John Brawley

5 hours ago, jonpais said:

So you're saying it was impossible to have a longer life battery like the Panasonic? (asking for clarification)

Just like you’re saying that it is possible ? 

How do you know ?

The pocket 4k, like every Blackmagic camera uses a custom thermoelectric cooling system.  The only other camera platform that uses this approach is made by Arri. 

In fact this means it draws more power in hotter weather and less power in colder weather. It’s got a larger screen and has custom made hardware boards made internally by BMD themselves. They make the boards. The place the components. They don’t outsource. 

I beleive they tend to use FPGAs because they’re faster to prototype and totally re-program, but they use more power than ASICs which is what the big guys tend to use. ASICS use less power but are limited to their original design / specs and take longer to develop. 

They draw more power than consumer cameras that are more purpose built and have longer development cycles, therefore have more time to optimise their power consumption . 

Feel free to keep droning on about it and reminding everyone how bad it is but it’s not going to change the fact that they use more power than other consumer cameras.

JB

1 hour ago, Myownfriend said:

According to a post by John Brawley on a the BMD forums, BMD cameras do have an automated pixel remapping function.

It’s an automatic process. Just like you can’t do it on Alexa. 

JB

9 hours ago, jonpais said:

I don’t know anything about doubling sales, but a longer life battery would most certainly be welcome - manufacturers of everything from cellphones to laptops to you name it realize the importance of battery life.

Nowhere is it written in stone that you should spend upwards of $300 for a solution to a problem that need not exist in the first place. 

How do you know that ?

Active thermoelectric cooling for example.

JB

4 hours ago, jonpais said:

Do you work for Blackmagic or something? ?

He’s one of the main and lead members of the camera development team.  He is especially responsible for Gen 4, now available in the beta of Resolve. 

jb

35 minutes ago, DBounce said:

What pictures? Ok, now that that's settled... back to obsessing over the specs?

Soon...

JB

Do we really and truely honestly all believe that the exact same sensor is being used in these cameras ? The X7, E2, GHx, P4k...

It’s a rhetorical question.  Which means I know the answer I’m just asking it here publicly...

And why does it even matter.  Even if they ARE the exact same sensor. Every camera manufacturer will get different results and we learn NOTHING from these kinds of theoretical sensor discussions. Look at the previously mentioned CIon / BMD 4K / Aperture images.  Are they the same ?  Look at digital Bolex, who used the KAF / KODAK senors.  Are they the same images as the Ikonoscope / Leica / Olympus E series cameras that also used the same sensor family ? 

Just because they share the same sensor or sensor linage or it comes form the same factory, it’s almost meaningless because the company making the images form that sensor have to do so much customising, calibrating and processing of the image, it becomes irrelevant other than the specs we already know...

Let’s just look at the pictures instead of obsessing over the specs....

JB 

27 minutes ago, Robert Collins said:

Of course there are... 

To quote Sony...

"The third category of sensors is completely generic – the sensors are created for use internally and to sell externally, to anybody. China or Taiwan or wherever.”

You can literally pick up a catalog of sensors from Sony, Samsung or Omnivision for that matter.

What Nikon does - getting a custom build with their design exclusive to them - is the 'exception' rather than the 'rule'

'Custom builds' are expensive to design and even more expensive to build because you both need a separate manufacturing line and you then have to ramp up to an acceptable yield.

If you consider that the cost of a generic M43 sensor is US$40-50 a piece and you expect to sell 100,000 cameras, your spend on sensors is roughly US$4.5m.

Most components in cameras are 'generic' or 'off the shelf'. Shutters by Nidal Copec, EVFs from Epson etc...

You're implying that you can buy a generic sensor, put it in a camera and you get a picture.

It's not that simple though.

Let's look at a camera that uses an off the shelf sensor in the sense you're describing.  The Apertus project.

Started conceptually in 2012.... https://www.apertus.org/

Still not really a production ready useable camera and they've been working on that for YEARS, even with the magic lantern help and the open source community.

In the time since they've started, the same sensor that has been developed into use for the AJA Cion, several BMD 4K cameras and other machine vision cameras.

If it really was off the shelf, then Apertus would have a camera that doesn't still have issues Apertus have.  They would have delivered by now with the significant effort going into it.

Why haven't they ?  Why haven't they just been able to plug the same off the shelf sensor used in other cameras into a body, program up some boards and get it working ?

Because there's no such thing as an off the shelf sensor.  Even the off the shelf sensors still need a huge amount of development to make them into a functional product that can ship.

It's certainly not the exception rather the rule at all.

