Epic and Scarlet release in 2010

Status
Not open for further replies.
Subhadip said:
As for Scarlet, I am not sure where the "7000 per month" rumour came from. 2000 per 6 months sounds very low. Scarlet 2/3" is not competing with something like a $17k Sony PMW-350. We are talking about Scarlet 2/3" Fixed, which is a ready-to-shoot camera at $4750. It is clearly a competitor to the Sony EX1 or Panasonic AF100. This is a much larger volume segment. Now, I don't expect it to have massive sales like Panasonic HVX200 or Sony EX1/EX3, but there's a bigger market than 2000 per 6 months, certainly. The Red One, at $30k or more, has sold 8000 units within 3+ years!
Click to expand...

These market segments are not as large as we might think they are. Yes, a successful affordable camcorder can sell 20,000 units within a year, but this is digital cinema. Very few people need it and very few people appreciate the advantages and are prepared to deal with the digital cinema workflow. 2000 units a year is very rare (red one?) and at the present state of the market and taking the saturation by Red One into account, 1000 units is a major success. Thousands talk about the Arri Alexa, but only a few dozens will buy one. Most will not even use it once, and some will only rent it. Rentals easily saturate the market in the digital cinema world.

The entire market for a 10,000 euro digital cinema camcorder looks smaller than 10,000 units a year and 4,000 units is probably closer to reality. The market is a lot smaller than that of a 10,000 euro camcorder.

How many units did Si-2K actually sell? After a reasonable marketing effort, Ikonoskop got 43 preorders or something like that. I have a list of about 1300 people interested in Drama and what would this translate to? 200 or 250 units perhaps. The Arri Alexa sales expect a few hundred units and IMHO, in 2010 they are being optimistic for a camera of that price. HPX3700s and F23s do not sell much either. F35s and high digital cinema cameras? They are not even moving.
 
Something I'm particularly interested in is how much the marketing for Scarlet includes the 1080p RGB out. Of course, we still have no idea what format (ProRes 4444 anyone?) or even what problems may come along with the in-camera debayer and downscale realtime. But the fact that it could be used - perhaps very well, perhaps not so much - as a "camcorder" rather than a digital cinema camera may allow it to bridge markets. RAW aside, I still see unique features on the Scarlet platform.
 
Raw work flow is not that cumbersome with Redcode anymore. Vegas Pro9 and Premiere CS5 accept native .r3d files realtime without conversion now. Next gen FCP will too. A year ago, two years ago, it was more of a problem. Is there a learning curve to take best advantage of it? Sure. But having single raw footage format that can be archived, graded, formatted and repurposed across multiple media types from 2k DCI filmouts to web streaming apps is I think Red's single most valuable attribute. One is not locked into anything about the original image other than the actual range of raw sensor data captured. That is only affected, as with film, by subject illumination, aperture, optical filtering, frame rate, and shutter speed. Everything else is metadata that can be manipulated at will, in camera or in post, to get the output look you want without affecting the original recorded file at all.

It is this well engineered practical RAW acquisition model that I am most interested in. Video is so passe. :Drogar-BigGrin(DBG)
 
Last edited:
Michael Olsen said:
Something I'm particularly interested in is how much the marketing for Scarlet includes the 1080p RGB out. Of course, we still have no idea what format (ProRes 4444 anyone?) or even what problems may come along with the in-camera debayer and downscale realtime. But the fact that it could be used - perhaps very well, perhaps not so much - as a "camcorder" rather than a digital cinema camera may allow it to bridge markets. RAW aside, I still see unique features on the Scarlet platform.
Click to expand...

Exactly. That is where most digital cinema cameras will fail. A real time RGB output is not the solution, in a raw camera the quality will be poor.

Most raw cameras do not even have adequate processing in camera (good gamma curves, proper color controls, image finetuning etc). Even the RAW workflows do not have the functionality of a good camcorder in image finetuning because the companies do not invest time in developing such features. RAW is used as an excuse for poor processing. You have curves and some controls, make your image. Nothing could be further from the truth. More options is not a substitute for the art and experience implemented by the camera engineer. That will never produce a filmic look or treat highlights well without a major investment in processing each shot. You can't expect the typical camcorder user to have the knowledge of someone working in the color correction business. Many uses are simply not worth the investment in time or the hassle.

