Even if all the goodies are added to both cameras and the price came out to be the same I still don't see just how this Nox camera is all that great compared to some of the other options.

Nox=8 bit
Red =10 bit

Nox=1.2" CCD
Red=24.4 mm CMOS or almost 1" wide

Nox=max of 25fps for HD or 2k (guess nobody will be using it for 30p or 60i production)
Red=max of 60 fps at 4k

Nox=2K
Red=4K

Nox=720p can do 37fps (no 50p or 60p HDTV production)
Red=720p can do 120 fps

Nox=2 audio channels
Red=4 audio channels

Ok both of these cameras are using a high quality single chip bayer pattern design so how can the Nox look so much better then the Red? Even if Red and Nox happen to cost the same when they are decked out you can do so much more with Red. At best I would say when looking at 2k quality the Nox could only be very very slightly better but so far the images I have seen do not show this. Why is it that the Nox is "cinema" while the Red is "video" if they both come from the same style of chip. About the only thing in the specs I see that is better then Red is the size of the chip. Even then it is only slightly larger so I doubt it would make that big of a deal.

One more thing, I want to wait and see how both perform in the field vs. the specs. I've always operated that way; otherwise, I'd freak out when I first saw the compression of HDV (19 and 25 mbps).

heath

Marc, I am not trying to be overly critical here but here is my honest opinion.

From looking at the images, it seems like the Red camera would produce better quality images. The de-bayer algorithm used in the sample images is not very good (in my opinion) and introduces color artifacts and other weirdness. It does not look as clean as Red footage.

Dynamic range looks similar (Red is ~12-13.3 stops at 1080p, though it depends how you count), and the images are fairly artifact-free (clearly better than NoX).

2- Workflow-wise, the Red project looks pretty good with their Redcine conversion tool and Final Cut integration.

SI serves the PC market well with native Cineform editing.

The Sony F23 fits right into existing tape/SDI-based workflows. You don't need to learn a new workflow.

I don't see the workflow advantages of Nox (unless Noxbox provides an easy SDI-based workflow). Perhaps I am missing something?

3- I simply don't see why you wouldn't get a Red camera instead. Could you perhaps provide some advantages / differences this camera has over its competitors (Red, SI, F23, Genesis, Dalsa, etc. etc.)?

The market is flooded. I can guess 10-100K people for a professional camera, 1-10K for a good film camera.

At the price of the Red I can see 10K of customers, with an ENG camera setup for pro video and Television stations many more people might also buy (not saying they will achieve this against the competition). With SI, color space, the Russian company, and others, the market is getting crowed at those prices. So, the option is to sell ay an cheaper price to those that don't want to spend $20K+, or more expensive to make money from the limited quantity of sales. If the new camera can be treated as an ENG camera as well, maybe they can pull it off in the ENG market.

For cheaper good cameras, prices points of $13K, $8K, $5K, $3K are still available for the, likely bulk, or people that can't afford $20K+. The SI camera would be much more suited at $13K, $10K or less. I think the inclusion of HDMI uncompressed, and sudden move to film like quality, on lower end cameras is an deliberate move by camera companies to head Independent Digital Cinema Cameras off at the pass, and dent sales. A lot of low end people that happen to be after extra quality will settle for the quality of the low end XDCAM (and Panasonic AVCHD Intra when available at low end) and the HDMI solution, with professionals stubbornly avoiding this and preferring high end camera brands. The move to higher end pricing and targeting the film market correctly is probably a good move while the old time film people are still around.


Marc

Who are the people behind this camera, is the Drake people involved, or Rai?

The F23 is one of the hottest cameras right now. That's all I hear Cinematographers talking about. I'm hoping DVi will have topic areas for cameras like the F23, Genesis, etc. (subtle hint, grin)

heath

How many times has RED delayed their shipping start? I cannot even count anymore.

Zero. We've never said we would be shipping and then delayed it. We've announced engineering targets which we've almost always hit.

At NAB we did announce a tentative shipping schedule.

We are doing things different. It's clear that you're not a fan of this approach. That's fine. We are developing the camera together with our (potential) customers and we'll ship when we are ready.

If you don't like that then buy another camera or wait till it is out.