JB

1 minute ago, Myownfriend said:

I think may have misinterpreted something because that article doesn't dispute anything about off-the-shelf sensors. That whole article was about Nikon makes there own custom sensors and has nothing to do with purchasing sensors made by other companies.

I see users of this an other forums often quoting the phrase "off the shelf sensor".  In a similar line of thinking, many have theorised that because Sony "make" a lot of Nikon's sensors, that they're essentially a Sony sensor in a Nikon camera body. Sony look, Sony tech and Sony DNA.  Much like the conversation in this very long thread about the source of the sensor in the namesake of the thread and it's DNA....

I suggest this article as a way of drawing parallels.  Sony make certainly fab the sensor but it's anything but a Sony.

JB

3 hours ago, IronFilm said:

Well it never got sold, for basically the same kind of reasons BMD had, the benefits of having global/rolling as an option wasn't worth it for the trade offs. 

 

I don’t think it ever even got made, let alone sold.  I think they only announced it.

It’s worth reading the very interesting article Andrew posted about how Nikon sensors get made..I would hope it may change our view of the idiotic phrase “off the shelf sensor”.... because no sensor is off the shelf....

JB

12 minutes ago, IronFilm said:

I found it interesting that Geoff Boyle managed to get not one but two Kinefinity cameras for his shoot out comparison!

 

 

And apparently the results are so bad he didn’t want to post them because he thought something was wrong. 

JB

There’s no problem making contact. I spoke to them at cinegear. 

But its its moot until there’s actual cameras shipping as advertised and there’s depth of inventory.

JB

1 hour ago, Savannah Miller said:

To be fair if you emailed kinefinity and told them you were interested in shooting a major network TV on their cameras they would give you 8 bodies and all the support in the world.

You’d think so.....

Ive never been able to get any kinefininty camera to even test.

JB

I don’t think you guys get it. 

From a TV drama point of view there is little appetite for RAW. No one wants it aside from a DP maybe.

It doesn’t matter if you think it’s better or the workflow can be worked around. 

DP’s don’t really get to make this choice alone you realise ?

And most productions want multiple cameras.

Lets imagine Mavo.  If I wanted to shoot Mavo on the show I’m on now I’d have to convince a rental company to get 8 bodies and kit them out.  Tell me a rental company that will do that and then get a couple of spares as backup.

in fact show me any rental company anywhere I can get 8 kinefinity anything cameras from... 

I’d have to convince a director of the creative reasons, and SHOW it was better. I’d have to convince a studio that hates varying anything from the templates of how they do things.  

It’s not going to happen.  You don’t understand how it works and the theorising over specs without actually referencing the pragmatic reality is just playground taunting.

These cameras are in fringe for many reasons. The pictures are only part of that story. (And yeah I’m talking about the C200 as much as Mavo)

JB

2 hours ago, mercer said:

Raw in the C200 is advertised at 15 stops. CLog3 is 13-14. So same or better.  

Doesn’t shoot ProRes. 

Inconsistently can’t shoot 12 bit and has to shoot compressed RAW to do it.

Doesnt as easily look like Alexa as the UMP does.

Doesnt do TC. Major major problem in a multi camera TV drama series. That makes it instantly a non starter by the way for any serious regular use.

Unergonomic to use the way I like to operate.

The only thing it can do better for the way i shoot is AF.  And I don’t really use AF.

JB

1 hour ago, Anaconda_ said:

Sorry for posting the same video again, but..... @John Brawley what are your thoughts on this hypothesis? (using ISOs lower than the native for 'cleaner' images)

 

I think Zero Dark Thirty has some of the best night work. It was shot at iso 200 on alexa. 

JB

31 minutes ago, Mokara said:

So, given a choice between a C200 or similar and this thing at the same price, you would choose the BM camera?

I can use pretty much any camera I choose to on a show like this.

Which is why comments like yours don’t make a lot of sense.  There’s always exceptions to totalitarian comments like that ?

In this case, I short a short test in pre of the kinds of shots we’d be doing on the show. We blind tested DXL, Alexa Mini, Helium and UMP. 

The very very experienced and visually accomplished director picked the UMP as the one he liked most for look based on the graded images from the four cameras we tested. He’s also worked with me before with the operating style you see above which he also likes very much.  

Based on look (from a blind test) and based on operating ballistics (as above) production chose UMP.  

Like I said on a pilot we could have had pretty much any camera we wanted.

What would a C200 give me in drama shooting that the UMP doesn’t have ? 

JB

1 hour ago, Mokara said:

Lol....the reason they don't "over charge" is because really low price is the only reason people would buy these systems. 

I guess you’re right...

11 hours ago, kye said:

Thanks @John Brawley - fascinating stuff!