The affordable digital cinema market is quite unique. It has to compete with regular camcorders. The digital cinema camcorder that manages to produce a good image out of the camera without post processing and be productive in everyday use, while maintaining the power and post correction a RAW workflow offers, will be the one to survive.
 
Do you guys really think a Scarlet will ship before this time next year? July 2011? I really don't think that's going to happen which is really a big delay. I thought that it should have been out now, atleast that was their projections some 2 years ago at NAB....

If Red doesn't think that the other big boys don't have something up their sleeves then they are wrong. Red is trying to steal the big boys lunch money. It's fine when it's up against the higher end market but when you are into the realm of the sub 10k, and certainly in the 4500 range, well that just is the ass market that will certainly make all of them pull no punches. Canon has the sensor but they just need to figure out how to get the best results for video.
 
Otis Grapsas said:
These market segments are not as large as we might think they are. Yes, a successful affordable camcorder can sell 20,000 units within a year, but this is digital cinema. Very few people need it and very few people appreciate the advantages and are prepared to deal with the digital cinema workflow. 2000 units a year is very rare (red one?) and at the present state of the market and taking the saturation by Red One into account, 1000 units is a major success. Thousands talk about the Arri Alexa, but only a few dozens will buy one. Most will not even use it once, and some will only rent it. Rentals easily saturate the market in the digital cinema world.

The entire market for a 10,000 euro digital cinema camcorder looks smaller than 10,000 units a year and 4,000 units is probably closer to reality. The market is a lot smaller than that of a 10,000 euro camcorder.

How many units did Si-2K actually sell? After a reasonable marketing effort, Ikonoskop got 43 preorders or something like that. I have a list of about 1300 people interested in Drama and what would this translate to? 200 or 250 units perhaps. The Arri Alexa sales expect a few hundred units and IMHO, in 2010 they are being optimistic for a camera of that price. HPX3700s and F23s do not sell much either. F35s and high digital cinema cameras? They are not even moving.
Click to expand...

Thanks for your insights, Otis. And I would definitely agree with you were the Scarlet 2/3" Fixed a 10,000 Euro camera. However, it isn't. It is half that price, a $4750 mass-produced digital cinema camcorder. A direct competitor to the EX1 market. Now, unlike the EX1, this is not a video camera, and sure, its demand will be far lesser, at least by logic. The Red One, at $18k body only has sold upwards of 2000 units per year. Surely, surely, a $4750 ready-to-shoot camera will sell more than that! I am willing to bet a couple of thousand users have posted on REDUser's Scarlet forum with a real intent to buy. Scarletuser.com has over 4,000 members. I won't say there's a very large userbase at all, no. I don't know where the 7,000 per month rumour came from, but that sounds far too high. I just feel you may be underestimating the potential market for Scarlet a tad. It is the Scarlet S35 that will more likely serve the 10,000 Euro market, not the Scarlet 2/3" Fixed. Scarlet S35 certainly would have a more limited userbase.

The likes of SI-2K and Alexa are in a far higher price range. I have no doubt neither of these cameras will sell more than a few hundred. I wish you best of luck with Drama, and 200 to 250 units is a very good result indeed. These are all niche products. Scarlet 2/3" is a commercial, mass-produced, affordable (under 5000 Euros) digital cinema camera.

Also, I feel the REDCODE workflow is easier than the DSLR H.264, for example. It is a far, far, far superior acquisition format, and it works just fine real-time preview on Premiere Pro CS5 or Vegas Pro 9 with all the latest color science / FLUT etc. Surely, FCP will catch up soon.

Justyn said:
Do you guys really think a Scarlet will ship before this time next year? July 2011? I really don't think that's going to happen which is really a big delay. I thought that it should have been out now, atleast that was their projections some 2 years ago at NAB....

If Red doesn't think that the other big boys don't have something up their sleeves then they are wrong. Red is trying to steal the big boys lunch money. It's fine when it's up against the higher end market but when you are into the realm of the sub 10k, and certainly in the 4500 range, well that just is the ass market that will certainly make all of them pull no punches. Canon has the sensor but they just need to figure out how to get the best results for video.
Click to expand...

As has been said time and time again, everyone fully expects the big companies to catch up. Red are not sleeping. They have been working tirelessly over the last 2 years. It is a very complex project - bringing a RAW digital cinema camera to a $5000 market. There have been challenges - several unforeseen. No one really knows when it will release. I just hope it is 2010.
 
