ALEXA MINI LF OUT :)

[Hrvoje Simic]

Arri really loves to make large mounts and with good reason, they are still not able to produce a sensor with enough resolution(5.5..6k bayer) in a range where mortals can buy lenses (s35 and FF).

Arri doesn't make sensors.

There is no "enough resolution".
There are aesthetics, criteria, choices to context and left-brain masturbation.

Arri's S35 image texture is closest to S35 film aesthetic. 6K is not.
Arri could have switched to larger resolution in the last 10 years had that been their priority. It was not.

 

[Björn Benckert]

Ps PVs new big mount the PV 70 is 40mm so ARRI could def have gone a little smaller but remember once your go to what ever the min depth is (and I have no idea what that is) ...that is it for the lenses designed for that depth for ever... if there is a need (due to some invention yet unforseen) to build a camera that needs a little more depth you can’t get those lenses on those bodies..

Best would ofcourse be to have a sensor that is mounted flush or even infront of the actual body and then the mount or what every you want to have in front of it builds out forward. Then any thing would apply, ND`s any type of mount etc. To have the sensor stuck far in any sort of cave infront of the camera is not forward thinking. The only natural way to progress is to get the sensor out of there. Kinefinity seams to be the most forward thinking in that regard their mount is pretty much flush with the sensor not a bad concept if you ask me. See the LF mayo.

http://www.kinefinity.com/mavo/?lang=en

From the newsshooter article

"HDE is free-of-charge for use with Codex Capture or Compact Drives, openly shared, and fast: ARRIRAW Open Gate 4.5K can be encoded at 24 fps on a modern MacBook Pro"

MacBook Pro specs: upto 2.9GHz 6‑core 8th‑generation Intel Core i9 processor, Turbo Boost up to 4.8GHz and Radeon Pro Vega 20 with 4GB of HBM2 memory
Or a $3.000 windows laptop.

Those are pretty moderate specs.

From the newsshooter article

"HDE is free-of-charge for use with Codex Capture or Compact Drives, openly shared, and fast: ARRIRAW Open Gate 4.5K can be encoded at 24 fps on a modern MacBook Pro"

MacBook Pro specs: upto 2.9GHz 6‑core 8th‑generation Intel Core i9 processor, Turbo Boost up to 4.8GHz and Radeon Pro Vega 20 with 4GB of HBM2 memory
Or a $3.000 windows laptop.

Those are pretty moderate specs.

Now the normal struggle when doing on set DIT work is normally, get material onto low cost drives, the kind clients wants. Normally that is single spinning drives, or small slow raids. Do the numbers shoot a couple of LF cameras and then you try to duplicate that material and stick it into something like 200-300MB/s drives. Then sure transcoding is also a thing and even if thats changing with GPUs. Before your computer pretty much locked up while transcoding. Sure you can copy in the background while resolve chews away and such. But I try to let one computer move the data and one process. And yes, no matter how you slice it having to render proxies on set takes focus of things I find more important, sure normally thats how its done, but i find it not to be the best environment for doing such things, maybe thats related to me being a post guy. But I find computers do better when they are rack mounted in a post facility then out in the mountains some place hooked up to a genie that some one really want to shut down and take home. Things like that.

I was talking about sensor! Arri had only 1 sensor since last 10 years.ONE
If you wanna count camera models please write how many models Red released in last 12 years.Less than Arri? Nope
Red One Mysterium, Red One Mysterium X,Epic-M,Epic-X, Scarlet-X,Epic Dragon,Red Epic Carbon...Im not even counting last 3-5 years.

I was not really doing a comparison with red. Just stating, Arri been trough a shit load of camera models yet the sensor tech has been the same.

 

[Marc Wielage]

On a lighter note, here's a satire shared to the Movie Set Memes group on Facebook by Tom Gault...

 

[Hrvoje Simic]

Now the normal struggle when doing on set DIT work is normally, get material onto low cost drives, the kind clients wants. Normally that is single spinning drives, or small slow raids. Do the numbers shoot a couple of LF cameras and then you try to duplicate that material and stick it into something like 200-300MB/s drives.