I was trying to work out the disk space required but there appears to be some confusion online around Prores 444 vs Prores 4444 which seem to be referring to the same thing?
Assuming 330Mbps for Prores 444 the 12 hours of dailies would be around 1.8Tb/day
I count there being 6 other master-quality copies of the footage made, which equates to over 12Tb per day, or 100Tb per episode of total storage for all involved.....  ouch!
(please someone double-check my logic / maths here...!)

When you shoot with the BMMCC + BM Video Assist 5" what codec do you shoot then?  BM doesn't list Prores 444 / 4444 on the tech specs.  Perhaps 422?  Otherwise I suspect transcoding would be required.

 

ProRes 4444 is the post version of the format (has the alpha channel) and in camera it's called 444.

On the Micro cameras when I use them I only use the VA as a small lightweight monitor.  I record internally to 1920 ProRes HQ.

Resolution and dynamic range / bit depth aren't related at all really.  You could maybe make an argument that extra resolution helps with surprising noise, but it's only those that want to make a petty technical point that would say so because it's a small advantage.  I doubt the difference between 720 and 1080 would make any difference at all.

If the costs come down, I can see Networks going to 4K because they DO want to future-proof themselves, they just don't want to PAY what it costs to process and store all those 4k files.

To me as a cinematographer I care a LOT LESS about resolution and a LOT MORE about bit depth.  I'd rather have 1920 12 bit files than 8k 8 bit files.

JB

14 hours ago, kye said:

 

Who has worked on serial productions with RAW / HQ cameras, and what can you tell us before we fill our HDDs with RAW 4K files?

As mentioned in the other 4K Pocket 2 thread...

In episodic TV, where I spend a lot of my time, we shoot three cameras almost all the time.  The show I'm on now shoots 8+1, meaning 8 days of main unit and a day of second unit (while making unit starts the next episode).

We tend to shoot about 2 hours per camera per day.  

So that's 6 hours by 8 days of footage. On second unit they shoot with at least two more cameras, if not three.  So that means on some days the data wrangler has 12 hours of dailies to process from set from two units.

The footage is backed up on set, then shuttled to the post house offsite.

They backup twice again and then transcode the footage for editorial.  We shoot in Atlanta but editorial are in LA cutting on AVID.  So they grade (time) the dailies, for editorial and render DNX files and send them over a very secure connection to LA.  All done Overnight ?

Editorial then sound sync the dailies, first thing, compare the shots logged vs what they've received and sign off that they have everything that we say we shot (down to the shot).  After the emails I see from editorial are notes querying the wrong slate number just so that they can verify this.

As well as doing all that they produce the pulled or select dailies and generate 264's of just those circled / printed takes and they get assembled into a H264 file that gets uploaded to DAX for about 200 studio execs and production people to watch / refer to.  There are also some secured rotating ipads that get the same 264 files updated to them for those of us that hate trying to scrub 264 files over the internet.   I get those myself, and the director of the episode gets one as does the writer. 

Right now the show I'm shooting is on air 9 weeks after we started shooting.  That's a clock winding down, because it airs every 7 and we take at least 9 days to shoot each episode ?

You can see how it escalates.  There's no safety net for time at all. An episodic TV series will generate a LOT of footage that can quickly overwhelm a production and there's very little room for a slow down at any point.

The studio I'm shooting for now is Fox.  They still only air in 720 so they only master in 1920.  

Despite the rhetoric about 4K, the reality is that MOST TV DRAMA production isn't shot 4K.

We shoot 1920 ProRes and I always make ProRes 444 the minimum because it's 12 bit.  That's all I really need from a grading point of view and RAW workflows offer very few genuine advantages.  Try designing a test to show that it does and I'll show you a TV network that shoots 4K.  (None of them do)

They're not even interested in mastering 4K for a future proofing scenario.

4K also quickly kills workflow because everything is that much harder with all the above numbers.

I think ProRes RAW is an interesting option, but honestly, over 12 bit ProRes, I don't see much need to shoot ProRes RAW.

I have some friends on Stranger Things at the moment.  They do 12 to 14 days per episode.  We have 8.  That's nearly double the amount of time to shoot the same screen time.  They tend to shoot single camera too, because that's the style of the show.  But that's a large part of why they need so much more time.  I guess you can argue it's success makes the investment worthwhile, but days shooting is the most expensive part of TV production.  Because of the cast generally and their availability.  A show that shoots for longer tends to find it much harder to hang on to good actors.

JB

On 7/23/2018 at 9:44 AM, kye said:

Thanks @John Brawley that makes sense, and the data logistics would be significant.  However, I thought someone had mentioned that Prores was preferred even in situations where the data rates were similar?  I probably should have included it in my original question

Maybe I'm remembering incorrectly, or perhaps that statement might have been in a different context?

