New Camera and Mount Upcoming ?

[CaptainHook]

In the release video Grant mentioned that the BRAW debayer was changed to account for the OLPF. Where does that fit in for those us who have a Rawlite filter in an original 12K?

No affect, it will get processed as normal. Because 12K cameras with the OLPF installed have a different camera type name the SDK can identify it has our OLPF installed and has been calibrated with it.

For those that have a Rawlite OLPF already, it’s very very similar in performance in both OLPF and IR. I have both and I won’t be changing the Rawlite options I already have installed.

Carry on.

JB

Thank you for sharing such clear and useful information.

In the release video Grant mentioned that the BRAW debayer was changed to account for the OLPF. Where does that fit in for those us who have a Rawlite filter in an original 12K?

It’s a bit complicated.

Basically the camera knows if it has a factory installed OLPF. It’s in the metadata and tells Reaolve to process it slightly differently. Let’s call it a more optimised result.

If you have a Rawlite OLPF, then as anyone who has one knows, there’s a slight change to WB and tint. In a BMD factory OLPF this slight change is flashed into the individual camera when they calibrate each sensor when they make the camera.

So Rawlight OLPF = better IR, reduced moire.
BMD OLPF = better IR, reduced moire, better inter camera matching and slightly sharper result for same moire reduction.

I’m not sure BMD have worked out if they would be OK with shipping an OLPF out to users like what Rawlite does or is it better to send your camera in so it can be re-calibrated and better optimised with the BMD filter fitted.

JB

And if you need one now, there's nothing wrong with getting the RAWlite.

We have it in our UMPs and it's great. The correction for grading is minimal, if at all. If you are not shooting with a bunch of these, some with and some without that OLPF, you may hardly notice.

Well if nothing else the NAB offering has made me convinced whatever's coming next from BM is a replacement for the Pocket line. Here's my thinking:

1.) The 12K is clearly going to be the flagship cine camera for some time
2.) Their broadcast offerings are generally comprised of hardware trickling down from other cameras
3.) Unless they do a 4K MFT G2, there's no refreshes left of the current Pocket line

Think the days of BM as an experimental pioneer are over. They're well established in cameras and serve three very clear demographics; content creators/indie cinema, broadcast, and pro cinema. They could cycle through products for these three markets from now until eternity. Reliable, but no surprises or weird leftfield cameras like the Micro.

I still have questions like, will BM ever go full frame? Will they ever make a box camera? Are they done with the 4.6K sensor? And we might get answers to those someday, but probably not as a full frame box format camera with two 4.6K sensors stitched together that gets randomly announced at NAB.

To those who have experience shooting with the 12K RGBWWW sensor and the Dual ISO sensor in the Pocket cameras, do you think there is any chance we’ll see the RGBWWW tech in an 8K Pocket with mFT mount? Physically it would fit, but would the BMD marketing department feel it’s a risk worth taking in your opinion?

I often shoot the BMPCC4K at ISO 1250 when ISO 400 is too dark and am satisfied with the result except when the lesser dynamic range is evident? If the proposed BMPCC8K was released, would you agree that it would provide a significant improvement for 4K/2K deliverables in moderately low light without the use of a dual ISO option?

If we’ve seen the last of the Pocket camera form factor, then an 8K URSA Micro or Cube form factor with RGBWWW sensor with mFT or BMD mount seems possible, so the gamble would be ceasing use of the Sony Dual ISO approach. I am trying to avoid purchasing another URSA Mini body.

I am trying to avoid purchasing another URSA Mini body.

I made the decision before this NAB that I would not purchase another one. So I'll be looking at other cameras now. And as much as I hate it that probably means the Red Komodo.

Disappointed in BMD for not being able to read the room. At least my room.

The news from NAB 2023:

NABShow Daily April 17

https://issuu.com/futurepublishing/docs ... d=31790704

NABShow Daily April 16

https://issuu.com/futurepublishing/docs ... d=31790704

It's interesting that with the U12K OLPF Netflix has released already approval and best practices, yet the camera is not yet even shown available for preorder on B&H or Adorama. Blackmagic made Netflix approval a top priority, obviously with bigger ambitions than just listening to "the room." This was the right decision for the longer term, as is the Ursa body. Some of the updates mentioned for updated USB and flash memory cards made sense and could still be ahead but I've used the Komodo too. If people feel that particular footprint is so important that they would exit the brand, then by all means they should.