Last edited:
Still photographers have been working with raw images for a decade or more now. I think the digital cinema cameras like Red are going to appeal to cinematographers who are more serious about their image quality and about the archival possibilities of stock footage for multiple uses than they will appeal to more casual "filmmakers". There are shooters who are recording the HDSDI monitor out of Red 1 now with a proper LUT to deliver instant HD video footage to clients in common tape or file formats. The RED 1 was not really designed for this, but Epic and Scarlet will be much more capable and flexible in this respect. Still it will require knowledge and skill significantly beyond the automatic everything convenience of a modern video camera.

Jarred said Red was gearing up to be able to produce up to 7000 cameras a month, but like any small manufacturer these days, they will build to order based on demand for specific models. I don't think anyone anticipates these cameras being a huge mass market consumer/prosumer product. Red does not need that to be a healthy and profitable company. Any camera selling for more than $3k is by definition a niche market product, the mass consumer market is 95% under $2k. Even with the runaway success of the HDSLR's, interchangeable lens cameras represent less than 15% of Canon's total camera market.
 
Last edited:
Justyn said:
Do you guys really think a Scarlet will ship before this time next year? July 2011?
Click to expand...


In response to the people who posted on this forum in 2008 that they were holding off on the purchase of an HMC150 for a Scarlet, I present the following reality check...

Prior to working prototypes you're going to see footage leaked. To my knowledge no footage has been leaked from a 2/3" Scarlet with fixed lens.

Months (one to five) after footage is leaked you'll see working prototypes in the hands of big names -- "working prototypes" as opposed to the foam mock-ups their "marketing" people are currently showing people.

Months (one to six) after the bugs have been addressed on those first working prototypes the manufacturing process will be commenced. In this you need to understand that in the best possible scenario -- it'll take six months before a single factory can produce a single component. That's the "best case scenario." The reality is that it could take nine to sixteen months before stuff actually gets produced.

Next, they'll probably run a somewhat limited first run production series -- a thousand, two thousand? Those cameras will go to select customers -- not you -- the general public.

So -- you -- probably won't be filming anything in "RAW" until 2012 or 2013.

Is this criticism of Red? Absolutely not. It's simply the real world.

The truth is that all the buzz about Red has pushed the conservative Japanese camera makers to come up with innovations -- and for that I'm grateful to Red. This AF-100 from Panasonic is a result of Red. Sony's (very, very likely overpriced) mock-up at NAB 2010 is a result of Red. Canon's answer to the AF-100 (yes, it's coming) is a result of Red.

So -- without even releasing Scarlet -- Red has already changed the world for the indie guys.

Bottom line -- keep being excited about Red -- and the Japanese companies who spend much time trying to figure out what we're thinking and up to will keep innovating -- and those innovations mean better hardware at lower costs.

And, in 2012 or 2013, I look forward to buying my first camera from an American company. The good news is that by then solid state hard drives with 50 terabyte capacities will be on sale at Best Buy.
 
I think those interested in digital cinema require interchangable lenses and the option to use large apertures so they will want the more expensive Scarlet. Based on the fixed Scarlet specs I have read (3.3um and f2.6 which I'm not sure about), I don't see a chance against the camcorders competition. The pixel is small (between 1/2" EX1 and 1/3" XF300) and the fixed zoom lens is a very small aperture. I would be surprised it it could compete with EX1 in low light or dynamic range, actually I suspect 1/3" performance from the fixed lens scarlet. What is 2/3" good for if your lens wastes the light and the shallow DOF and you have no option to change it? 2/3" 5um pixels and f1.4 is professional camcorder performance in low light, dynamic range and shallow DOF. I don't see that happening with 3.3um pixels (2.3x less light per pixel) plus f2.6 (3.5x less light per pixel). That's a 3stop (8x) light gathering disadvantage vs a 2/3" camcorder. If the aperture is f1.6 it will compete with camcorders well, but don't expect a 10x f1.6 2/3" zoom on a 5,000 euro camera, even if it's a fly by wire varifocal. I also believe the Scarlet fixed lens is not feasible at that price until sales become 50,000 a year or so due to the 2/3" lens and investment in integrating the extra features that the common consumer fixed zoom camcorder has.