 

I'd love to explore it more, but I got some flack for discussing workflow issues like this as being non-core 4K issues in this thread so let's talk ab out this somewhere else some other time.  

JB

4 minutes ago, kye said:

I'm curious to understand a bit more about how RAW differs to Prores in post.

To preface this, I'm a Resolve user and I don't know much about PP or FCPX, so maybe there are larger differences there?

In terms of Resolve, when you shoot DNG sequences:

• the sequence appears in the media browser and other windows as one item, the same as a clip, and behaves like one throughout
• the RAW panel is available (which isn't for other file formats obviously) but I think the defaults are just to defer to the camera metadata, so you don't have to change anything here if you don't need / want to (this is the part I'm less familiar with so maybe there are things you need to do here?)
• everything else behaves the same way a clip would

Assuming the above is correct and I'm not missing something, where is the extra difficulty?

Is it in extra processing power required to debayer the files perhaps?  Or does PP / FCPX not handle image sequences as clips?

People online talk about image sequences from time lapses as an extra bother because you have 'lots of files' or 'you have an extra step to combine them' but in Resolve that's completely automated.

The issue is logistics. 

The show I’m on now shoots in Atlanta but editorial is in LA.

Dailies needs to be available next morning synced and ready to edit.

We shoot three cameras on almost every scene.

We shoot about two hours per camera per day of footage.

Ar least once per episode we shoot 2 units so you can double that again. 

That’s a lot of footage that has to be graded, transcoded to DNX (Avid) and stored safely in three locations (archives).

Every day.  For weeks and weeks.

RAW data rates would swamp a production.  

The network I currently shoot for AIRS in 720 and masters at 1080.  They don’t need RAW.  There’s ZERO advantage, especially if you’re shooting a nice 12bit file like ProRes 444.

JB

52 minutes ago, webrunner5 said:

Gulp!    https://www.bhphotovideo.com/c/product/1317914-REG/sandisk_sdcfsp_256g_a46d_extremepro_cfast_256gb_525r.html?ap=y&gclid=EAIaIQobChMIgZf_rtSw3AIV3rjACh1wyAH3EAQYAiABEgJbivD_BwE&smp=y

Been getting good results with Angelbird. 

https://www.bhphotovideo.com/c/product/1380442-REG

JB

On 6/12/2018 at 10:20 PM, User said:

I've recently been roped in as a set photographer. I was using the ancient mirror-slappin' 5DM2 (with a decent collection of Canon lenses)... so no shooting during sound recording.

Until now I've paid relatively little attention to the mirrorless market for photography... the other set photographer was packing a Fuji XT something.

Is the Fuji XT (something) a smart investment on the road ahead. Lenses? Good video would be a bonus. Any thoughts?

Mostly I see Fuji on set. Some Sony and some Olympus (EM1 Mk2)

Olympus can do 60FPS RAW in silent and have ProCap, basically a stills buffer if you’re trying to get an action event. (Capture frames BEFORE you hit the shutter release)

Sony is better in low light and for resolution (post crops).  Olympus is better for native lens reach, better IBIS (lower shutter speeds in low light), smaller, cheaper, lighter weight and better weatherproofing. I think also better rolling shutter because the sensor is smaller (faster integration time)

jb

What I recall is 99% of people that complained about media issues didn’t use cards from BMD’s list.  I know that they also struggled to keep up with the SSD list because it takes a while to certify media and the manufacturers were constantly changing the performance.  I vaguely recall one of the SSD drives that WAS recommended wouldn’t work when they went to the next generartion of the same branded media.  That’s not on BMD.

It is why they changed to CFAST though even though everyone seems to think SSD are the perfect solution and always work.  Go look on BMD’s forums at the number of complaints over SSD’s with the original BMCC.

I personally worry about USB-C as a port for reliable use.  I would hope someon comes out with a cage or mount that locks that plug down.

I’ll be sticking to CFAST.

JB

29 minutes ago, MurtlandPhoto said:

I certainly was around for that. Those were early days. BM was still young & inexperienced and SD cards were not up to the same read/write standards that they are today. The landscape has massively shifted since then. 

There were data rate issues with BMD recommended SD cards ?

jb

It costs millions.

From what I've seen (first hand experience that is) at several companies it's a two year development cycle at minimum to make a camera. 

That's dozens of engineers.  Many iterations of sensor fab customising alone. Not to mention the in-house development costs. 

And there's no such thing as off the shelf sensors.  Just ask Apertus. They still haven't made a feasible camera after how many years ?

JB