Arri Alexa and Mysterium-X...

[Tim Morten]

The issue of altering sensor white balance via gain is that any increase in gain forces premature clipping on that channel, which is un-recoverable afterwards.

This is really the heart of the matter why "baked" white balance is bad, and why no one here consider's Red's WB to be "baked". Dwight, if that's not clear to you, I'm not sure how it can be more clearly communicated.

- Tim

 

[Scott Crawley]

Just to further clarify...

I stated that releasing enhancements via firmware was a financial risk. If you are not charging for those enhancements it becomes a triple negative and a no-brainer.
-1 Bricks come back for service.
-2 You can't recover R&D costs
-3 You won't profit from that R&D as you would if you put it into a new model.

 

[Steven Caesare]

I never said it was a "gimmick" I said that (unless mathematically lossless) by strict definition compressed is not raw. I also very clearly stated that I believe it was not only an acceptable but a USEFUL marketing TERM. Twisting that into I said it was a "gimmick" is exactly the kind of misrepresentation that disturbs me on this thread/forum.

I apologize if you felt I was quoting you when I said that I felt that Red's usage of the term "compressed RAW" wasn't a marketing gimmick. I was not quoting your exact words, hence no quotation marks.

However, you did state that:

"once you decide to discard some of the information you are almost by definition NOT providing the RAW sensor data"

and

"Compressed RAW" is an oxymoron."

as well as

"everyone is messing with the TRUE RAW and has debased the term"

Note that there IS no one definition of RAW. Wikipedia says:

Providing a detailed and concise description of the content of raw files is highly problematic. There is no single raw format; formats can be similar or radically different. Different manufacturers use their own proprietary and typically undocumented formats, which are collectively known as raw format

Furthermore, reagrding compression it admits:

Most raw formats do not use compression or implement light lossless data compression to reduce the size of the files without affecting image quality. But some others use lossy data compression where quantization and filtering is performed on the image data.[14][15] Many recent cameras let photographers choose between no compression, lossless compression or lossy compression for their raw images.

So I'd suggest that your statement that "Other than MATHEMATICALLY LOSSLESS compression..." everything else doesn't qualify to be called RAW is not what the industry holds as as the same opinion.

I provided examples why spatially compressed Bayer sensor data (I'll avoid callint it "RAW" for your benfit), is still significanlty benefical for all the same reasons people have been choosing RAW workflows for their DSLR, and more rectly, video workflows.

This is why I object to your catagorizing it as a "marketing term", there are solid technical benfits to a compressed RAW format.

-sc

 

[Michael Hastings]

This is really the heart of the matter why "baked" white balance is bad, and why no one here consider's Red's WB to be "baked". Dwight, if that's not clear to you, I'm not sure how it can be more clearly communicated.

- Tim

If I understand Graeme correctly setting a non-native white balance has consequences (baked or unbaked) and "unbaked" simply postpones that choice.

I think the counterpoint that wasn't that hard to understand was that a lot of people are fairly confident and comfortable with making that choice when they are standing there at the location.

I like the option to be a little careless and I don't understand ARRI's reasoning (it seems that it has yet to be explained) but if we didn't have that option or if there is some other benefit, I wouldn't have a big problem committing to a white balance setting. (Not to mention, knowing that we still presumable have quite a bit of room for color grading adjustment with these almostRAW formats - much more than was available back in the Betacam days or even current video in general.)

The other question I have - and I think it is an important one - is that if "baking" WB in has an effect when you set it to a non-sensor native WB isn't the inverse true? - that if you set WB to the sensor Native WB, i.e. if the sensor is native 5000 and you set the "baked-in" WB at 5000, then there is no effect and you have maintained the "RAWNESS" of the RAW?

 

[Graeme Nattress]

Actually, it's not just the postponing of the choice of white balance. Consider that the order of operations has the white balance followed by a colour correction matrix. These two operations concatenate nicely. If you separate them by an integer representation, like in recording (even uncompressed) you will have the matrix magnifying rounding errors in the integer conversion. This is undesirable.