Zero. We've never said we would be shipping and then delayed it. We've announced engineering targets which we've almost always hit.There are some companies with terrible reputations for announcing dates and missing them by a mile. RED is not one of those companies.

I, for one, have been really impressed with the aggressive approach RED has taken - along with their ability to hit those aggressive targets.

Hi @ all,

here are some more answers to the questions that have been raised:

Q: ...and only does 8 bit, which hampers the 12-stop latitude claim...I'd rather have compressed 10-bit than uncompressed 8-bit.
Q: Is 8 bit even acceptable for digital cine?
Q: Ok both of these cameras are using a high quality single chip bayer pattern design so how can the Nox look so much better then the Red?

A: Of course it is acceptable. Explanation: if you use 10-bit material you have four times the colour information and you are more flexible when you do colour grading. Now here comes a big BUT:
noX records 8-bit uncompressed, which is NORMALLY not as good as 10-bit. I explicitly state “normally” because with noX things are different.
In practice noX footage has proven to be extremely good and easy to grade. The only thing you have to bear in mind is that you use the right LUT on set (which is very easy handled).
This may sound a little bit esoteric, but our 8-bits are “somehow better”. Difficult to describe why, but this is what people told us that worked with noX material. You will see this with our clips and new stills which will be published near-term.

Q: Not to be an ogre but how is this any different then Red or the Silicon Imaging camera?

A: We don’t arrogate to judge cameras from other manufacturers. This should be done by the people working with these cameras. Every camera has its pros and cons, so the user finally decides which one is the best for his requirements. Let’s wait until several filmmakers worked with both cameras and how they assess them.

Q: Maybe I'm the only one here, but from the sample stills that these guys have posted on their website, I'm really not impressed. Where is the high contrast? I would really like to see some tight mosaics. What about some low light shots?

A: Regarding the sample footage we are currently working on new material, which will show more of the capabilities of the camera. This will be published soon on our website.

Q: What about a Apple workflow? You mention still image sequences and AVI files but what about quicktime files?

A: The Apple workflow is fairly simple. You record uncompressed RAW files with the noXboX and then render single frame tiff files directly to your Mac using SDI. Once the footage is on your Mac you can easily transform it into Quicktime files.

Q: What is the storage medium?

A: noX records on three 2.5” notebook HDD. They are to some extent shock-resistant and much more reliable than standard 3.5” HDD. If you use the “fail-safe” option (in the software) your footage backs up automatically, so in the unusual event that one of the three HDD should crash there is no risk of loosing any footage.
A standard noX package allows you to record 90 minutes of uncompressed noXRAW footage or 45 minutes in "fail-safe" recording mode.
As an option we offer completely shock-resistant flash drives with a running time around 16 minutes.
After recording to HDD you can backup your footage using the noXboX (Digital MAZ for noX) which uses a RAID 5 server.
Furthermore, any part of your footage on this RAID 5 server can be written to a inserted SATA-drive with one touch at any time.

Q: if you buy a Cinealta the body alone is around 70.000 - 100.000 dollars, also WITHOUT LENS, and storage.
If you buy the camera that's anounced here ,it's 50.000 dollar, ALSO WITHOUT storage and lenses. See the point?
Professional cameras ALWAYS come without lens, storage, displays,...
EVERY camera always costs a bit or much more with all the accessoires with it...

A: Partly correct. The noX comes WITH storage, battery and display together with the basic camera. The only thing you need to start filming are lenses. As we found high quality cine lenses to be most complementing to the pictures of noX, it is more sensible to rent them rather than to buy them for yourself.

Note to Chris:
You are right. It wasn’t meant to be directly targeted against any other company or person. If anyone was offended by my statement, I truly apologize because this wasn't my intention. But as we in Germany say: “You only wear a shoe that fits”.

Marc

What a great reply under such immense judgement from the members here - it makes me quite contended to see that people are pushing out cameras (whatever their target market might be) and providing people with further options to consider and tools to use to elucidate their visions; whether this cam will 'lose' to the RED and SI cams or not is yet to be seen but im sure there are filmmakers out there who will find this camera the perfect product for them.