The consumer will not see the adantages of the intraframe format if cameras of the same cost have better low light, a viewfinder, larger monitors, a mic, cheap peripherals etc. I'm not sure that part of the market wants to deal with a raw workflow either.
 
A correction. Based on a fixed scarlet FAQ (it's from 2008) the lens will be f2.4/T2.6 and not f2.6. So the loss vs f1.4 is about 2.94x plus 2.30x for pixel area = 6.8x less light vs a 2/3" 1080p sensor with f1.4 lens.
 
Otis, what do you think about the coverage area for 5k on Epic. No one has released any real facts yet. Some guy has a chart which has approximated calculations, which puts it around 31mm while wikipedia says the chip coverage seems to be around 30mm.

I'm curious on your thoughts... considering Angeinuex, Cooke, and most Ziess cinema PL glass only covers 30mm while just a few 31mm... since RedProPrimes are 31 or something like that, do you think they will put it over 30mm to simply try and get more people to buy thier lenses for Epic use?

I'd hate to see all those great lenses not being used on 5k epics. I think it would be stupid for them to not make it compatible and keep the sensor at a coverage of 30mm.
 
Ryan Patrick O'Hara said:
Otis, what do you think about the coverage area for 5k on Epic. No one has released any real facts yet. Some guy has a chart which has approximated calculations, which puts it around 31mm while wikipedia says the chip coverage seems to be around 30mm.

I'm curious on your thoughts... considering Angeinuex, Cooke, and most Ziess cinema PL glass only covers 30mm while just a few 31mm... since RedProPrimes are 31 or something like that, do you think they will put it over 30mm to simply try and get more people to buy thier lenses for Epic use?

I'd hate to see all those great lenses not being used on 5k epics. I think it would be stupid for them to not make it compatible and keep the sensor at a coverage of 30mm.
Click to expand...

With so many pixels, a little crop in post will not be a problem. Red could offer a mode with a slightly reduced resolution perhaps.

I haven't found a 2/3" lens (9.6mm wide in 16:9) that will not cover Drama (10.56mm wide in 16:9) yet. That's 10% extra. I don't think the 3% from 30 to 31 will be a problem.
 
F1.4 is the standard optical block aperture for a 2/3" 3-chip HD camera without the lens. Most of the 2/3" EFP zoom lenses fall in the F1.8/T2 to F2.9/T3.1 range.
I don't think you can extrapolate relative sensitivity from pixel size alone for any two cameras. Two different Sony cameras, the PDW700, and the PDWF800 measured 1600 ISO and 800 ISO at 0 gain as tested by a cinematographer for HD pro magazine. These cameras have the same sensor and optical block assembly but different processing circuitry.
Canon's new XF300 camcorder outperforms the Sony EX1 in low light too for both sensitivity and noise despite the smaller pixel and smaller sensor size.
 
Razz16mm said:
F1.4 is the standard optical block aperture for a 2/3" 3-chip HD camera without the lens. Most of the 2/3" EFP zoom lenses fall in the F1.8/T2 to F2.9/T3.1 range.
I don't think you can extrapolate relative sensitivity from pixel size alone for any two cameras. Two different Sony cameras, the PDW700, and the PDWF800 measured 1600 ISO and 800 ISO at 0 gain as tested by a cinematographer for HD pro magazine. These cameras have the same sensor and optical block assembly but different processing circuitry.
Canon's new XF300 camcorder outperforms the Sony EX1 in low light too for both sensitivity and noise despite the smaller pixel and smaller sensor size.
Click to expand...

The single sensor 2/3" digital cinema cameras have no prism, they can use film lenses and any aperture. f1.2 f0.95 etc. f1.4 is quite common.

Sensitivity has nothing to do with image quality. It's just a decision by the manufacturer and it tells you nothing about the image quality of each setting. The 0dB setting is relative itself, that's why many cameras have negative gain settings. CMOS cameras operate at high gain, analog and digital, so 0dB is even more relative on those. The basic sensitivity of a consumer CMOS/CCD on a small digicam is no better than an ISO20. They don't even have that setting on the menus, the lowest ISO is noisy already but they have ISO1600 or ISO3200:) That's marketing.