Graeme

 

[Jeff Coatney]

Just to further clarify...

I stated that releasing enhancements via firmware was a financial risk. If you are not charging for those enhancements it becomes a triple negative and a no-brainer.
-1 Bricks come back for service.
-2 You can't recover R&D costs
-3 You won't profit from that R&D as you would if you put it into a new model.

This really only applies to cameras that are high volume consumer or prosumer cameras with limited internal capabilities and low price-points. Professional cameras are durable goods and the R&D costs you mention are reflected in the high price-point.

Digital or electronic cameras have the added hurdle of addressing a sector of the market that was previously served by two products: the film that ran through them and the light-tight box with the clockworks that moved the film through for exposure.

Digital or electronic cameras have absorbed "the Film" permanently, eliminating what is essentially an analog version of the firmware update: improvements in emulsion and stocks. With film cameras, every time you slapped in a new roll, you basically got an updated camera with better performance.

The benefit of downstream firmware improvements is the system can evolve or improve even after the camera has left the factory and features and capabilities can be added or accessed without the expense of building an entirely new hardware platform. This adds to the lifespan of the product and gives the end-user greater value for their purchase price.

The Red DSMC system adds to these benefits by also managing the negative effects of Moore's Law. High cost, low volume equipment like digital cameras face a limited lifespan due to improvements in processor power and ever lowering costs.

By modularizing the components most likely to suffer this increasingly limited life span, Red has enabled the end user to upgrade only those parts of the camera that can take advantage of the positive effects of Moore's Law and limits the need to buy or upgrade components that are unaffected by it, like I/O, battery connections, the housing and peripheral attachments.

 

[Jeff Coatney]

Actually, it's not just the postponing of the choice of white balance. Consider that the order of operations has the white balance followed by a colour correction matrix. These two operations concatenate nicely. If you separate them by an integer representation, like in recording (even uncompressed) you will have the matrix magnifying rounding errors in the integer conversion. This is undesirable.

Graeme

You're saying that baking in values at the recording stage increases the probability for error propagation?

 

[Graeme Nattress]

You're saying that baking in values at the recording stage increases the probability for error propagation?

Yes, that's what I'm saying. Now that error may just show up as quantization in deepest shadows. I'd also be worried about premature highlight clipping in any channel with increased gain.

Graeme

 

[Jeff Coatney]

Yes, that's what I'm saying. Now that error may just show up as quantization in deepest shadows. I'd also be worried about premature highlight clipping in any channel with increased gain.

Graeme

So you could have unrecoverable noise or unintended clipping that could be the result of error propagation and not user-induced exposure choices.

 

[Scott Crawley]

This really only applies to cameras that are high volume consumer or prosumer cameras with limited internal capabilities and low price-points. Professional cameras are durable goods and the R&D costs you mention are reflected in the high price-point.

Digital or electronic cameras have the added hurdle of addressing a sector of the market that was previously served by two products: the film that ran through them and the light-tight box with the clockworks that moved the film through for exposure.

Digital or electronic cameras have absorbed "the Film" permanently, eliminating what is essentially an analog version of the firmware update: improvements in emulsion and stocks. With film cameras, every time you slapped in a new roll, you basically got an updated camera with better performance.

These are good points, but every time you bought a new roll of film you paid for it, and it didn't matter if it were a new improved stock or not. To get the same improvements with digital you will either have to invest in a new brain, or maybe you'll have the option to buy a firmware upgrade. That is one of the more hidden costs of working digitally.

How many free firmware upgrades would you reasonably expect for your "high price point" of $28,000 for a new Epic? The Alexa of course costs more than that, but what does a Panaflex cost? Is $28,000 or $60,000 a lot in comparison?

If the new improvements can be applied via software only it would seem that the firmware option would be prefferable, as it costs us all less. I don't think the typical U.S. consumer will see the same value and satisfation in a software update as they do when they buy the next, shiny new thing, so from a perception standpoint I expect companies to continue to lean toward hardware replacement as upposed to updating units in the field.