And in response to the whole idea that DVINFO only really serves a low end target market - i think that for now that is true however the notion that the lines between the high end, high budget filmmaker or a low budget indie composer are becoming blurred with the introduction of ever cheapening cameras such as RED and SI - and even the noX is 20,000 dollars cheaper than one of its counterparts - is really reflected on this forum and you can really see that indicated in Chris's consistent inclusion of these newer (albiet cheaper) projects that are aimed squarely at proffesionals - not the 'dreamers' - cutting a long story short i think we are headed in the right direction.

Marc,


Are you guys using an Optical Low Pass Filter? Feel free to email me if you don't want the answer on a public forum :)

aaron@colorspaceinc.com

What a great reply under such immense judgement from the members here - it makes me quite contended to see that people are pushing out cameras (whatever their target market might be) and providing people with further options to consider and tools to use to elucidate their visions; whether this cam will 'lose' to the RED and SI cams or not is yet to be seen but im sure there are filmmakers out there who will find this camera the perfect product for them.

And in response to the whole idea that DVINFO only really serves a low end target market - i think that for now that is true however the notion that the lines between the high end, high budget filmmaker or a low budget indie composer are becoming blurred with the introduction of ever cheapening cameras such as RED and SI - and even the noX is 20,000 dollars cheaper than one of its counterparts - is really reflected on this forum and you can really see that indicated in Chris's consistent inclusion of these newer (albiet cheaper) projects that are aimed squarely at proffesionals - not the 'dreamers' - cutting a long story short i think we are headed in the right direction.

Tyson,

I agree, 'it's all good'. The more choices and competition the better for us, the consumers. I don't think that NoX or Marc is getting unfairly hammered. Sure people are asking tough questions, but I seem to recall RED getting scorched pretty good in the early days. If you want to sell an unproven product for 50-70k, you've got to expect some scrutiny.

yes you do - but im saying that i was glad that he remained a calm as he did under this scrutiny - and i am also not denying that he does need to answer some very very tough questions.

The problem here was the original presentation... "A new star rises..." and then a (seemingly unintended) stab at RED marketing.

I guess the best question is... what does the noX do that RED cant?



ash =o)

Finally, digital cinema has reached the level of film in matters of picture quality.

Sometimes it's not about counting the pixels, but making the pixels count.

What is the noX? She is a full HD / 2K cinema camera and she produces astonishingly film-like pictures.

Most current HD cameras take pictures which look like video, so we created a digital alternative for people using film.
Furthermore we wanted to find the fine balance between the resolution which is needed to look good on the big screen and the resolution which is still manageable without collecting loads of terabytes.
Finally we wanted to create a workflow which is easier, faster and more flexible so there is more time and energy to focus on what really counts - filming visual dreams.

In all sincerity: we did it.

Many HD cameras claim to produce beautiful film-like pictures, but to be honest – not many are able to keep this promise. Check out the sample stills of the noX to see that digital pictures can look like film, especially regarding skin, hair and DOF. Stop doing video – start doing photography again.


Marc O. Hardt

Marc,

Sorry to chop up your original post, but it seems to me that most of your statements above go unfounded. In today's world outsouced manufacturing and a broad array of available components, it wouldn't be very hard for any company to release a digital video camera like yours. I think there could very well be more competition for this segment in the future.

That said, what is it that makes the images that this camera produces decidedly film like? Many productions have used HD cameras, HDV cameras and Film cameras seamlessly together. How does your camera produce images that don't look like video?

You created a workflow? You have said nothing on this since this post.
Though your camera looks to be about the same size as a film camera, and honestly a bit larger. What is a camera op. and asist. used to an arri or panavision going to think when they approach this camera? What about steady cam or handheld set-ups?

The nice thing about 35mm is it is completely scalable, you use the same film stocks and lens setup no matter what camera your shooting with. I think Red can offer the same scalability with its modular design. Arri and Panavision both have their own digital cine cameras, that look and act very much like their film cameras. Why did you guys go with an lcd? This would suck on a bright day or bright set. Would there be a viewfinder option for the camera op? Why would DPs and cam ops choose your camera?

Are you guys using an Optical Low Pass Filter?
I don't believe they are. That's why it looks so aliased.

