KOMODO....

[Bastien Tribalat]

I really wish everything would just go to CFEXPRESS cards already. 1400 max write and 1600 offload speeds. This will future proof the memory aspect of the camera and keep us from waiting around with long offload times. Plus the form factor is perfect.

No. Not really. It depends SO MUCH on the interface (USB 3 ? TB 2 ? TB 3 ? Voodoo magic ?), the bus speed, the computer storage speed and other computer components.

For comparison : last may on Cannes Film Festival (where I worked on a multitask gig part editor, part DIT and part camera operator for a Press Agency), offloading our CFAST 2.0 (from San Disk) with the Lexar Pro USB3/TB2 reader onto a workstation PC (i7 6900K, 64 GB of DDR4, GTX 1080, SSDs and so on) via USB3, the speed maxed out around 390 MB/sec. The card is rated at 500, the connection at 625 (5 Gb/sec).
We also had XQD cards. They were rated at 400 MB/sec. The transfer speed maxed out at 200 MB/sec. Same other conditions.

FURTHERMORE, CFExpress uses an XQD slot and maybe that slot is still a licensed property of Sony and one manufacturer might not one to pay those royalties.

https://youtu.be/3x9W2x8bweM

https://www.bhphotovideo.com/bnh/co...earch=yes&sts=ma&Top+Nav-Search&Ntt=CFexpress

Well, according to your B&H link, it's in preorder state and they don't have a release date.

CFAST2.0 are more established on the market, fast enough for 6K Redcode and most probably the slot isn't licensed OR is cheaper to use.

 

[jamesbridges]

Gotcha...
Jarred’s quote says it’s still a unicorn, looks like it is not.

If you watch the video it says it can replace XQD... seems legit

 

[Bastien Tribalat]

Gotcha...
Jarred’s quote says it’s still a unicorn, looks like it is not.

If you watch the video it says it can replace XQD... seems legit

Yes it can replace XQD. No doubt about that. BUT, the only manufacturers (that I know of, hey I'm just a user I'm no insider to any part of the industry ) other than Sony who use XQD are doing so on products that are using Sony sensors so maybe there's a relation (not technical, financial like "if you put a XQD slot you get a discount on the sensor") and there's no reason for RED to prefer using XQD (and therefore CFExpress).

Man, on the FS7, the XQD cards tale forever to format. On the C300 MKII, CFAST2.0 cards take maybe 2 seconds. Maximum. I don't know for CFEXPRESS since to my knowledge no one can buy it yet. I don't doubt it will be an awesome card tho but I do think going CFAST2.0 is cheaper for the manufacturer and is very reliable.

 

[Savannah Miller]

higher rolling shutter

Does this mean slower rolling shutter? I feel Z-Cam and Blackmagic are on par with a lot of RED cams these days. Rolling shutter was one of the main complaints of the BMCC and Blackmagic doesn't seem to have these problems anymore.

Ursa Mini Pro G2 has alexa-like rolling shutter performance.

 

[Brian Boyer]

So what do you guys even mean by this?

The sensor has a noise floor. Once enough photons trigger a signal above that you have something. After that, you'll need twice the signal to gain a stop, all the way up to FWS.

This is all you'll ever capture—a linear signal response to incoming light. This is the data, this is what's captured. There is no roll off here.

Now you're free to go into your software of choice and apply any OETF or log curve you like. "Change the ISO" too. You can take that captured linear data and push and pull it along a logarithmic function all day long—for presentation. Clearly, this will change how much information is revealed in the shadows (along with noise), or how much contrast/tonalily/compression you want in the highlights.

But this is our creative choice, right? What am I missing?

Interestingly enough, you answered your question in your explanation. FWS, or full well capacity, or however people refer to it, is the clip point on the high end, where the highlights are.

Some sensors clip at a lower light intensity, some at a higher one. Or, to use the other terminology, some have a smaller FWS and some have a larger FWS. The larger ones clip at a higher light intensity than the smaller ones. Is this not right?

It's the difference in the highlight detail (the number of stops above middle gray before clipping, or FWS, occurs) the sensor in a Sony NEX series can capture versus what Venice can capture.

We want the well to hold more photons before it gets saturated. Does that make sense or have I got it all wrong?