The image quality of a system depends on the pixel size. This is universally acceptable in the scientific community even though there are some differences in the noise between different sensor types. Why would a Scarlet 2/3" pixel be better than a Canon pixel or a Sony pixel of the same size? I would actually expect the opposite based on the resources these companies have in design. The influence of the pixel size is so strong that you can estimate the quality of a sensor without even testing it by its pixel size. It's not something that needs a debate. Sony EX1 appeared to be more sentivitive than it should be, but when changing gain you can see how much noise reduction is active. Canon 5dII also appeared to be more sensitive that it should be for its pixel size but leveling the field by using ACR shows its performance is just what you would expect from the pixel size. They are all photon measuring devices using similar affordable CMOS technology that is common today. Don't expect any miracles.

Cameras will have small differencing in procesing and noise reduction, but nothing major:

Canon XF300 assuming typical 1/3" sensor size: 2.5um pixel
Sony EX1 assuming typical 1/2" sensor size: 3.75um pixel

Light gathering advantage of EX1 due to pixel area: 2.25x

F1.6 for XF300, F1.9 for EX1.

Light gathering advantage of XF300 due to maximum aperture: 1.41x

Combined light gathering advantage of EX1: Only 1.6x. Less than one stop, close to 0.5 stops.

I posted a sample from the slashcam low light test above. The theoretical 1.6x advantage is visible in the image and it looks like more than one stop in image quality (notice the blue noise of the 1/3" XF vs the 1/2" EX on the vinyl record). I bet Sony also use heavier noise reduction so the difference looks even greater than 1.6x. In that image I composited the result of this test for 5 different consumer camcorders and the larger pixels do have an advantage. I always found this to be true in any commercial sensor I have tested for my project. Pixel size is a good guide for estimating true image quality (not "sensitivity" or noise at "0"dB).
 
My old Canon GL2 camcorder has an effective ISO range at zero gain of 50-200 with 100 ISO the sweet spot for daylight balanced lighting. It falls apart quickly if you push it beyond 200 though, even with gain added.
I have always shot video cameras manual exposure like film after calibratng for ASA/ISO rating with scopes and a light meter, then fine adjusting the rating for best overall visual dynamic range. Typically this means under exposing by 1 to 1-1/2 stops compared to the auto-iris settings for better highlight retention but still retaining decent noise performance in the shadows.

It is going to be interesting to see where Scarlet 2/3" falls, but given the progress with Red 1 and MX I expect it to fall around 320 ISO native with a useful range from 200 to at least 1600.
R1-MX is showing useful range out to 3200 or better as the software improves.
 
Razz16mm said:
My old Canon GL2 camcorder has an effective ISO range at zero gain of 50-200 with 100 ISO the sweet spot for daylight balanced lighting. It falls apart quickly if you push it beyond 200 though, even with gain added.
I have always shot video cameras manual exposure like film after calibratng for ASA/ISO rating with scopes and a light meter, then fine adjusting the rating for best overall visual dynamic range. Typically this means under exposing by 1 to 1-1/2 stops compared to the auto-iris settings for better highlight retention but still retaining decent noise performance in the shadows.

It is going to be interesting to see where Scarlet 2/3" falls, but given the progress with Red 1 and MX I expect it to fall around 320 ISO native with a useful range from 200 to at least 1600.
R1-MX is showing useful range out to 3200 or better as the software improves.
Click to expand...

1.5 stops makes sense. The GL2 did not have any advanced gamma curves. It was normal gamma, no highlight protection at all.

The 2/3" model does not need to be as fast as the s35 model, it will have deeper DOF in wide apertures. 2 stops slower is ok IMHO.

Lets set a requirement for the fixed lens Scarlet: Due to the f2.4 limit, the 2/3" fixed lens Scarlet will need to match the quality of:

- An f1.9 EX1 while operating at an ISO 1.5x higher (with a pixel 1.29x smaller). This requires it to be 1.94x better than the Sony Exmor in sensor technology. It's only one stop, but exmor is quite good.

- An f1.6 XF300 while operating at an ISO 2.25x higher (with a pixel 0.57x smaller). This requires it to be 1.28x better the Canon sensor in sensor technology. This is probable. The Mysterium X 5.4um sensor used on Red One performs quite well for its pixel size.

- An f1.4 2/3" Kodak while operating at an ISO 2.94x higher (with a pixel 2,78x smaller). This requires it to be 8.17x better in sensor technology. I believe this is impossible, it's a full three stops, ISO500 vs ISO4000.