Cheap client drives are 2.5" spinners, mostly in 80-130MB/s range.

 

[Misha Engel]

Arri doesn't make sensors.

They also don't make: batteries, glass, lenses, carbonfiber, PEEK, FPGA's, magnesium, copper, screws, etc.. you get the point.

There is no "enough resolution".

Forgot to mention for 4k RGB delivery

Arri's S35 image texture is closest to S35 film aesthetic. 6K is not.

Texture, it's a CMOS bayer-sensor. What is s35 fim aesthetic(copy of a copy of a copy of a copy, etc...) soft images...

 

[Misha Engel]

Now the normal struggle when doing on set DIT work is normally, get material onto low cost drives, the kind clients wants. Normally that is single spinning drives, or small slow raids. Do the numbers shoot a couple of LF cameras and then you try to duplicate that material and stick it into something like 200-300MB/s drives. Then sure transcoding is also a thing and even if thats changing with GPUs. Before your computer pretty much locked up while transcoding. Sure you can copy in the background while resolve chews away and such. But I try to let one computer move the data and one process. And yes, no matter how you slice it having to render proxies on set takes focus of things I find more important, sure normally thats how its done, but i find it not to be the best environment for doing such things, maybe thats related to me being a post guy. But I find computers do better when they are rack mounted in a post facility then out in the mountains some place hooked up to a genie that some one really want to shut down and take home. Things like that.

We stopped using spinning/tape for mobile use, it's just to fragile.

 

[Hrvoje Simic]

They also don't make: batteries, glass, lenses, carbonfiber, PEEK, FPGA's, magnesium, copper, screws, etc.. you get the point.

I don't get the point.

Your statement was that Arri is "not able" to do something. As in "incapable" compared to others. While they dominate the high-end acquisition.

Forgot to mention for 4k RGB delivery

Left brain dominant fixation, attempting to glue value on content by counting dots of the canvas.
Forest for the trees.

Higher pixel count is a technical parameter advantage.
Aesthetics are something else. Perceived value of aesthetics is relative to feel and taste.

Texture, it's a CMOS bayer-sensor. What is s35 fim aesthetic(copy of a copy of a copy of a copy, etc...) soft images...

"Soft" is a valorization relative to reference and criteria.
If reference is S35 film, 2K was "too soft" and 3K digital acquisition is about there. 6K film scans are not introducing 6K scene detail but reducing the effect of digital quantization on organicity of film's texture. That's why the over-sampling there.
If reference is 6K or 8K digital acquisition, then yes - Alexa 35 seems "soft". Then again "- for what". For zooming 300% and staring at pixels - most likely. For turning a wide shot into CU also. For typical usage in narrative - no it isn't.

Directors, DOPs and actresses in CU don't seem to mind it not being "sharper".

We can appreciate and value Red and other camera makes without undervaluing the achievements of Arri team, with different priorities but excellence still.

 

[Patrick Tresch]

A lot of people say “fuck 4K I shoot 2K” and I tell them “Why 2k? I shoot 4K since 2012”
A 4K master on a 4K projector is something beautiful. Let’s say like a 35mm contact print VS a S16mm print. Now if I want to shoot s16mm I change the camera.

 

[Jeffery Anderson]

For better or worse, the Mini LF is going to be everywhere.

As will the Venice.

Especially on larger productions.

Monstro has a lot of competition now, and it's lack of internal ND isn't helping.

DSMC3 will probably address a lot of these issues, and I'm guessing RED has another large format sensor up it's sleeve (8K Gemini), that will send schockwaves across the industry (as it will compete with Alexa 65 for half the price, while also being much smaller and with a much easier workflow).

It also wouldn't surprise me if they DO announce an 8K Gemini, that Monstro drops in price - which will make it much more appealing/accessible.

Just throwing things at the wall here...

if you want a 4K finish, aren't you're still gonna need to shoot 5K for DeBayering to RGB?