 

[jamesbridges]

Yes it can replace XQD. No doubt about that. BUT, the only manufacturers (that I know of, hey I'm just a user I'm no insider to any part of the industry ) other than Sony who use XQD are doing so on products that are using Sony sensors so maybe there's a relation (not technical, financial like "if you put a XQD slot you get a discount on the sensor") and there's no reason for RED to prefer using XQD (and therefore CFExpress).

Man, on the FS7, the XQD cards tale forever to format. On the C300 MKII, CFAST2.0 cards take maybe 2 seconds. Maximum. I don't know for CFEXPRESS since to my knowledge no one can buy it yet. I don't doubt it will be an awesome card tho but I do think going CFAST2.0 is cheaper for the manufacturer and is very reliable.

Makes total sense, I generally format my cards ahead of time(I use SXS Pro+ btw), my other cam op using a FS7 I’ll have to check in that one. With those write speeds, CFEXPRESS seems to be the future.
I’ll wait for Jarred to mention financials, sensor discounts, etc. for me, it’s total speculation, but your comment could be true. I wonder how much the Canon Mount costs?

 

[Christoffer Glans]

Where do you suggest the 'rolloff' component lives in the signal chain? How could such a thing exist in a digital sensor?

I'm not saying the digital sensor has such a thing, but that the OLPF filtering light to maximize the color science interpretation of the RAW data. If the filter manipulate the light transfer to the photosites, you will get a different read between different filters, just as you get with filters in front (or behind) your lens. The "sensor tech" improving highlights include both the sensor, OLPF and base color science. I think you misinterpreted the line of thought on the roll-off thing, never said it was specifically only at the sensor level, it's the sum of all parts. The highlight/skintone OLPF retain the quality in highlights so that roll-off in the RAW decode handles better. If you don't have the OLPF doing this, it's harder to do roll-off handling of those stops with any kind of quality to it. There's a reason we have a harder time getting the roll-off quality any good on the low-light OLPF, which is closer to what you refer as "letting it all in". BM doesn't even have an OLPF doing much, so they're closer to the low-light OLPF. Recovering and handling highligths with a much more pleasent roll-off requires those stops to be compressed in a curve that... rolls off. When grading I'm doing the same in shadows, but not the extent as done with highlights.

The key here is that camera manufacturers need to decide if they want to keep the stops handled as a straight stairway or roll-off highlights to create a more pleasent DR to the eye. I'd prefer handling the DR like that, then just giving it straight. It's faster and more easy to handle in post. This was the point I meant.

 

[Christoffer Glans]

Does this mean slower rolling shutter? I feel Z-Cam and Blackmagic are on par with a lot of RED cams these days. Rolling shutter was one of the main complaints of the BMCC and Blackmagic doesn't seem to have these problems anymore.

Ursa Mini Pro G2 has alexa-like rolling shutter performance.

My bad, yes, I meant a slower rolling shutter, i.e slower readout. The Ursa Mini Pro G2 is, in my opinion, a very impressive camera from BM. Should be praised more than people seem to give it. Pocket 6K, however, is not really close to the performance of the G2, since the readout of the Pocket is way slower. It's also the reason why it can't go above 50 FPS in 6K mode. Now, if Ursa Mini Pro G3 becomes a 6K 35+ or 8K FF, that would be some serious competition.

 

[Bastien Tribalat]

Makes total sense, I generally format my cards ahead of time(I use SXS Pro+ btw), my other cam op using a FS7 I’ll have to check in that one. With those write speeds, CFEXPRESS seems to be the future.
I’ll wait for Jarred to mention financials, sensor discounts, etc. for me, it’s total speculation, but your comment could be true. I wonder how much the Canon Mount costs?

Yes it is.
(only thing I'm mostly sure is that Sony has some sort of licence control -and monopol?- on XQD format and that it may apply to the in-camera slot as well)

SxS Pro+ are my favorite cards of all. I have NEVER had any problem with them, it's super reliable, in my experience it's build like a freakin tank. Only downside : it's expansive !

My bad, yes, I meant a slower rolling shutter, i.e slower readout. The Ursa Mini Pro G2 is, in my opinion, a very impressive camera from BM. Should be praised more than people seem to give it. Pocket 6K, however, is not really close to the performance of the G2, since the readout of the Pocket is way slower. It's also the reason why it can't go above 50 FPS in 6K mode. Now, if Ursa Mini Pro G3 becomes a 6K 35+ or 8K FF, that would be some serious competition.