In the shallow DOF department:

- The f1.6 XF300 is equivalent to f3.4 in the Scarlet sensor size (vs f2.4).

- The f1.9 EX1 is equivalent to f2.7 in the Scarlet sensor size (vs f2.4). Practically EX1 DOF.

- The 2/3" will obviously be superior than f2.4.

This is why I believe only the interchangable lens Scarlet makes sense.

But if Red manage to bring the fixed lens Scarlet to market at 4750 USD (less than 3700 euro!) it will be a miracle. Think about it. A 2/3" camera with zoom lens from a small company, in an expensive shell, with an intraframe format, using high quality components and connectors, with a sensor specifically developed for one product, at a similar price to an entry level 1/3" from Sony/Canon/Panasonic that reuse sensors and components in many different systems, have a wide array or VLSI technology and licensing agreements, source parts for peanuts in very high units and operate at very low margins. If we hadn't been hearing about the fixed lens scarlet for years, I would say you are pulling my leg.

EDIT: If I were Jim Jannard I would bring it to market, even if I lost millions in the process and in every unit I sold. It doesn't feel good to dissappoint people and Jim can afford to do it.
 
Last edited:
Michael Olsen said:
The increasing likelihood of a real competitor has changed my thinking from "I'm saving for Scarlet" to "I'm saving for the next camera which will give me what I want".
Click to expand...

Amen brother!
My money is on Canon - they came late to the original 24p DV party, and late again to the HD party.... but they sure did deliver, IMHO.
I still own and regularly use my XL2 along with my HV20, they have served me well and frankly I've held out on upgrading because of the Scarlet announcement.

The AF-100 and the VG-10 signal the entry into the 35mm camcorder game for 2 of the big 3...... Canon's participation is inevitable, albiet late.
Canon won't cripple 24p on the entry-level model as Sony has done with the VG-10, if the HV20 and its kin are anything to go by.
And I bet Canon includes Zebra and manual everything on the entry-level model - just like the HV20, even if it is only a crappy 1/8" audio jack input - that's still useable with a Beachtek or similar.
Here's hoping Canon devises a short flange distance mount similar to m43 or Sony E which can then be adapted to any longer mount, and slaps it on the 35mm HD Canon camcorder so I can use my Nikon glass on it!

I'm brand-agnostic - all I really want is something like the HV20 in the $2k ballbark but with a large sensor and an adaptable short lens mount that will take my Nikon glass and manual EVERYTHING, and whoever manufactures it first will get my $$$$ and my vociferous word-of-mouth on-sell to my collegues. :Drogar-BigGrin(DBG)
 
I will be interested to see your camera when it is finished too Otis.
If Jannard didn't have a solid track record and the resources it would be difficult to believe.
So far Red is blazing a pretty unique trail with Scarlet, but it will be one option in a broad market and not necessarily best suited to everyone. I'm opting for the interchangeable version with a range of 16mm lenses. But then I like the idea of seeing how this vintage glass performs and creating a useful reference database for others that might be interested in doing the same.

I am more of a zoom kind of shooter, hate changing lenses unless absolutely necessary. But I have a few interesting primes too, mostly wides down to Century 3.5mm fisheye.
 
As has been said time and time again, everyone fully expects the big companies to catch up. Red are not sleeping. They have been working tirelessly over the last 2 years. It is a very complex project - bringing a RAW digital cinema camera to a $5000 market. There have been challenges - several unforeseen. No one really knows when it will release. I just hope it is 2010.
Click to expand...

I wish that people who talk so highly about Scarlet, a camera that doesn't even exist, took a wild guess and figured out why the big companies haven't "caught up" to this wonderful new technology. They haven't come out with something similar because all Red did was offer pie in the sky. A little camera so revolutionary that was only in their minds at the time two years ago. The big companies haven't caught up with the technology because the technology is in its very infancy for everybody. The whole concept is in development and it will take a long time for everybody, from Red on up to fully implement and deliver to market such kind of camera. Anybody can promise a car that flies out of your garage. But to deliver on that promise is another story. To create a camera with the Scarlets specs takes a lot of work and effort. Red, Panasonic or any other company will have a lot of work ahead to deliver on it.
 
Status
Not open for further replies.
Back
Top