I am trying to avoid purchasing another URSA Mini body.

I made the decision before this NAB that I would not purchase another one. So I'll be looking at other cameras now. And as much as I hate it that probably means the Red Komodo.

Disappointed in BMD for not being able to read the room. At least my room.

depending on what you shoot, save yourself a headache and avoid the Komodo. This coming from a former Komodo owner.

It's really about a camera that has timecode and genlock in a smaller form factor. I am done trying to cart around the Ursa body for every corporate / narrative job I have. I honestly have no use for the Komodo however in the under $10,000 price point it's really one of the only cameras out there that has it. Sony blew it with the FX6 in not adding the timecode genlock that the FX9 has.

The OLPF is actually a significant announcement. You might think not but here is why.

I personally think too this is significant news and the fact it has been now been approved by Netflix, in spite of what we all say, is very important for BMD. This camera always deserved to succeed IMV, it was an incredible achievement, especially at it's price point. Well done BMD! I won't resurrect our disagreements about OLPFs, John, I promise.

The most important aspect of the OLPF announcement is that it means future cameras & sensors will have a Blackmagic made OLPF. This means that the next use of the 12K sensor in a new body will also have an OLPF.

I'm positive behind the scenes Blackmagic is working on the next big design. Grant said as much in his previous quote about investing in future products. We just didn't see the new stuff at NAB this year.

Right on time, to distract those looking to dissect a made up failure of BMD to deliver your dream camera at NAB, I wanted to briefly return to a note I was making earlier about camera balance, where I said the ideal should be to have the camera sit as low as possible and that you could take your hands off and not have it move.

Well…this is at NAB, one of the SOC instructors for an SOC / NAB event on operating. (SOC is the Society of Camera operators. Like the ASC for operators)

https://www.instagram.com/reel/CrJSFMiA ... MyMTA2M2Y=

In other news…I still have the same friends at BMD and they still have are doing the same things they’ve always done. They’re still making and designing cameras and workflows and thinking about what they can bring to the table.

Their last big project was dropped at the very height of a global pandemic if you all recall? Literally launched a camera that was so game changing most still don’t understand what’s different about it and focus on how they don’t need the headline resolution. It’s funny now to see older projects that were derided for failings that users thought were huge problems then becoming more recognised later and becoming cult cameras.

Some of the voices here made the same doom and gloom comments in these forums in the comedown of not getting what they want and seeings signs where there are just choices that don’t align. Everyone’s an expert on how to be a successful camera designer. I feel like it’s an annual ritual to point to the Homer. https://uxdesign.cc/the-homer-simpson-c ... 7da30ba348

BMD have always done things in their own way. I’ve often really disagreed with many of the choices made in design and spec but I’m also a pretty niche user and have to accept my “wants” are not what everyone else wants.

The pandemic in some ways was good for business but mostly was not. Supply chains are a massive headache that cause DELAY to both shipping product, but also new, emerging and R&D.

They’re still very much in the camera business. I’m very excited for their future.

JB

Color me mildly disappointed but not surprised that Blackmagic didn't have something more substantial to announce, but I figure with using their own sensor on the Ursa Mini 12k things might slow down a bit on releases (and I recognize that BM doesn't necessarily release on tradeshow schedules.)

I ordered the Ursa Mini Pro the first day it was available, same thing with the G2, and I've been very happy with them. Great image and, for me, a mostly perfect package. I'm in the camp that likes the shape and absolutely don't want a box camera. It's such a nice package for handheld and shoulder work and I haven't used anything else I liked better. Tried the Sony FXs, the Komodo, Canon C-Series, have used various other Reds, the Varicam LT, etc. and nothing within 2x-3x the price seems like a good alternative. For the doc and doc-style commissioned work I tend to do it's been fantastic.

Really my only complaints are I wish it had: even a little better high ISO performance, like useable 6400 and 3200 without FPN; better audio controls on the body with hard stops and limiters; a bit better durability / repairability / availability of replacement parts.

