ARRI ALEXA LF - 4k, FullFrame announced

[Björn Benckert]

My observations from the past few years are that more and more focus pullers are using monitors, especially with HD and wireless focus systems. I have to say though, I've been blown away by those who have mastered the art of focus pulling. Their keen sense of distance, the ability to anticipate variations, exceptional concentration. I drool with envy when I see the great ones pulling off difficult shots.

Bigger productions use the ultrasonic stuff.

But again, You only want so and so thin DOF. If you get a set of really fast lenses you usually dont want to shoot them wide open if you are not all out of light or strive to get the thinest fov possible. When moving to monstro I think most people will find them self stoping down more than they did on smaller sensors to reach the same sweetspot.

 

[Kenneth C Merrill]

When I used to pull on the Alexa Classic I felt very confident in my ability to pull by distance and lens markings alone. Nowadays shooting in 4K+ I find myself using the monitor more and more.

 

[David Peterson]

Bigger productions use the ultrasonic stuff.

But again, You only want so and so thin DOF. If you get a set of really fast lenses you usually dont want to shoot them wide open if you are not all out of light or strive to get the thinest fov possible. When moving to monstro I think most people will find them self stoping down more than they did on smaller sensors to reach the same sweetspot.

Yeah, people rave on about how bigger sensors are "better in lowlight" but totally forgetting (or maybe never even ever knew the basics) that when it comes to practical usage you trade off a lot of that better lowlight as you need to compromise with a slow T stop to get a usable level of DoF

 

[Mike P.]

It still drains the actors and the crew, looses time for other shots and burdens the data management and editing.
No one gets to delete anything on a serious shoot. Botched takes are marked and kept.

It's rare to do only one take on a narrative set (even if it's nailed the first time). I'd also argue that focus is most often blown because the actors don't hit their marks, not because the AC was out to lunch or struggling. And improvisational performances and "keeping the cam running" isn't entirely uncommon these days either. Data is cheap, and marking the best take(s) is done by the script supervisor on set (hell, you can tag takes via clip metadata nowadays)... it's not rocket science.

I also don't think it burdens crew nearly as much in practice since they're generally *not* doing different set-ups between takes -- you do a set-up, shoot it until its done, then do a new set-up. My point is, the reset-time and difficulty to 'shoot it again' has dropped considerably with digital; it's essentially real-time. Similarly everything is lighter and requires less physical effort than it ever has before (re: amount of lighting and weight). It's sounds like you're making a mountain out of a mole hill, and I'd argue that the aesthetic of big sensor images are well worth it for the majority of content (maybe not action sports or docs, though.)

And sure, no one deletes anything on set, but the drives can be re-used, so the cost isn't that big of a concern for even smaller productions (and the bigger the production, the less it's a concern).

No one pulls focus from 42 inch monitors. In many cases focus pullers don't even look at the monitor at all during a take and look at the scene and watch the distances. Autofocus is useless for those types of shoots.

Maybe no one you know does, but I've seen it being done more and more often (and it is *definitely* easier.) People who pull by eye still get mucho respect (because that's a legit art and there are no shortcuts to learning it) and hence are probably on all the biggest productions... For the rest of us mere mortals, I see no problem with someone using all the assists at their disposal.

Besides, who are we kidding; all of the up and coming ACs cut their teeth on 5Ds and T1.5 Rokinons with crappy marks -- they're more than capable of pulling decent focus on VV/FF35, especially with real glass

Autofocus is a topic for another thread, but if the advances seen in the last 3-5 years are any indication, it won't be long before it's a mainstay on the average production (especially when using drones or gimbals or steadicam.)

VV/FF35 are here to stay... Venice, ALF, Monstro are all going to trickle down into their entry-level and midrange counterparts. And as long as there are 4k+ s35 crop modes, there's no need to put the genie back in the bottle.

 

[Hrvoje Simic]

Bigger productions use the ultrasonic stuff.

Cine tape shows you the distance, and experienced focus puller already has the delicate feel for distance, instant reactions and precise muscle memory. As with the monitor, you have to see first the result to react, which in many cases is too late and focus lags behind the action. Watching the scene directly enables to anticipate the outcome and focus puller can flow with the actor in the speed and style director intended. Cine tape and monitors are there to check and correct if necessary and simplify things, they cannot compensate real time action feel because you don't have the same sense of depth and presence and connection with the scene, necessary to anticipate the motion through space.