We're trying to be more green, so my preference is for updates, but I think that is minority thinking.

 

[Andrew M.]

But .5k is used for look-around and is not available for RAW recording or for scaling.

Graeme

5K is not available. on EPIC just for now I hope
I remember RED1 4.5 K format was not available for some time at first and now we have full 4.5K and still we have look-around at top and bottom.

As to look-arround area on each side of the frame it is not so important as top and bottom area, you know hanging mikes etc.

Arri is 2.8K wiht 3.5 K look-arround. If they will have spare DSP cpacity, free firmware upgrade and modular way, maybe it will be full 3.5K one day?

Did I say free.....)(zzz

 

[Michael Hastings]

Actually, it's not just the postponing of the choice of white balance. Consider that the order of operations has the white balance followed by a colour correction matrix. These two operations concatenate nicely. If you separate them by an integer representation, like in recording (even uncompressed) you will have the matrix magnifying rounding errors in the integer conversion. This is undesirable.

Graeme

But that doesn't seem to answer my question/proposition ".... that if "baking" WB in has an effect when you set it to a non-sensor native WB isn't the inverse true? - that if you set WB to the sensor Native WB, i.e. if the sensor is native 5000 and you set the "baked-in" WB at 5000, then there is no effect and you have maintained the "RAWNESS" of the RAW?

 

[Jeff Coatney]

But that doesn't seem to answer my question/proposition ".... that if "baking" WB in has an effect when you set it to a non-sensor native WB isn't the inverse true? - that if you set WB to the sensor Native WB, i.e. if the sensor is native 5000 and you set the "baked-in" WB at 5000, then there is no effect and you have maintained the "RAWNESS" of the RAW?

It would imply that its the introduction of the integer itself that creates havoc with error correction, so regardless of the color temp setting, you never escape the potentially negative effects of "baking in". In this scenario, you may have the potential of protecting the Raw, but you'll never know for sure.

 

[Graeme Nattress]

5K is not available. on PIC just for now I hope
I remember RED1 4.5 K format was not available for some time at first and now we have full 4.5K and still we have look-around at top and bottom.

As to look-arround area on each side of the frame it is not so important as top and bottom area, you know hunging mikes etc.

Arri is 2.8K wiht 3.5 K look-arround. If they will have spare DSP cpacity, free firmware upgrade and modular way, maybe it will be full 3.5K one day?

Did I say free.....)(zzz

It's probably HD-SDI limits because that's how the raw is output. DSP would be a limit for 3.5k to 2k or 1080p scaling rather than the fixed 2/3 scaling they use currently.

Graeme

 

[Tim H.]

Ok, several pages later and I'm still sticking needles in my eyes as I read this thread. This place is freaking amazing. I've never seen so many people pull things out of their ass and try to pass it like fact. I've never seen so many wanna be technical gods who are debating how they know for a fact what one company is doing over another and how one sensor design over another is going to kill the other.

People - as sad as it sounds if you want to win you have to win the hearts and minds and that usually comes from marketing, not necessarily the better product. Just look around you. Part of Reds success has come from it's gorilla out of the box strategy. Jim is a master at viral marketing and he perfected it with Oakley and now successfully applied it to Red.

Doesn't the endless debate about mine is bigger than yours get old after a while? Page after page of going in circles trying to convince everyone that your pov is right? Who fn cares. Why don't you guys go out and shoot something and make a name for yourself. I want to see some of the names on this board win some awards and industry recognition. Show me the Academy Award, ASC Award, Spirit Award, Best Little League Playoff Video Award, something! anything! If some of you spent half the time shooting instead of arguing points on here you might actually get somewhere.

I actually enjoy a good argument (I think we all do as you can see from these forums) but what makes me crazy is that some of you argue about points of view you don't even have. Meaning you don't even have an ounce of real world experience yet you think you can argue the subject as fact. Think about someone in the operating room with no medical school experience trying to argue to the surgeon how he should better perform the surgery. That's Insane.