I decided to show this photo here, it is an private photograph, all copyright applied, but it may illustrate something I think they maybe getting at.

This is from an old camera I tested a few years ago, with up to 20 stops of latitude. It was taken from my place, the blurriness is an quirk on the shot. The exposure is similar in appreciation to exposures you get in Hollywood cinema. It was close to noon on an extremely bright summers day, around an 1000KM from the equator (stacks brighter than what you are normally used to in Europe/States). The leaves on the cane, to the side, normally have large broad burnout on an consumer camera, and are quiet glary/white reflections to the eye, but are well exposed here, while maintaining an lot of features in the shadow under the mango tree (turn up the brightness ion your monitor to see).

This is taken using an latitude extension feature sadly neglected in Cinema cameras around here. The old camera, is an maybe $69 credit card camera, with fixed set lens. It uses one of the cheapest sensors (in dollars) and one of the worse I have seen on some measurements of the earlier age of sensors, with SN of 30db from memory and plenty of image artifacts at times, but 20stops of latitude when it wants to. It is an shame not more expensive Cinema camera manufactures could at least take advantage of cheap technology like latitude extension, even when the senor had it. Puts the prices in the Digital Cinema realm into an bit of perspective.


Thanks

Wayne.

Wayne: very cool.

Perhaps the video equivalent of the Holga camera? (A medium format still camera that is charming because of its technical flaws... vignetting, softness, etc.)

Embarrassment, blush :) , it is only the exposure latitude (and maybe color) we are looking at here. Looks like I bumped the camera on closer examination. It was one of the senors I was examining for the Digital Cinema camera projects, but the Fill factory Ibis5a offered around 60db SN at that stage (when using external ADC circuit that is properly setup) and 90db of latitude extension. Now days that is all old hat, and there are better cheaper sensors.

Wayne, when you say, "latitude extension" are you referring to an antibloom on the chip? The main problem with antibloom is that the only chips that can utilize it effectively generally have a small fill factor to begin with. Antibloom also typically involves a trade off in full well capacity, which when you already have a low charge capacity, will result in less exposure latitude.


Or is this something else? 20 stops of latitude sounds great!

No not that, I see this feature more on CMOS sensors, but it is features like dual/multi-slope and autobrite, as two examples, that allow an much greater number of stops. The sloping technology exposes the pixel, then resets it for an different exposure time on each pass, the two results are combined on chip into an new pixel value. Autobrite, from what I can tell, adjusts the gain (up or down) of the pixel to bring it into exposure. For the dual-slope Ibis5a had reasonable fillfactor (less microlens) and good well capacity, for an cmos sensor. QE and SN ratio, the Altrasens scores better in cmos sensors of the time (well it got delayed an lot).

Interesting to note, that the sensor company in the camera above, and the Fillfactory company that makes the Ibis5a sensor, are now owned by Cirrus Semiconductor.

I have only read about dual slope (and not much at that), so I wonder if you can comment on how it effects uniformity, linearity, and frame rates.

Can't remember too much, it was 3 to 4 years ago I looked at it. With different exposures, of the slopes colors respond differently, so it does nit look right (I wonder if color correction can be applied). Obviously there is going to be different performance throughout the range, but there is an technology that retains conformity. The technology requires correct setup so it blends together nicely and doesn't jump. But Andrey can tell you much more accurate detail then I can, he used to use the Ibis, but was not impressed with it's internal ADC, that could not produce sufficient enough quality for him.

I remembered last night, I think there is an anti-blooming use of multi-slope.

Sounds to me like a version of antibloom where by the charge which is drained to avoid blooming is recorded so that the full-well capacity is now full-well + charge drained. It sounds to me like it would reduce uniformity, linearity and probably frame rate, as well as adding another transistor possibly reducing fill factor. But the trade offs could be worth it, especially since highlight have always been a weak point with imaging sensors.

I suspect it might decrease maximum shutter, but you can still do normal frame rates. Linearity, what was that again, am I assuming right? I think it works by reseting the pixel.

Linearity is essentially a euphemism for predictability. But yes, that sounds correct to me.