My G2 has been having some hiccups so I'm looking at either replacing it with another G2 or a 12k (or another mfg entirely, if only there was something that seemed better), and I'm warming up to the idea of trying a 12k, but I wish it felt like more of an upgrade than a sidegrade/slightly different version of the same thing.

Find a 12K to rent for a day or two and you'll see that it is an upgrade, not a sidegrade.
Those colours are a dream come true. It can play with the big boys, and I don't mean Red or even Venice.
And FPN? If you want to find any (if at all), you'd need to starve it of light and then turn up exposure to insane values. Doing that, I can find flaws in any CMOS sensor, including Arri's (well, with the exception of not having tested the S-35 yet).

Eh, I might give it a shot. The color is already quite nice on the G1 and G2 and not an area that I felt like was deficient in the first place. I'm used to working within the G2's limitations for lower-light situations; I just wish it did better and it's something I'd happily pay for. If I shot narrative mainly it would be less of an issue I think, but for the more doc types of work I do you just wind up in needing more exposure sometimes and the UMP falls apart dramatically in a way that isn't very salvageable when you push it. Obviously any camera has issues if you push the sensor too far but the UMP really wants a lot of light. On the flipside, I think it handles contrasty noonday sun incredibly well.

In the meantime I've been using an S1H as b-cam/backup/low light option and it works great for all that but color matching to the Ursa Mini is a little rough.

The news from NAB 2023:

NABShow Daily April 17

https://issuu.com/futurepublishing/docs ... d=31790704

NABShow Daily April 16

https://issuu.com/futurepublishing/docs ... d=31790704

NABShow Daily April 18

https://issuu.com/futurepublishing/docs ... d=31790704

Allright, got a 12k rental on the way to give it a fair shot at least.

Planning to try a few compressions but curious what any of you all like for balances of compression/resolution. Like say, for UHD delivery if one would be better off doing 12k Q5 or 8k Q1. On my G2 I shoot 4.6k braw at 8:1 and 5:1, and then UHD 422 or 444 for clients that want prores. I feel like there was... some reason I used constant bitrate instead of constant quality on the G2 but I made that decision back when I got the G1 so it's lost in the sands of time.

I'm thinking if I did have one of these I'd mostly be shooting 8k for 4k delivery, but 12k seems to have some not too bad data rates if you use higher compression. Ideally I'm looking for equal/better than the G2 quality if possible while still being able to do some HFR without jumping in and changing codec/resolution and sticking with one CFast at a time. I tend to use 48 quite a bit, some 60-96, and then anything above that I don't mind a little more jumping around in settings.

This depends, most of all, on your media. IMHO, if your media is fast enough, constant quality is the better option. But you need to be sure that even the most demanding scene will not cause dropouts. Now, what’s a demanding scene? It’s all about contrast and amount of fine detail. Contrary to popular belief, it’s not about motion, since BRAW is an I-frame codec. OK, to some degree motion blur will reduce detail, so if there’s any effect, it’ll rather reduce demands, other than for GOP codecs.
Now, contrast and detail are not only about scene, but lenses too. The 12K sensor is outresolving most but the very best lenses. So, if you don’t use very good glass at optimal aperture, you’ll be hard pressed to see a difference between 12K and 8K when downsampled to 4. I’d say, carefully testing it under different conditions with your own lenses is essential.
But also check John Brawley’s commentaries in this subforum, he is the authority on these cameras and spreading most valuable information.

It's interesting that with the U12K OLPF Netflix has released already approval and best practices, yet the camera is not yet even shown available for preorder on B&H or Adorama. Blackmagic made Netflix approval a top priority, obviously with bigger ambitions than just listening to "the room." This was the right decision for the longer term, as is the Ursa body. Some of the updates mentioned for updated USB and flash memory cards made sense and could still be ahead but I've used the Komodo too. If people feel that particular footprint is so important that they would exit the brand, then by all means they should.

If you’re just looking for a Netflix approve BM camera, then the URSA Mini Pro 4.6 G2 (UMP46G2) has been approved and hard tested by them already, while the U12K OLPF is still being tested. Also, the UMP46G2 is still in the market and a very good camera even after many other models have come and gone. It is still a very viable option.