I pulled focus on a bunch of projects, the longest on half of season of one TV series, so spoken with some mid-level experience let's say. Long lens fully open and camera and actors moving with time ticking and minimal options for re-takes was fun. I've also worked with a bunch of ninjas and that is fascinating to see up close.

Yeah, people rave on about how bigger sensors are "better in lowlight" but totally forgetting (or maybe never even ever knew the basics) that when it comes to practical usage you trade off a lot of that better lowlight as you need to compromise with a slow T stop to get a usable level of DoF

Which is why a speedbooster rocks.

 

[Björn Benckert]

It's rare to do only one take on a narrative set (even if it's nailed the first time). I'd also argue that focus is most often blown because the actors don't hit their marks, not be because the AC was out to lunch. And improvisational performances and "keeping the cam running" isn't entirely uncommon these days either. Data is cheap, and marking the best take(s) is done by the script supervisor on set (hell, you can tag takes via clip metadata nowadays)... it's not rocket science.

I also don't think it burdens crew nearly as much in practice since they're generally *not* doing different set-ups between takes -- you do a set-up, shoot it until its done, then do a new set-up. My point is, the reset-time and difficulty to 'shoot it again' has dropped considerably with digital; it's essentially real-time. Similarly everything is lighter and requires less physical effort than it ever has before (re: lighting and weight). It's sounds like you're making a mountain out of a mole hill, and I'd argue that the aesthetic of big sensor images are well worth it for the majority of content (maybe not action sports or docs.)

And sure, no one deletes anything on set, but the drives can be re-used, so the cost isn't that big of a concern for even smaller productions (and the bigger the production, the less it's a concern).



Maybe no one you know does, but I've seen it being done more and more often (and it is *definitely* easier.) People who pull by eye still get mucho respect (because that's a legit art and there are no shortcuts to learning it) and hence are probably on all the biggest productions... For the rest of us mere mortals, I see no problem with someone using all the assists at their disposal.

Besides, who are we kidding; all of the up and coming ACs cut their teeth on 5Ds and T1.5 Rokinons with crappy marks -- they're more than capable of pulling decent focus on VV/FF35, especially with real glass

Autofocus is a topic for another thread, but if the advances seen in the last 3-5 years are any indication, it won't be long before it's a mainstay on the average production (especially when using drones or gimbals or steadicam.)

VV/FF35 are here to stay... Venice, ALF, Monstro are all going to trickle down into their entry-level and midrange counterparts.

What you gain in sensor size you loose in speed on the image circle of the lens. So a small high preforming sensor and a fast lens looks very similar to a larger sensor and a lens that has enough sized image circle. If they both have focal lengths to match the FOV. So difficult to say what is the look of large format, if you loose that short DOF due to change of lens speed.


Monstro with it´s wider sensor get its full width into the circle of FF So it can utilise the speed of FF glass even though the sensor is actually got more width than FF. That will give it a "large format look" when using FF glas but if you move up to larger glass you start quickly to loose speed and the fast / large format look starts to drop of again.

 

[Hrvoje Simic]

It's rare to do only one take on a narrative set (even if it's nailed the first time). I'd also argue that focus is most often blown because the actors don't hit their marks, not because the AC was out to lunch or struggling.

I didn't suggest one take, just saying not to take lightly the notion of relying on more takes than usual. It is a much bigger deal than it may seem and changes many things.

Focus puller cannot depend on actor hitting the mark. He can mark the distance based on references and has to be focused 100% during a take to react. Actor has to process emotion, text, facial expression, posture, body movement, rhythm and tone and subtle nuances of it all in real time, and do that throughout the day and repeat if necessary, so with all that to handle it is not uncommon to miss the mark precisely.

Focus is missed either due to momentary lack of focus or stamina throughout the day or a very difficult shot to pull off. Imagine hours and hours throughout the day concentrated on that, watching and fine tuning when every mistake pops out.

And improvisational performances and "keeping the cam running" isn't entirely uncommon these days either. Data is cheap, and marking the best take(s) is done by the script supervisor on set (hell, you can tag takes via clip metadata nowadays)... it's not rocket science.