I agree.
I think the fact that it doesn't have dual ISO, doesn't go higher than ISO3200 (and it's a bit noisy) and has a base ISO of 800 (I think) turns people off because nowadays and especially in this price range people tend to buy specs lists instead of an actual camera.
But when you think about it, this camera offers the same ISO settings (more or less) as an Arri Alexa (not LF, I've never used one so I don't know about those settings ).

 

[jamesbridges]

Makes sense Bastien, licensing hardware is a whole new world for me to try and discuss. Imagine if Sony makes licensing could be very expensive, if for sale at all w/o some serious strings attached. I am only speculating that the camera design was some time ago as well.
True, SXS Pro+ is fantastic, so far it’s been a very reliable recording medium for me...correct, downside-$$$

 

[AndreeMarkefors]

Does that make sense?

It's a bit of a tangent to what I said. A diode that has a greater well capacity will give us a stronger signal and higher dynamic range, given that the noise floor is constant.

The captured signal on sensor has no correlation to the distribution of stops around 18% gray. That comes later via OETF interpretation.

I think you misinterpreted the line of thought on the roll-off thing, never said it was specifically only at the sensor level, it's the sum of all parts.

Well... the sensor can't really tell if there is a filter in front of it, or the sensor. It just gets charged in a linear fashion. I maintain that it has nothing to do with any rolloff.

I agree that if you can filter out some undesirable characteristics of certain light wavelengths, you will get a cleaner result that will most likely help when you:

apply your OETF and any other curves and/or image manipulation tools/tricks that lead you to your preferred look. Of course it's possible to compress/extend/roll off highlights. It just doesn't happen on the sensor.

 

[Manfred Baulig]

Speaking of sucker punching, and I know it is going off topic but anyone know if DSMC3 will have a 70mm-ish sized sensor in its roster ?

Please...just don't. Aperture 8 and above here we come. We'll need a lot more of light to get anything sharp in the picture.
Also...why does everyone wants dsmc3????
Dsmc2 is still evolving with dsmc2 teradek module and now the one from vaxis. When they release a system it always takes around 1-2 years till all the important parts are available. You also can buy everything new...for those who really need it...they can release a 70mm sensor in the dsmc2. Form factor. But have fun with gathering enough light.

 

[Christoffer Glans]

It just doesn't happen on the sensor.

Which was never really stated. Sensor tech, as said, is the sum of all parts, not just the sensor itself. Point being, how the sum of its parts behave is the point made with what we prefer, no one is actually shooting and grading with pure RAW data, everything has been set to a specific standard and with the OLPF, some standards are baked into what the RAW data is capable of.

 

[Bastien Tribalat]

Please...just don't. Aperture 8 and above here we come. We'll need a lot more of light to get anything sharp in the picture.
Also...why does everyone wants dsmc3????
Dsmc2 is still evolving with dsmc2 teradek module and now the one from vaxis. When they release a system it always takes around 1-2 years till all the important parts are available. You also can buy everything new...for those who really need it...they can release a 70mm sensor in the dsmc2. Form factor. But have fun with gathering enough light.

Am I missing something here ? A 70mm sensor being roughly four times the size of 35mm sensor (for the sake of the argument) why would it need more light ?
You mean we would have to close the aperture a lot more to get depth of field ?
If so, doesn't seem to have been so much of a problem on Nolan's films or on Rogue One for example (which was entirely shot on Alexa65 and has, at least to me, a GORGEOUS picture). THAT BEING SAID, it's obviously true that it would make focusing a bit more difficult but it was never meant for one man crew. You need a (serious) focus puller

 

[Manfred Baulig]

Am I missing something here ? A 70mm sensor being roughly four times the size of 35mm sensor (for the sake of the argument) why would it need more light ?
You mean we would have to close the aperture a lot more to get depth of field ?
If so, doesn't seem to have been so much of a problem on Nolan's films or on Rogue One for example (which was entirely shot on Alexa65 and has, at least to me, a GORGEOUS picture). THAT BEING SAID, it's obviously true that it would make focusing a bit more difficult but it was never meant for one man crew. You need a (serious) focus puller

I come from photography and at aperture 8 nothing was sharp. I had to experience analog Hasselblad and the digital beginning. On some stuff for commercials we had to shoot at aperture 64.