I think lower compression ratios would be a good thing... and they are already teamed with Panavision for color science

Be interesting to see a Full Frame 36x24 5K sensor with very low compression ratios ...and if possible Highspeed recording (5K @ 240fps)

 

[Karim D. Ghantous]

if you want a 4K finish, aren't you're still gonna need to shoot 5K for DeBayering to RGB?

4.8K for UHD, so yeah, 5K. Raven was 'only' 4.5K, however, there is no up-sampling to 4K performed on that sensor, as opposed to a certain other sensor which I will not name. IMHO, a RAW 3.8K file is not such a bad thing, but it won't allow down-sampling. Scarlet-X, IIRC, is 4K at 16:9, and if so, it's technically a UHD camera, at the least.

Be interesting to see a Full Frame 36x24 5K sensor with very low compression ratios ...and if possible Highspeed recording (5K @ 240fps)

What's full frame though? It's always been a relative term, at least in cinema. ;-) However - someone above mentioned an 8K Gemini I think. That would, as was said, blow everyone's minds.

But what is even more important is the ability to shoot 240fps at full frame, as opposed to windowing. So really what we really want is Gemini 5K VV. Or not! Personally I prefer the smaller formats. YMMV.

 

[Jeffery Anderson]

What's full frame though?

36x24

But what is even more important is the ability to shoot 240fps at full frame, as opposed to windowing. So really what we really want is Gemini 5K VV. Or not! Personally I prefer the smaller formats. YMMV.

having larger photosites would get more detail that's what the Alexa is and why they stick with the ALEV III sensor. Not that the Monstero is a bad sensor by a long shot, but I think it could be interesting to see what type of dual ISO they could get with larger photosites and with highspeed you would need that extra light, so if the camera could have a native ISO of 1280/5000+

...it's be interesting

 

[Hrvoje Simic]

HA! I honestly never thought of it this way. I know the AE from my last feature was so stoked to have the proxies already made for them (indie budget, no DIT just dumping on set). And Producers were also happy for not paying for rendering out an entire movie from R3D files. Huge win for everyone involved...except poor DITs.

Offload
Backup
Sync audio & video
First light, Sometimes rough look
QC
Shooting day reports
Camera reports
Upload to server
Media organization
Cooperation with DoP for assuring optimal technical parameters of camera utilization and signal chain
Director assistance for media reviewing (when video assist is busy/moving/incapable to go through masters etc)
Pampering clients with media related special requests
Art dept/ wardrobe dept / MUA / FX /actor wants to check something from that day/camera/card/scene/shot/take
Pampering camera masters and backup so whole production can sleep well
Helping out camera dept.
...
...
...
...and there is transcoding. The part where one presses "render" and juggles time/focus/material/workflow tactic while cards and folks are coming and going, with minimal room for mistakes ( those which don't cause disaster).

 

[Björn Benckert]

Cheap client drives are 2.5" spinners, mostly in 80-130MB/s range.

Yes I was referring to the good scenario, something like lacies dual drive raids or such. But yes the normal route is into slow single spinning disk. which simply takes lot of time. Like twice the time the DIT needed for the regular mini or XT. With monstro data rates are the same as weapon dragon so as I see it there is a difference there. It was a slight bump going to DSMC2 from the epic bodies with mini mags, but not much. The LF comes with a X2 factor in that regard and thats a lot. Producer will raise eyebrows for media costs and DIT fees no doubt.

 

[Hrvoje Simic]

The LF comes with a X2 factor in that regard and thats a lot. Producer will raise eyebrows for media costs and DIT fees no doubt.

True but depends on the profile.

Dilletante producer choosing an uncompressed 4K LF, led by irrational desires and disassociated from reality will raise eyebrows at the media cost/transfer time/DIT fee because the meeting between the offset to reality and reality is often rude awakening. The one in touch with reality won't have those feelings because being aware of reality in the start he will factor in the requirements for LF if choosing LF.

The former is facing much more challenges than DIT fees in his project due to reality distortion field and messed up system of values predetermining his decision making.