It is not rocket science but some get the thick end of the stick if that is overdone. I hear from a bunch of folks that film brough more focus, preparation and efficiency when media rolling was analogue to money rolling away. It is also pretty frustrating and tasking getting tons of data to handle because someone interpreted digital media as a compensation for preparation.

I also don't think it burdens crew nearly as much in practice since they're generally *not* doing different set-ups between takes -- you do a set-up, shoot it until its done, then do a new set-up. My point is, the reset-time and difficulty to 'shoot it again' has dropped considerably with digital; it's essentially real-time.

That is true and reset time alone is not a huge burden but neither the only thing. Stamina, focus, shooting schedule, data handling and post are. Set ninjas can handle the repetitions but you don't want a drained actor. Also, if the shot is physically difficult many folks get screwed. Less stamina you have things get progressively more heavy later. And it is a long shooting day. And the more time you waste on repetitions the less you have for other setups. Or > problems.

Similarly everything is lighter and requires less physical effort than it ever has before (re: amount of lighting and weight).

Ultra light cameras if handheld actually take more effort during shooting for a camera operator because there is no inertia to smooth out the motion so shooter has to use body muscles more to keep the shot steadier. Depends on the type of shoot/shot. But it drains you significantly nevertheless. It is not the carry weight as much as camera control and balance.

It's sounds like you're making a mountain out of a mole hill, and I'd argue that the aesthetic of big sensor images are well worth it for the majority of content (maybe not action sports or docs, though.)

I never argued against the larger format. I'm all for different options, from small to large formats.

Besides, who are we kidding; all of the up and coming ACs cut their teeth on 5Ds and T1.5 Rokinons with crappy marks -- they're more than capable of pulling decent focus on VV/FF35, especially with real glass

True, I've learned much on old stills glass with short focus turn, so many things were easier with a cine lens and a proper follow focus. But a stop less is still a challenge and things change when you have a whole set and crew around you, actors and scene to follow with energy and a tight schedule. And you don't get many chances to miss.

VV/FF35 are here to stay... Venice, ALF, Monstro are all going to trickle down into their entry-level and midrange counterparts. And as long as there are 4k+ s35 crop modes, there's no need to put the genie back in the bottle.

Agreed.

 

[Tom Gleeson]

Arri did a presentation here in Sydney last night with their new Alexa LF Camera and the reel footage looked superb. This camera will be an astounding tool. There are a few downsides as the data rate at Full frame is 2TB an hour! Compare that with 8K Red at 5:1 compression you could double that at two hours of recording with 2TB. At Redcode 8:1 which will still give you excellent results 2TB will give you well over 3 hours of recording. As Arri increase the resolution of their cameras they really need to rework their ARRIRAW codec. Even on bigger budget productions I have worked on with Alexa they have not used ARRIRAW. I did not get a chance to ask so I don't know if you can dual record Pro res proxies and ARRIRAW at the same time?

The LF like the SXT camera also does not have the internal switchable ND system although you can fit an internal ND. Anyone used to switching internal NDs using the Mini will have to forsake this on the LF. I did speak with one senior DP at the event who said this was a deal breaker for him. Of course Red can't point any fingers as they have not provided this feature. Sony Venice wins this round although the Kippertie internal ND system may address this when released?

If the idea is to go big in these cameras then the Monstro chip has an edge with its 41mm wide sensor compared to the Arri LF full gate 36.7mm and the LF's 16X9 mode at 31.7mm.

The camera also sucks over 120W so it will be the "Eater of Batteries." None of these issues are deal breakers although at $140,000AUS for a basic shooting kit it dampens my enthusiasm just a little but I expect to see a large proportion of major productions adopt this camera this year along with the pricey but extraordinary Signature Primes.

I would love to be shooting a big enough project that I had to choose from the PanaVision Monstro DXL, Arri LF, Sony Venice or Red Monstro. Such exquisite agony

 

[Bill Totolo]

Arri did a presentation here in Sydney last night with their new Alexa LF Camera and the reel footage looked superb. This camera will be an astounding tool. There are a few downsides as the data rate at Full frame is 2TB an hour!