This kind of extreme latitude extension is used in some security cameras. Each frame is taken twice at different shutter speeds and then the DSP merges them. Very usefull when at night you have a car with headlights driving towards the camera and you can still read the number plate. I have no idea how well this technique would hold up in terms of overall image quality but I've been tempted to buy one of these camera just to try it out.

It is better if it does it on one pass, somebody suggested doing it the other way on the Digital Cinema thread discussions before. But, this was my point, is better tonality one of the things they get, is it from dynamic range/SN, and the other things (35mm DOF, fill factor).? You can beat out an better sensor through these things then plain small 8+ Mpixel pixels. I wonder if they are using an FIllfactory sensor from Cirrus Logic.

We haven't heard from them, for an while, back on discussion.

www.hdrc.com

Yes, that is one of the ones on file.

I agree with you, price is an sticking price in the Digital Cinema camera market for most around here.

An 2K camera can be made with significantly better performance than Red, except in resolution, or Silicon Image. So, until I see exactly what they are up to I am not discounting their claims of offering something more. As we can see above, even cheap technologies have something more to contribute to the cinema camera market, over what is presently being done.

Maybe they have worked out a good effective work flow, and don't need support of Apple, or popular NLE's. The cinema film camera industry got buy with specialist digital work flows for sometime, and this is the crowd they are chasing. I would like to hear what they say, before they run away (which has happened with other cameras before).

Technically, all this is already available, but spread across different cameras, with aftermarket mods (eg, Hydra, DOF adapters).

Here's a thought:

Three half-inch CMOS sensors capable of extreme spatial offset in order to take Hydra, which is built in, not aftermarket.
Removable half-inch zoom. It could be essentially the same one that would have been built in as a fixed lens, but modified to be removed.
HDMI 1.3a, with full support for higher bit depths.
Progressive 1,920 x 1,080, with support for 0-60 frames per second.
AVC-Intra, 100 Mbps, 4:2:2, for normal operation, HDMI or Hydra for 4:4:4 uncompressed output.
DOF adapter, with stationary advanced plastic screen (a la the new Movietube) and fixed achromatic relay lens, designed especially for the camera.
A shoulder-mount-style viewfinder.
Why both a half-inch zoom and a DOF adapter? Well, I'd suggest that you don't always want shallow depth of field.

I find this all fascinating. Candace, I'm curious, what would you do with 0 fps, J/K. It seems like apple will support anybody who gives them an exclusive contract, possibly with a future buyout plan? Lawrence, sorry for my ignorance, what is Hydra? Is there a link to another thread you could give me?

I see both of the specs you guys have posted as pretty specialized towards the indie market. The indie market is pretty small compared all of the other various forms of media production. If Sony or Panasonic could make a decent buck by marketing a cam with a 35mm lens mount, they would, but they can't. The reason why RED could be succesful, in my opinion at least, is not there price point, but the fact that they let Peter Jackson shoot sample footage for NAB. If you can get some major names to use your camera, then people will want to use your camera. I think the ammount of braodcast, theatrical and distributed material shot with the Panavision and Arri digital cine cams will vastly outnumber red, if only becuase DPs, camera ops, and ACs are used to the way these cameras work, and because productions will continue to have the budget to support these cameras. I would hope that the majority of the Red cameras sold will have PL mounts and be equiped with real motion picture glass attachted to them.

The beauty of Red is not just the specs of the cam but also, and this is where I think it's appeal really takes off, that it's designed so that modular elements in the camera can be replaced with upgrades. A new chip comes out, replace it rather than the whole camera! This in my opinion is what makes Red viable!

There are just many things why RED is so fascinating.
Unbeatable price point (considering the quality), the support, the upgrades, the 35mm sensor, the modularity, resolution, workflow,...

And RED is a viable camera. I don't want to sound like a fanboy, I don't have the money to buy a RED, but the REDcamera is real, and it's coming.
Peter Jackson didn't shoot his 12 minute short with thin air.

And looking at what the RED team did in 15 months, it's very normal they got a little delayed.

[W]hat is Hydra?It is a modification to the Panasonic HVX200 (other cameras may be Hydra-ized in the future), done by Reel Stream. A succinct desciption can be found at the beginning of the Hydra thread (http://www.dvinfo.net/conf/showthread.php?t=89958).

