and post side of things. You can't really call something a consumer camera and sell it to that market if it needs professional knowledge and gear to handle it at the other end. Consumer divisions tend to not think that way, they think of entire workflows because if they don't, half the consumers aren't going to figure it out. And handling large amounts of RAW data, dealing with transcoding it to color video in post, sort of falls beyond the borders of the consumer market.

So it may be less of an engineering problem as it is a marketing problem. I would think dealing with 3K+ RAW moving images, recording them, processing them, color-correcting them, falls into the prosumer and higher market, in which case such a market generally is willing to spend more than $3000 for a camera.

 

[Terry VerHaar]

...
You say RAW is problematic to build cheap, but that was one of the things on the potential "maybe not needed" list.
...

If you give up (REDCODE) RAW, you are traveling down the road towards many other cameras. 3K is also harder to process (thus requiring a more sophisticated, and expensive camera). So, if you eliminate that, you go even farther down the road toward many other cameras. Right now you can probably buy a used Sony EX-1 for less than what a Scarlet will cost. Great camera, without RAW and without 3K. Yeah, it's not exactly modular. And, it doesn't have HDRx™™ - but that's ANOTHER one of those "expensive" features anyway, so you have to give that up, too.

So where does that leave us with a "stripped down" Scarlet? Maybe a used EX-1 or even a Canon 7D? Depends on what you think is "essential," doesn't it? :undecided:

 

[Leaf Baimbridge]

Think about it this way. If you don't need 95% of the features, buy something cheaper with the 5% you need. No disrespect intended, just basic logic.

If it existed I would likely do so.

 

[Joseph Hutson]

You say RAW is problematic to build cheap, but that was one of the things on the potential "maybe not needed" list.

I'd rather have an HD camera with REDCODE RAW than a 3K cam with the 7D's H.264...

 

[Ash Worth]

I'd rather have an HD camera with REDCODE RAW than a 3K cam with the 7D's H.264...

I'd rather have a HD camera that I can actually have.

 

[AndreeMarkefors]

That's unethical business practices, right there...

(Starting off with a quick reply to this. I won't discuss it further, though.)

No, it isn't.

Many premium car makers have tuners that develop chip software for turbo engines. Usually a consensus is reached about how far an engine is allowed to be pushed with warranties intact.

In Europe, for Volvo (just like many german car makers and so on), there are tuners that sell chips that comply with Volvos guidelines. In some countries Volvo cooperates with them as a liaison. But normally the customer organizes everything.

The throttling of the engine is the most economical way for Volvo to offer two price points. Just like Intel chips many times are the same, but sold at different price points, where the speed is controlled in BIOS, and not by the chip itself.

Anyway. Won't discuss further. I just thought the above was an unnecessary remark.

 

[Ken Waller]

Deal of the Day had the Flip HD for $39 Ash, you can actually have one if you want it now.

How long can these cry babies bitch about camera prices. Yawn! The Fixed will still cost you less than shooting on 16mm and getting a HD video transfer. Plus you still have the camera afterwards.

 

[AndreeMarkefors]

I'd rather have an HD camera with REDCODE RAW than a 3K cam with the 7D's H.264...

I'd probably want to quote a couple of other things as well, but it's all around the same theme...

When I say I could live without RAW, what I mean is I'd settle for Redcode at an appropriate compress rate.

To me, RAW is two things:

1. Access to the actual sensor data at time of capture
2. Possibility to revert an image to original state, through active meta data.

Someone mentioned that RED has a solution for the future in the lossless REDCode version that complies with both of the above. But for the time, I don't consider the lossy REDCode versions completely RAW.

Don't even think that I'm interested in starting a fight about this—I'm not. But things are what they are.

3k+ video/film RAW is almost not possible at the moment. I believe that REDCode RAW is a very good and to the point solution.

For a cheaper camera they could possibly remove the highest end compression rates with low compression?

HDDSLRs are not a viable option since the lack even "normal audio", auto focus, a screen large enough to verify focus and suffer from moiré and aliasing artifacts. And h.264 doesn't help, but it's not the main problem. High enough data rates+editing in 10-bit, would be OK for entry level film/video.

 

[AndreeMarkefors]

Deal of the Day had the Flip HD for $39 Ash, you can actually have one if you want it now.

How long can these cry babies bitch about camera prices. Yawn! The Fixed will still cost you less than shooting on 16mm and getting a HD video transfer. Plus you still have the camera afterwards.

I'll never understand how someone can reduce themselves to something so small and laughable, over so little.

Ken, if your post isn't deleted, and soon forgotten... how can you hope that anyone will take anything you say, or opinionate seriously?

 

[Joseph Hutson]

How long can these cry babies bitch about camera prices. Yawn! The Fixed will still cost you less than shooting on 16mm and getting a HD video transfer. Plus you still have the camera afterwards.

Come on, man. Grace?

 

[Christian Jadot]

Deal of the Day had the Flip HD for $39 Ash, you can actually have one if you want it now.

How long can these cry babies bitch about camera prices. Yawn! The Fixed will still cost you less than shooting on 16mm and getting a HD video transfer. Plus you still have the camera afterwards.

This really isn't really a fair arguement. The scarlets price has slowly crept upward since its announcement. It started 3k for 3k, from there the price has almost doubled, pushing the price out of some peoples budget. This is why people are bitching. If it increase much more past 6k, I will not have the budget. I would take a "stripped" down version for $1-2k less. (I would still want Raw though )

 

[AndreeMarkefors]

This is why people are bitching.

I'm pulling this quote out of context. The following is not necessarily targeted at you Christian.

We are not bitching.

This discussion, for me, is about a couple of things:

• defining what is technically possible to achieve for a certain price point.
• defining what is necessary, or what makes a camera useful for certain markets (biased towards <$4000), and perhaps identify key things that would make a camera stand out in i certain customer segment.
• arriving at a consensus, so that we can define for RED, or any other company reading, what would make an attractive offer as an entry level film/video tool.

It has been stated many times that RED has taken input from their forums, and I believe that to be true.

That doesn't mean that I think RED will build the camera some of us are hoping for. Looking at their business plan, they don't have the time or capacity right now. And when they might finally have time, they probably still don't want to build it. Who knows?

Someone stated that RED is high resolution and RAW. Both good things. And if that is the end all be all with RED—clearly they won't deviate.