But you can still shoot ProRes, no?

 

[Adrian Jebef]

Yes. 4K ProRes Log C encoded. Same ease of use that made the original Alexa ridiculously efficient for 80% of productions. The LF is a beast but the imagery is fantastic. Low-light shooting is insanely clean. Lack of internal NDs is a bummer but it won't be a deal-breaker. Biggest announcement besides the camera is the LPL mount available to the entire Alexa line of cams. Flange distance is the same as EOS. I expect this to enable some very creative lens choices in the near future. Very exciting.

 

[Tom Gleeson]

Adrian,

Can you shoot the Pro res simultaneously with the Arriraw and have ready made proxies?

 

[Björn Benckert]

Got an email with some used arri cams offerings. What was interesting was that.

The order they cost is the following:

Most expansive ALEXA mini. Price: 52,000 EUROS
Then Alexa XT. Price: $36,600 (w/ARRI check out)
And least costly Amira Camera with Premium License Price: $27,500



And it correlates quite well with what I see on shoots. Seams like people are not really into getting those big studio cameras any more.

I was on one big 14 day commercial shoot where we had 6 cameras and 5 of those where minis. Even though they where on large cranes dollys etc. The XT was sitting on a dolly and even the shoulder cameras where minis. And to be honest they had real difficulty rigging those in a nice way. The cameras where provided by one of the best rental houses in Europe but the shoulder cams looked kind of like an early adaption of red. Battery sleds behind camera, tons of cabling to FIZ and video transmitters etc, problems to get the EVF´s in good position, poorly balanced and so on. So I was a but surprised that they did not just get amiras or XT´s as they simply where not restrained by budget and could get what ever they wanted. But seams like the trend is that the AC's now want the minis, and for some strange reason they have absolute no problem at all with all the complications it comes with. Something they face palmed over when it was the same for early red adoptions.

but I guess the LF push down the price for the the XT´s.

 

[Phil Holland]

but I guess the LF push down the price for the the XT´s.

Agreed and it has certainly impacted the used market rather interestingly. For instance the typical Mini package used seems to go for $55K and I just saw one offloaded for around $25K during the week of the announcement. There was another for $35K that also moved, but didn't track the sale so much.

Classics and Pluses are going for sub $10K through $16K typically now. Used XTs are all over the place. Haven't seen too many Amiras for sale in the face of this announcement, but considering that sizeable variance on Minis there's going to be deals to be had for some who are interested in them.

 

[C. Burkhart]

Got an email with some used arri cams offerings. What was interesting was that.

The order they cost is the following:

Most expansive ALEXA mini. Price: 52,000 EUROS
Then Alexa XT. Price: $36,600 (w/ARRI check out)
And least costly Amira Camera with Premium License Price: $27,500



And it correlates quite well with what I see on shoots. Seams like people are not really into getting those big studio cameras any more.

I was on one big 14 day commercial shoot where we had 6 cameras and 5 of those where minis. Even though they where on large cranes dollys etc. The XT was sitting on a dolly and even the shoulder cameras where minis. And to be honest they had real difficulty rigging those in a nice way. The cameras where provided by one of the best rental houses in Europe but the shoulder cams looked kind of like an early adaption of red. Battery sleds behind camera, tons of cabling to FIZ and video transmitters etc, problems to get the EVF´s in good position, poorly balanced and so on. So I was a but surprised that they did not just get amiras or XT´s as they simply where not restrained by budget and could get what ever they wanted. But seams like the trend is that the AC's now want the minis, and for some strange reason they have absolute no problem at all with all the complications it comes with. Something they face palmed over when it was the same for early red adoptions.

but I guess the LF push down the price for the the XT´s.

XT's or SXT's ?

It's important to make that distinction.
SXT's are the latest generation, the XT's are older.
A typical SXT goes for 55k on the used market today.

 

[Björn Benckert]

XT's or SXT's ?

It's important to make that distinction.
SXT's are the latest generation, the XT's are older.
A typical SXT goes for 55k on the used market today.

Not sure. Says XT. Also said "super low price" so...


But it was not that long ago the XT´s came out and my point was more that people seam not to want the big studio cams, or they are willing to pay more for the mini. So as RED allready have a "mini" it seams a bit unlogic to build a big cam which is not possible to charge more for.