All I'm saying is, it makes everything more expensive. Try Monstro on 85mm at 1.4. for product shootings we always choose the helium over monstro.

 

[Christoffer Glans]

I come from photography and at aperture 8 nothing was sharp. I had to experience analog Hasselblad and the digital beginning. On some stuff for commercials we had to shoot at aperture 64.

All I'm saying is, it makes everything more expensive. Try Monstro on 85mm at 1.4. for product shootings we always choose the helium over monstro.

70mm is not a problem for productions using the Alexa 65. Of course it's more expensive to shoot with such a camera, but so is using the Alexa 65. It's not meant to be used for casual stuff or low budget shoots, it's meant to be a camera used for the highest-end productions, while still available for those who want to shoot in medium format. Medium format productions usually use wider images, opening up the image without losing DOF. It looks fantastic, just check out Rogue One, Altered Carbon, The Revenant to name a few.

The whole point is that DSMC3 should be made up of three sensor sizes to be a rational lineup. 65mm medium format, Full-frame and super35. By having three sensor sizes, you choose the one which works best for your type of work. And Jarred already said that DSMC3 is coming, at earliest by the end of 2020. My point was that DSMC3 need to be a radical push forward in digital cinema, since alternative brands keep catching up with cameras that cost a fraction of a Red. Either, the DSMC3 lineup needs to cost much less in order to compete, or they need to be at the cutting edge of digital cinema technology. Until then, let Komodo disrupt the sub-10K market to get more people to use Red and be within that infrastructure. It makes logical sense.

 

[Bastien Tribalat]

I come from photography and at aperture 8 nothing was sharp. I had to experience analog Hasselblad and the digital beginning. On some stuff for commercials we had to shoot at aperture 64.

All I'm saying is, it makes everything more expensive. Try Monstro on 85mm at 1.4. for product shootings we always choose the helium over monstro.

When I was in cinema school (2008-2011) I shot medium format film (on Mamiya gear) but I can't remember any of the settings. Was awesome tho.
I shot a bit with the Panavision DXL (which has a Monstro VV sensor) and some Primo 70 whatever primes and it was fine by me. BUT it was NOT product shooting, just random people talking and drinking at Cannes Film Festival at the Panavision presentation of the camera (which for the fun fact was taken from the shoot of Mission Impossible Fallout for that day and had to go back overnight to where the film was being shot at that time).

My point : bigger sensor=more difficult to focus and stuff BUT on a cinema shoot (I mean with a full crew) it's not that much harder in my opinion and it looks so awesome (but that's a matter of personal taste. I do love using the lack of depth of field to bring... Depth to a shot I feel it's more effective than 3D)

 

[Brian Boyer]

Also one thing to keep in mind, is even though Komodo is around the corner, we don't know what Blackmagic is working on.

Because, if the competition, two years from now, release cameras with 8K full-frame, fast readout 14-15 stop sensors with superb color science and at least 60-90 fps max, but at a sub $10000 price-point, there will be a hard time justifying Red anywhere else but the very highest level productions (Disney/Marvel/Netflix/Amazon). Komodo could, however, be a disruptor to gain a market share before that, if done correctly.

How else can Red stay on top other than with brand recognition?

I know there has been a lot of talk in this thread about RED vs. BMD vs. Z CAM but I have a different (probably incorrect) take on the situation: RED is not in competition with either of those companies.

Insofar as they all sell digital cinema cameras, yes, they're technically in a broad competition. But it's not a competition in any real sense. The closest RED came to competing with BMD is with Raven, which is never more. I don't think Komodo is a Raven follow-up.

Disclaimer: my business acumen hovers somewhere around zero. However, in true internet forum fashion, I'm not going to let that stop me from pontificating on the subject. The way I sees it...

For the sake of this argument, a company's competition is based, in large part, on two things - who their target market is and whose products that target actually buys/uses. If your intended target isn't using your product, it's time to reevaluate your audience.

From all that I've observed, heard and thought to have understood from RED themselves, their primary target has always been, and still is, feature film and narrative tv production.

That's not to say those are their only targets, simply that's who they were targeting when they started the company, and still court with products like Monstro. For sure, they appreciate and bend over backwards to accommodate various types of shooters and shooting scenarios, as there are many and many overlap.


Let me digress a bit for a little history...