This a fairly thorough video on the 12k compression and resolution at different settings -

I tend to shoot 4k at Q5, and some 8k at Q5.

This depends, most of all, on your media. IMHO, if your media is fast enough, constant quality is the better option. But you need to be sure that even the most demanding scene will not cause dropouts. Now, what’s a demanding scene? It’s all about contrast and amount of fine detail. Contrary to popular belief, it’s not about motion, since BRAW is an I-frame codec. OK, to some degree motion blur will reduce detail, so if there’s any effect, it’ll rather reduce demands, other than for GOP codecs.
Now, contrast and detail are not only about scene, but lenses too. The 12K sensor is outresolving most but the very best lenses. So, if you don’t use very good glass at optimal aperture, you’ll be hard pressed to see a difference between 12K and 8K when downsampled to 4. I’d say, carefully testing it under different conditions with your own lenses is essential.
But also check John Brawley’s commentaries in this subforum, he is the authority on these cameras and spreading most valuable information.

Constant quality is variable bitrate encoding, and with wavelets, which I am virtually certain BRAW is, that although I-frame, it still considers motion as well as detail, just not motion in the sense of interframe GOPs, but intraframe motion within individual frames that looks for repeating patterns in the content, representing them mathematically. The wavelet transform breaks down the image into smaller pieces called sub-bands that can be compressed more efficiently than the entire image frame.

Anyway, for 12K and 8K I find Q5 is usually okay whereas Netflix says Q3, or 8:1 and 5:1 for CBR. Where I noticed a slight deficiency with Q5 was rare, but nevertheless could present as a loss of a minimal amount of fine detail within a flat, neutral colored wall or surface of low contrast approaching the clipping point from bright or sunlight, the gradation from shade to sun on a flat neutral surface, but not loss of fine detail within busy areas such as print patterns or foliage, wood, skin or clothing. In other words, not anywhere that it really matters.

Constant quality is variable bitrate encoding, and with wavelets, which I am virtually certain BRAW is, that although I-frame, it still considers motion as well as detail, just not motion in the sense of interframe GOPs, but intraframe motion within individual frames that looks for repeating patterns in the content, representing them mathematically. The wavelet transform breaks down the image into smaller pieces called sub-bands that can be compressed more efficiently than the entire image frame.

No, BRAW isn't wavelet, it's 8x8 DCT. You can read this even in the BRAW patent files.

BRAW is DCT based.

The compression is based on focus. Not movement. So how fine is the detail and is it in focus. That’s what drives the compression.

I only shoot 8K if I really need the faster rolling shutter.

Otherwise shoot 12k. Use all the photosites. Especially for UHD delivery. That’s when you’ll notice the difference.

Don’t think of it as 12k. Think of it as all the sensor. Yes, you can shoot 8k and 4k but you’re really shooting a lot less of the sensor.

I usually shoot Q3. Q5 can be very good too depending on what you’re shooting.

Try both Q3 and Q5. It’s not the same as on a G2 either.

Q3 usually runs about 1TB per hour. Same as Arriraw.

Also, if you want a single card HFR you’ll have to shoot Q5 drop to 8K. 2 cards makes it a lot more reliable at 12k.

JB

Thanks for the info everyone. Not sure I'll be able to stomach the data needs so G2 may remain my primary, but I'll give it a bit of testing and see how I feel about the 12k Q5 vs Q3. For about a quarter the stuff I do 1Tb/hr Q3 would be doable, but for the rest I'd need something skinnier. Be fun to mess around with it regardless.

Thanks for the info everyone. Not sure I'll be able to stomach the data needs so G2 may remain my primary, but I'll give it a bit of testing and see how I feel about the 12k Q5 vs Q3. For about a quarter the stuff I do 1Tb/hr Q3 would be doable, but for the rest I'd need something skinnier. Be fun to mess around with it regardless.

On 12KQ5 I've been able to get 70+ mins from a 256GB card on a simple talking head.

JB

Wow, that's wild. Excited to get the camera in my hands and do some tests!

Constant quality varies wildly with the scene. Shoot for a few seconds and the camera will give you a pretty decent estimation of the time left for that specific one.

Constant quality varies wildly with the scene. Shoot for a few seconds and the camera will give you a pretty decent estimation of the time left for that specific one.