ARRI - Some of you think that Arri has lost it's way or changed? Of course they have changed. They have to. The film industry has changed. Arri is a huge ship and it will take some time to turn it but at least they have started. They are still supporting their HUGE film market which I hate to break it to those who can't wait for it to go away because it scares you due to the fact that you have no clue how to expose a negative or operate a film camera but it's not going anywhere for a long time. TV might transition more and more away and that makes perfect sense. However, films shot on film and scanned to 4K look damn good, especially when they are then projected digitally.

So why should Arri worry so much about 4K right now? They have a huge film market and make damn good film gear. Some of the top names in our business are perfectly happy shooting film. So Arri made a camera to serve the TV market. 2K is plenty for the current HD market and the Alexa fits it perfectly. They are also getting a pipeline/workflow established so future cameras can expand on it. I can see the argument about 2K not enough for film, especially if it's going to be projected in 4K but we are not there just yet and if I were picking a camera for a film that I know might go out on 4K delivery then I wouldn't use the Alexa. I'd shoot with an Arri 35mm camera and scan at 4K. So pick the right tool for the job. Anyway, some of you bitch that Arri is too stuck in the old ways and then you bitch that they are changing up their marketing strategy. At least they are trying to adapt.

Red doesn't need to worry. They appear to have a very good tool coming, and if it's as good as they say it is then the industry will gobble it up. Period.

 

[Michael Hastings]

I think it is less a question of whether or not they can, but if they will. I certainly don't know the answer to either question. Before Red came along with their broad public beta program called Red One, such a thing would be an extreme rarity, reserved for fixes rather than enhancements.

I don't think this is a function of the RED program, it is simply a function of the computer/electronics industry in general. In the 90s firmware upgrades were rare, whether it was the size and availability of flash memory, the lack of easy distribution via the internet or whatever...

Free firmware updates were available that provided fixes and higher data rates and storage capacities on my circa 1999 Ikegami EditcamII - one of the first true "computercams"...

... in the 2000s firmware upgrades (including feature upgrades) have become the norm across most sectors of electronics whether it is computers, laptops, ipods/iphones, dSLRs, HDDVD/Bluray players, XBOX, PS3, GPSs, etc. etc. etc. - hardly a RED innovation.

Major improvements on mass market things like consumer camcorders/point and shoots/dSLRs come in new bodies for three reasons 1) they are generally much more "debugged" than say a redone and 2) they have generally been able to incorporate new features like improved ASICs, graphic cards, networking etc. in the new hardware as well. They do it incrementally because those products are much more price sensitive to a $50 or $100 price difference than a $5K, $20K, $70K camera and they have multiple price point/market segments 3) while the modularity of the "brain" system with Epic scarlet is nice I would suggest that its main import is in being able to configure the package in substantially different ways for different purposes. Making a sensor module that is interchangeable might make a little sense on the $5K plus high end dSLRs (although it would create greater weight/size and sealing issues - critical features on those cameras) it would not make sense economically for the lower end 7D/t2i type since the sensor module would probably cost nearly the same as the whole camera due to the much higher mass production of the camera (not to mention the headaches of supporting millions of consumers dismantling the cameras to replace a sensor).

The fixed scarlet is a tacit admission that there are substantial efficiencies to a non-modular package.

 

[Michael Hastings]

It would imply that its the introduction of the integer itself that creates havoc with error correction, so regardless of the color temp setting, you never escape the potentially negative effects of "baking in". In this scenario, you may have the potential of protecting the Raw, but you'll never know for sure.

I'm not sure if that is true, and I think I'd rather have an in-depth actual explanation from Graeme rather than speculation on what it implies.

Since I've asked twice (or is this three times?), if I were being snarky, given the history of this thread I'd ask if he was avoiding the question. :biggrin:

Big grin, Big grin, big f---ing grin - I'm just teasing..

 

[Jeff Coatney]

I'm not sure if that is true, and I think I'd rather have an in-depth actual explanation from Graeme rather than speculation on what it implies.

Me too. That's why I used "imply"...