I think why people have historically misunderstood RED is their initial price point of $17.5K for the R1 body when Sony was up over $200K. RED's stance was high-end image capture doesn't have to cost so much, especially for 1080p, which was a major compromise on image quality in their eyes.

The added benefit to consumers was, while not cheap, that level of image capture was relatively affordable for the first time. The downside for RED was, due to such a low price by comparison, they've had a hard time shaking the "budget" digital cinema camera company distinction.

I remember Jim getting pissed at people going on about wanting RED to price Epic at $5K when it first came out, even though there was nothing else like it, so RED could corner the market and essentially wipe out all other competition. My business acumen may be ~zero but even I knew that was a dumb idea.



So, back to why RED isn't in competition with those other two...

RED sells cameras to all kinds of shooters for all kinds of productions but I think their main focus is still feature films, short films, independent films, narrative TV productions, fashion films, music videos, etc.

Basically, anywhere S35 film would've traditionally been used before digital became so affordable. Remember, their mission was to be a suitable replacement for S35 film. Obviously, things evolve, goals shift and definitions change. However, I look at their current lineup and I don't think they've deviated all that much.

The truth is, those higher end endeavors I think RED is still very much courting have not flocked to any of BMD's offerings. Ever. Sure, they've been used on those projects here and there to great effect. But, those are the exceptions. That group is simply not all that interested in BMD cameras, nor will they be in Z CAM. They're just not.

Don't believe me? Look at it from the other side. BMD cannot release a $15K body. Their usual market won't buy it because it's too expensive. They might as well buy a RED at that price. And the higher end target market won't buy/use it because it's BMD, which they've never seriously considered.

These companies have two different audiences, which translates to them not being in any direct competition.


One of the lessons RED had to learn was pricing your stuff too low lets in the riffraff. These are people who have more money than talent or who don't have either but got their hands on a RED and blanketed YouTube with horrible looking pictures while boasting about having a RED.

That happened a lot and really hurt RED's image quality rep in the early days, IMO.

Getting into the sub-$5K cinema camera market is a lose-lose proposition for RED. Those people are too hard to please and are never satisfied. They always want unrealistically more for substantially less.

The cameras are so cheap that many people are on to the next thing when it comes out, especially if the old camera has already paid for itself after just a few gigs.

That's why I think Komodo was not designed for the budget-minded shooter, despite the suspected $6K - $6.5K price point. I think Komodo is designed to fill a niche that the Z CAM could reluctantly, but not abundantly, fill if Komodo didn't exist.

Komodo isn't designed to appeal to the budget-minded shooter. It's designed to appeal to the types of productions that need what Z CAMs offer in terms of size, weight, battery life and resolution, but which would rarely resort to a Z CAM to get them. It will fill a hole that a Z CAM could on paper, but won't in reality, for RED's primary target market.

Komodo is priced according to what it offers to the high end compared to what their current, more powerful and capable cameras do. Jarred said on FB that people considering selling their current DSMC2 cam to get Komodo will likely be disappointed because they still outperform Komodo at most things.

Komodo will certainly be more than enough camera to handle the majority of bread-and-butter A-cam needs on most productions. That's probably what Soderbergh is proving right now.

I think this idea of it disrupting RED's current lineup to better compete with BMD and Z CAM is due to misunderstanding what the camera is and who it's for.

Budget-minded shooters, on the other hand, will seriously consider Z CAM over Komodo because they'll get perceived bang for their buck. However, I don't expect Z CAM to do any better in the states than Kinefinity does. Another reason why RED is not in competition with these cameras.


For years now, people have been saying RED was going to get squeezed out by this and that budget cam. But, many of the cameras that people said would contribute to that scenario have fallen off the radar. It hasn't happened for the same reason the Alexa isn't getting squeezed out by these budget cameras for the higher end stuff.

RED and Arri aren't in competition with them.

 

[Brian Boyer]

It's a bit of a tangent to what I said. A diode that has a greater well capacity will give us a stronger signal and higher dynamic range, given that the noise floor is constant.

The captured signal on sensor has no correlation to the distribution of stops around 18% gray. That comes later via OETF interpretation.

I mention 18% gray as a known reference from an end-use standpoint so people can understand the practical impact of what I'm proposing. It doesn't change the fact that some sensors can handle greater light intensity before clipping than others, correct?

 

[Fahnon Bennett]

IBC starts on the 13th. I'm assuming we'll know everything before the show is over.