Seams like a better idea to make a really small brain and then let the open market sort out the rest. To me its a bit surprising that 3rd party manufacturer´s have not jumped at monstro and built more "eng style back packs" for it. Terradek and the others are still not having a line for weapon which I think is quite odd. If you are able to make a video transmitter, then to make a housing that fits on the back of weapon seams quite simple. But the prototype was shown I dont know how long ago and still nothing.

 

[Tom Gleeson]

Here in Australia last year the Mini outsold the SXT by ten to one. It will be interesting to see how if at all Mini sales are affected by the LF camera. I have been tempted in the past to buy into Alexa but the opaque road map makes it a tough call for investment. If I had bought in years ago it would have been a good call but in the twilight of the ALEV Sensor and with the lack of upgrade options spending the big bucks to be then left behind two years later is just too scary

 

[C. Burkhart]

Not sure. Says XT. Also said "super low price" so...


But it was not that long ago the XT´s came out and my point was more that people seam not to want the big studio cams, or they are willing to pay more for the mini. So as RED allready have a "mini" it seams a bit unlogic to build a big cam which is not possible to charge more for.

Seams like a better idea to make a really small brain and then let the open market sort out the rest. To me its a bit surprising that 3rd party manufacturer´s have not jumped at monstro and built more "eng style back packs" for it. Terradek and the others are still not having a line for weapon which I think is quite odd. If you are able to make a video transmitter, then to make a housing that fits on the back of weapon seams quite simple. But the prototype was shown I dont know how long ago and still nothing.

Maybe because the window is too small from one model to the next at RED for third party's to be willing to invest. They probably want to see what pans out as a popular cam from RED and which has some sustainability.

 

[Blair S. Paulsen]

DXL2 & Amira love child...

DXL2 & Amira love child...

(snip) Seems like a better idea to make a really small brain and then let the open market sort out the rest. To me its a bit surprising that 3rd party manufacturer´s have not jumped at monstro and built more "eng style back packs" for it. Terradek and the others are still not having a line for weapon which I think is quite odd. If you are able to make a video transmitter, then to make a housing that fits on the back of weapon seams quite simple. But the prototype was shown I dont know how long ago and still nothing.

+1,000

I've been suggesting it for years now. My analogy is separating the saucer section of the Enterprise on Star Trek.

The main brain has sensor/recording media bay/lens mount/power tap and multi-pin or pogo connector.

The rest of the "works" are housed in a shell that offers desirable features.

Make the "shell" in no more than two versions!!! One would be better. ACs/CamOps/DPs/Artists have to know what to expect in a rental kit and how to operate it - no surprises.

Take the most useful, reasonably cost efficient features of the Panavision DXL2 - crossbreed with Amira ergonomics. Market a $8-15K one piece module that has things like XLRs/manual audio gain/wireless video/FIZ transceiver/LCD ports/UHD spigot (SDI and HDMI)/etc.

DON'T BE AFRAID TO INCLUDE FEATURES THAT SOME MARKET SEGMENTS WILL NEVER USE - OTHERS WILL! (There are limits to this of course. That said, not including HDMI since it's not "pro" enough, or considering XLRs on camera a complete waste, etc are examples of "in the bubble" thinking antithetical to serving a broad enough market.)

There have been some excellent posts in this very thread about what a great owner/operator cameras REDs are, in no small part due to their versatility. Let's take the next logical step - build a backend with killer ergonomics and known support for a wide range of genres/shooting styles.

BTW - If anyone has anything like this in development, I suggest you have enough ready to ship once it's out of beta that you can flood the rental houses. If people see potential ubiquity, it motivates them to make the effort to orient themselves, knowing that knowledge will be useful on a regular basis.

Cheers - #19

 

[Nick Morrison]

Maybe because the window is too small from one model to the next at RED for third party's to be willing to invest. They probably want to see what pans out as a popular cam from RED and which has some sustainability.

All DSMC2 cameras are the same shape/size...so ergonomics certainly aren't a limiting factor, I don't think...

 

[Jacek Zakowicz]

I think the third party developers are a bit intimidated. There have been so many cases of the accs frying cameras- some probably think it's not worth the risk to their brand....