 

[Scott Crawley]

... in the 2000s firmware upgrades (including feature upgrades) have become the norm across most sectors of electronics whether it is computers, laptops, ipods/iphones, dSLRs, HDDVD/Bluray players, XBOX, PS3, GPSs, etc. etc. etc. - hardly a RED innovation.

I didn't say it was a Red innovation, but their ongoing upgrade program is a permutation of it; the result of the fact that it was a beta program. It is definitely anomalous. I concede that firmware upgrades do sometimes include feature enhancements, the primary function of most is still repair, be it a bug fix or security update. To use one of your examples; Apple has made software updates an integral
part of what they do. Their process is streamlined and largely idiot proof. They do offer features in some of their updates, but on an iPod you might see three updates per model total. After that you need to kick in and buy another. Digital cinema and iPod are apples and oranges, but the point remains; free upgrades are of limited scope and number. To do otherwise cuts into the bottom line. It can't go on forever.

But to get back on track I'll agree that Arri could indeed, possibly upgrade ArriRaw in a future update. I won't hold my breath if that's OK with you.

None of this has much to do with my original assertion that Mr. Bravin's main purpose of posting here was not market research. That usually happens at the beginning of the development cycle, not the end.

What is his title anyway?

 

[Jeff Coatney]

Ok, several pages later and I'm still sticking needles in my eyes as I read this thread. This place is freaking amazing. I've never seen so many people pull things out of their ass and try to pass it like fact. I've never seen so many wanna be technical gods who are debating how they know for a fact what one company is doing over another and how one sensor design over another is going to kill the other.

People - as sad as it sounds if you want to win you have to win the hearts and minds and that usually comes from marketing, not necessarily the better product. Just look around you. Part of Reds success has come from it's gorilla out of the box strategy. Jim is a master at viral marketing and he perfected it with Oakley and now successfully applied it to Red.

Doesn't the endless debate about mine is bigger than yours get old after a while? Page after page of going in circles trying to convince everyone that your pov is right? Who fn cares. Why don't you guys go out and shoot something and make a name for yourself. I want to see some of the names on this board win some awards and industry recognition. Show me the Academy Award, ASC Award, Spirit Award, Best Little League Playoff Video Award, something! anything! If some of you spent half the time shooting instead of arguing points on here you might actually get somewhere.

I actually enjoy a good argument (I think we all do as you can see from these forums) but what makes me crazy is that some of you argue about points of view you don't even have. Meaning you don't even have an ounce of real world experience yet you think you can argue the subject as fact. Think about someone in the operating room with no medical school experience trying to argue to the surgeon how he should better perform the surgery. That's Insane.

ARRI - Some of you think that Arri has lost it's way or changed? Of course they have changed. They have to. The film industry has changed. Arri is a huge ship and it will take some time to turn it but at least they have started. They are still supporting their HUGE film market which I hate to break it to those who can't wait for it to go away because it scares you due to the fact that you have no clue how to expose a negative or operate a film camera but it's not going anywhere for a long time. TV might transition more and more away and that makes perfect sense. However, films shot on film and scanned to 4K look damn good, especially when they are then projected digitally.

So why should Arri worry so much about 4K right now? They have a huge film market and make damn good film gear. Some of the top names in our business are perfectly happy shooting film. So Arri made a camera to serve the TV market. 2K is plenty for the current HD market and the Alexa fits it perfectly. They are also getting a pipeline/workflow established so future cameras can expand on it. I can see the argument about 2K not enough for film, especially if it's going to be projected in 4K but we are not there just yet and if I were picking a camera for a film that I know might go out on 4K delivery then I wouldn't use the Alexa. I'd shoot with an Arri 35mm camera and scan at 4K. So pick the right tool for the job. Anyway, some of you bitch that Arri is too stuck in the old ways and then you bitch that they are changing up their marketing strategy. At least they are trying to adapt.

Red doesn't need to worry. They appear to have a very good tool coming, and if it's as good as they say it is then the industry will gobble it up. Period.

Tim, you are on fire. Stop, Drop and Roll!