Time remaining is based on the previous 5 seconds of content.

When you power cycle or format a card it will tell you a theoretical maximum.

I’ve found with Q5 and Q3 they can sometimes be even more efficient than the published tables say they are.

JB

No, BRAW isn't wavelet, it's 8x8 DCT. You can read this even in the BRAW patent files.

Okay, thanks for the correction and sorry for the misinformation.

I'm sure no one is waiting with baited breath for my impressions after renting the 12k, but here they are nonetheless:

• Perhaps this is a given since it's the same body, but felt very similar to filming with the G2, from how it exposed/wanted a lot of light to working with it in post
• Q5 12k would be doable for me — for many, possibly even most, scenes I tested it was estimating 30+ minutes per card and regularly over 40, which is half what I'd get out of the G2 but still manageable. Detailed scenes pushed that up a lot of course but for me those are ones where I can handle the higher data rates
• Q5 12k did seem better than Q3 8k
• Sensor definitely gets noisy at higher ISOs, but I think it holds onto color better than the 4.6k in low light
• Agree with folks that the color is a step up from the G2's already great color
• While tests and BM rate the 12k as having lower dynamic range than the 4.6k, in practice I did not find it to be that noticeable
• Was surprised to find that with Q5 I could do some HFR single card work in 12k depending on the scene. Was not reliable though — even 8k Q5 if it was, say, trees or grass it'd start dropping frames single card
• Pleasantly surprised at how smooth editing was — in both Resolve and Premiere with Autokroma
• Really feel like the camera should have CFexpress instead of CFast. I'm sure they have good reason for keeping Cfast, but it'd really open up what you could do without spanning cards or using an SSD
• It'd be amazing if it could record super sampled UHD ProRes

Overall I was quite impressed with the sensor and camera — really look forward to where Blackmagic takes it in the future and kind of amazing that what they've done here. For my own use, the main issues that are probably going to keep me on the G2 or another camera for the time being are those last two above — I just need to have a solid ProRes or other more standard option in-camera, and I feel like the media type is a major compromise.

… I feel like the media type is a major compromise.

One could be forgiven thinking that a 12K camera was the ideal opportunity to introduce the pain of supporting CFexpress media cards. Sure that’s a bit of a burden for users that owned several CFast2 cards, but 12K/8K recording would have been the honey to help the bitter pill go down.

You have to remember that the camera dropped in mid 2020. 3 years ago basically.

It was late but it had been in the works for a few years. You can’t really just change something like that on the spot. CFexpress wasn’t really as supported as it is today. And it cost more at that point in time from memeory. It’s easy to forget that it wasn’t so certain a format at that point in time.

I’ve been told 12k superscaled in-camera ProRes would be computationally impossible. Or not possible when they designed the camera presumably for the price point too. Don’t forget this sensor is very unique. There’s really nothing else like it, and a lot of stuff happens in-camera like the scaling. I believe the processor overhead would be too much.

I also find the DR to be “similar” to a G2.

And see what I mean once you start using the files? Everyone thinks their computer will choke and the media will be onerous. But it just isn’t like that.

And yes the colour. It’s hard to describe but to me it’s more nuanced and subtle. Like there’s more individual hues visible.

JB

You have to remember that the camera dropped in mid 2020. 3 years ago basically...

JB

Very true.

Yeah that all makes sense — feels like a bit of bad luck with timing really. And I don't think that any of these things are showstoppers, just some tough tradeoffs for my own needs. Amazing what they've done with this sensor and camera.

The video title is: Soon Blackmagic will announce their first L-mount cine camera!

What's your thoughts?

Eh, maybe. Whever this comes up most folks tend to forget that the majority of BM cameras are based on Sony sensors. Our friends at Sony are already clocking hundreds of dollars for every Studio and Pocket sold. What’s a couple bucks more to license a mount?

Good Luck

That video describes the L-mount camera to be full-frame. Now we have the Super35 covered by the UMP12K camera. Will this mean BMD is going with a 16K version of their RGBWWW sensor that presumably would have a special place in the VFX world? Or will they rethink that 12K S35 sensor and issue a full-frame 8K RGBWWW sensor with 4.4 micron photosites? Would be a higher dynamic range I should think. Back to the $9,995 price?