Why not just have high res chips to begin with? Pixel shifting isn't ideal. If you are going to make a camera from scratch, why follow the example of a modded camera.

Why not just have high res chips to begin with? Pixel shifting isn't ideal. If you are going to make a camera from scratch, why follow the example of a modded camera.A few reasons:

Through radical spatial offset, Hydra's resolution exceeds (albeit only slightly) 1,920 x 1,080 and may reach or exceed 2K. In effect, the three CCDs become one big virtual sensor, even though, physically, the imaging area is still third-inch and therefore uses the built-in zoom.
Although there are small 1,920 x 1,080 chips (the Canon HV20 has one) and a half-inch chip should have more than twice the surface area of a third-inch one, there is still a trade-off between sensitivity and pixel density -- you could opt for greater sensitivity with a half-inch chip, or greater pixel count, or some of each. Maybe with improved technology you could have a lot of each.
You probably wouldn't want to increase native resolution but eliminate a design's ability to pixel-shift. I believe that Reel Stream's Juan Pertierra has said that although the JVC HD100 could be modified to deliver 4:4:4 uncompressed data, it couldn't be pixel-shifted beyond its 1,280 x 720 native resolution. Spatial offset, therefore, can be used as a hedge against the future.

Say folks, we have drifted far off topic here... the subject is not RED nor Hydra, but rather the noX. Thanks in advance,

I think that some folks would like to see Nox differentiate itself from its competitors (Red, SI, F23, etc.) and/or step out of their shadow. You have to ask yourself why you would get this camera when some of the cheaper products seem to do the same thing better.

I think Nox is claiming that they will do it, from an cinema filming perspective, better, and I would still like to see exactly what they mean.

Ah yes, it's not about counting pixels, it's about making pixels count. All this talk about RED's advantages has been in the numbers: Price, Resolution, Bit depth, etc. NOX is claiming that the numbers don't matter, it's the pictures that do.

I can't claim to notice a very noticeable difference between d-9 10 bit 4.2.2. and dvc pro 8 bit 4.1.1.

This is not to say that the d-9 wouldn't be more suitable for intense effects or compositing, but bit rate aside, either format, when exposed properly is going to produce a beautiful image.

On a similar note, I have an SACD player and honestly, I have some extremely well recorded cd's that do sound better than some reprints I have on SACD.

Sometimes the numbers do not tell the whole story.

Maybe the numbers don't matter, maybe 8-bit 2k raw does offer a significant workflow advantage over 10bit 4k compressed. I don't know, but I haven't seen any post by a NOX rep that would indicate what they mean by saying: "...it's about making pixels count." And neither have they explained anything about their workflow, other than the fact that they have one. This won't be the be all end all camera, we all can agree with that, but whose to say that there aren't a few skilled DPs out their that wouldn't like to give this thing a run.

If you look at sample images from both cameras, Red to me looks better in particular aspects.

The debayering in Red is significantly better. It doesn't have the weird zippering artifacts that the NoX images show. And the bad thing about the zippering is that any keying or secondary color correction will pick that up.

Noise-wise, the Red footage looks cleaner.

Red of course has more resolution. Nox does looks sharper, while the Red looks blurry at 4K. When you downsample I think this difference goes away.

Otherwise, the images look similar to me (aesthetically).

2- But to get back on topic... I think people would like to know what differentiates Nox from its competitors? (You don't have to mention particular competitors.) This question wasn't quite answered, other perhaps than to say that the footage is easy to grade (but it looks like one of the competitor's products is even easier, so...).


Q: Not to be an ogre but how is this any different then Red or the Silicon Imaging camera?

A: We don’t arrogate to judge cameras from other manufacturers. This should be done by the people working with these cameras. Every camera has its pros and cons, so the user finally decides which one is the best for his requirements. Let’s wait until several filmmakers worked with both cameras and how they assess them.

Good points Glenn. The demosaicing is indeed less than perfect, and coupled with lack of the necessary optical low pass filter produces a range of artifacts that just don't look good. I'd be very keen to see some NOX images with an OLPF, or as raw so that I can demosaic them myself.

Graeme