 

[AndreasOberg]

And to poke a hopefully dead bear, all this crap about resolution not mattering over and over again from the Arri purest

Yes, I completely do NOT agree about resolution not mattering. I had a discussion with Arri a while ago and they asked "how often do you really crop in?". And I said "eh, very often". We have just been camera assistants on Blue Planet Live and some of the best shots were 8k crops to HD delivery so that we can zoom in almost 400%. And this is already using a 1500mm lens giving us an equivalent of 6000mm lens. So yes, for wildlife resolution is extremely useful.

Otherwise looks like a nice camera, and it is not so expensive.
Andreas

 

[Mike P.]

36x24

Only if she's 5'3"...

But yeah, I kinda agree that bigger photosites typically make everything better (DR, low-light, cleanliness)... BUT I love that 8k = 35mp, which is perfect for VV/FF35 sized photos. Monstro is ~41x22mm I think, so 3:2 crop would be 33x22... so close to proper FF35.

 

[C. Burkhart]

Arguing about resolution here again.
Well, guess it's Reduser.

 

[Uday Bhaskar Karra]

Arguing about resolution here again.
Well, guess it's Reduser.

Yes. This is Reduser.net. Are you sure you are posting in the right forum!

 

[Sergio Perez]

Obviously I have a Horse in the race, and most here are red owners/operators, or well into the ecosystem. But honestly, if not just to shoot "cinema"- and that is the 24fps/25fps kind, why would one choose the Mini LF over Monstro?

Arguably you can say Arri has more accurate, or more subjectively "filmic" look out of the box. Or that it "might" have half a stop more on the highlights (which I doubt). But seriously Monstro just completely outperforms in almost every aspect:

Resolution- Yes, resolution. It does matter for a variety of many commercial shoots. Film and motion picture production today goes way beyond the platform of simply a cinema platform. We are in a multi media world with a variety of distribution formats.

Mature Raw Workflow: Redcode Raw is a joy to work with. And with Nvidia GPU acceleration, the "working with raw and high resolution is a pain" is a thing of the past.

Cost of the system: the body, the accessories and Red Media in total are cheaper than Arri's solution.

Storage: Codex data is less compressed, and equals obviously more storage concerns. Nothing of big worry to the big productions, but to everyone else, this adds (a lot) up.

Lens Versatility: Red has interchangeable mounts for diverse type of lenses. Arri's system is much less versatile in this aspect. Not every production needs expensive PL Primes. Adapters are and always will be a hassle.

High Frame Rates at full resolution: The Monstro is basically the Alexa LF non Mini rival, not a Mini LF one. At 40 frames per second at 4K open gate, this is Arri's very own "Scarlet". Red Monstro is everything the Alexa LF is, but in a Mini form factor, which is the ideal form factor for versatility.

Stills for print: I've done a project with both Monstro and Helium where 8K Footage was used for both Magazine Prints and Broadcast Graphics. This is unthinkable in Arri. For me, this is exactly what a replacement of film should be: also able to pull out frames for print.

I will not argue the subjectivity of which one has the best overall picture for film . Arguably nowadays, this goes much more to the skill of f the cinematographer and the Color Grader (there should be an Oscar category for the grading). But if versatility and adaptability to the demands of modern day of Audio-visual Production is taken into account, I would definitely say Monstro is a step above every single other competitor out there. Its really a camera looking ahead and embracing the High End world of different delivery platforms we live in, and doing it with class leading performance. The Arri Mini LF is a brilliant cinema, or 24/25fps UHD 4K camera. One of the best at what it does. But that's it.

 

[rand thompson]

Pre-NAB First Look: ARRI ALEXA Mini LF





By AbelCine

The ARRI ALEXA Mini LF is the hottest ticket in town and we were very fortunate to not only sit down with ARRI's Marc Shipman-Mueller, Product Manager of the ALEXA LF and ALEXA Mini LF, but also to get our hands on the camera.

Read more: https://www.abelcine.com/articles/blo...