This thread is a rinse and repeat thread. The repeat process happens every time a major cinema event comes and goes. NAB was going to see the next camera, then Cinegear was going to see the next camera. and on and on it goes.

It's literally been years since a new cinema ( not broadcast ) camera was announced. And the 12k was a new sensor put into the asme old body style. Honestly I am not convinced that Blackmagic is coming out with a new body style camera at all this year.

Eh, maybe. Whever this comes up most folks tend to forget that the majority of BM cameras are based on Sony sensors. Our friends at Sony are already clocking hundreds of dollars for every Studio and Pocket sold. What’s a couple bucks more to license a mount?

Sony will never ever license E mount.

The Sony that make sensors is a different company entirely to the one that makes cameras. Just because they share a name, it really means nothing.

L mount is at least a maybe possibly yes could be done. Leica own it, it's not like MFT where there's a consortium and so far the L mount companies are all kind of Leica aligned. But it's the most likely option because it would make sense for Leica to have a partner like BMD involved.

JB

Quite possibly you’re right, but there is IBC 2023 mid-September, so we could have an announcement for delivery in 2023.

Quite possibly you’re right, but there is IBC 2023 mid-September, so we could have an announcement for delivery in 2023.

But as discussed many times previously they don't really use trade shows for announcements, only when they're ready to ship.....

JB

And the "full frame" comment seems weird. Well, even more than it being a term to begin with. As every camera design is "full frame" if you use the full designed opening, whether film or digital. (And I came into video from pro stills.)

Most of us choose a camera depending on the lenses we have or wish to use. Yea, I've still got a few "full frame" Nikor lenses, but why would I want to use them for VIDEO capture?

Especially weird as an L mount even knocks out the ol' Canon DSLR crowd.

List me as choosing to identify as "officially skeptical".

Sent from my SM-S908U using Tapatalk

And the "full frame" comment seems weird. Well, even more than it being a term to begin with. As every camera design is "full frame" if you use the full designed opening, whether film or digital. (And I came into video from pro stills.)

Most of us choose a camera depending on the lenses we have or wish to use. Yea, I've still got a few "full frame" Nikor lenses, but why would I want to use them for VIDEO capture?

Especially weird as an L mount even knocks out the ol' Canon DSLR crowd.

List me as choosing to identify as "officially skeptical".

Sent from my SM-S908U using Tapatalk

L mount was designed with a 135 format sensor size in mind.

Like EF, which hasn't stopped BMD from making non-135 format sensor sizes with EF mount.

I think L-mount at least has room to grow for those that are interested in AF and other modern lens niceties as well as having nice stills cameras available that natively use the same lenses.

And it ticks the box better for adapted lenses too, much better than EF does...

JB

Well noted. Though I still identify as "O-S".



Sent from my SM-S908U using Tapatalk

I get that it’s silly to want a camera thats specifically tailored to me, but i would just like a tinier ursa with the 4k sensor and mft mount. That would make my life easier, just being able to use the lenses i already own, a slightly nicer form factor that isn’t a cube, have internal nds, multiple video outs, viewfinder support, just using 1 big battery and not having to rig anything. That would be just be my dream.

panasonic kinda managed to do it in 2010, camcorder like shape, mft, internal nds? Would be cool now too.

Eh, maybe. Whever this comes up most folks tend to forget that the majority of BM cameras are based on Sony sensors. Our friends at Sony are already clocking hundreds of dollars for every Studio and Pocket sold. What’s a couple bucks more to license a mount?

Good Luck

Sony also makes money on almost every L mount camera. Panasonic cameras and Leicas use Sony sensors.

Sony also makes money on almost every L mount camera. Panasonic cameras and Leicas use Sony sensors.

While Sigma has their own sensors for photography, even the Sigma fp is using one from Sony.

The video title is: Soon Blackmagic will announce their first L-mount cine camera!

What's your thoughts?

Before we get carried away, the source for this is literally just a YouTube video saying 'trust me bro', right?

Before we get carried away, the source for this is literally just a YouTube video saying 'trust me bro', right?

What? Are you implying here that you can't just trust any random YT dude talking about some hear say?
You've destroyed my naive faith now...