ARRI 4K Super 35 Camera Will be Launched Next Month

[Mike P.]

I must of missed the pamphlet; where is it? [Edit: LOOOL! Never mind; I just noticed a whole other page of this thread.]

As others have said, it could that the “texture mode” is merely to match ALEVIII cameras (cause the Minis, LF, 65 aren’t going anywhere any time soon).

 

[Patrick Tresch]

Texture mode… seems like Arri is going instagram…

 

[rand thompson]

Texture mode… seems like Arri is going instagram…

Patrick,

According again to the Brochure, all previous Arri Alexa camera came pre-programmed with a default predefined texture. This will give you more control over your own textured looks.

 

[Björn Benckert]

I can't see how downsampling is related to the DR. Bit depth - yeah, but not the DR, at least not in those values to be conserned of. I'm surprised how long it took Arri to hold out on this camera, it was known that the original plan was to release it in 2020.

17 stops of DR - wow. That's what camera progression should really look like.

If you count Usable stops then downsampling really makes a difference in the dark end of the DR. Downsample with interpolation from 8k to 4k and your image loose 3/4 of the noise which drastically adds usable latitude in the darks.

 

[Björn Benckert]

Patrick,

According again to the Brochure, all previous Arri Alexa camera came pre-programmed with a default predefined texture. This will give you more control over your own textured looks.

Hmm.
Quite likely "preprogrammed" is the same as clean as they could make it on the ALEV sensor. Now when things has been improved they are worrying that DP's will not like the more whistle clean look so they added a noise feature. As I see it the big selling point for Alexa cameras is that they looked very good straight out of the SDI port. So when connected to a on set screen DP's liked it, good amount of texture and a very nice rec709 lut to go with the log c signal. If this texture feature is put into something like a side car when filming raw, then sure, its a nice feature. I'f its actually burnt into the capture then I cant say I'm all positive.

 

[Khurum Khan]

it looks really good

 

[Matt W.]

Hmm.
Quite likely "preprogrammed" is the same as clean as they could make it on the ALEV sensor. Now when things has been improved they are worrying that DP's will not like the more whistle clean look so they added a noise feature. As I see it the big selling point for Alexa cameras is that they looked very good straight out of the SDI port. So when connected to a on set screen DP's liked it, good amount of texture and a very nice rec709 lut to go with the log c signal. If this texture feature is put into something like a side car when filming raw, then sure, its a nice feature. I'f its actually burnt into the capture then I cant say I'm all positive.

I agree.

The grain texture of the Alexa is quite different in each channel and finer in the green channel than in blue and red (vs film where the texture is in my experience the same per-channel but the amount is quite different). It has a much nicer texture than some other cameras but this seems like marketing talk.

 

[rand thompson]

Bjorn,

I think it will probably be like a SideCar file that you will be able to alter in the ARRi Camera Raw settings in Resolve to some extent. Maybe you will be able to extensively alter the texture with an update to the ARRIRAW Converter. But if this new TEXTURE is something other than allowing more noise back into a cleaner image and is instead the ability to control the amount, size...and other variable of the grain/noise, could that really be saved into a sidecar file?

 

[Adrian Jebef]

If you count Usable stops then downsampling really makes a difference in the dark end of the DR. Downsample with interpolation from 8k to 4k and your image loose 3/4 of the noise which drastically adds usable latitude in the darks.

Lol that's not really how it works. DR isn't about the noise. It's about the signal. Just because you reduce noise doesn't mean there's any usable signal down there to begin with.

All the downsampling rhetoric just keeps the conversation about resolution being the dominant factor in image quality. I know this is RU but perhaps we can give some weight to all the other parts of the process.

 

[James Darren]

Internal ND is very nice . I wonder if it will be electronic ?

No it won't be electronic ND's

 

[James Darren]

I'd expect Alexa Mini LF level pricing.... 70k for just the body

The full package is a bit more than that, but you're close

 

[Björn Benckert]

Lol that's not really how it works. DR isn't about the noise. It's about the signal. Just because you reduce noise doesn't mean there's any usable signal down there to begin with.

All the downsampling rhetoric just keeps the conversation about resolution being the dominant factor in image quality. I know this is RU but perhaps we can give some weight to all the other parts of the process.

Double Lol

I wrote "Usable DR" normally that is the part of the range between the sensor floor and ceiling that you consider usable which varies with your mastering resolution, screen size and a few other things.

The more downsampling you apply the more to the left you can expose and still have an acceptable level of noise in your final image.

The full DR or what ever you want to call it, the range between the absolute sensor floor and overexposure is of course not affected by resolution.

 

[Craig W. Bickerstaff]

I specified that Arri were surprised by the "Film community", not the camera's overall success, Film was still completely dominant in Hollywood at the time and digital had barely made a dent.

That's not true either, the first Alexa film was Anonymous in 2011 by which point digital had already gotten a pretty solid momentum in the industry following the Panavision Genesis, the Red camera and Arri had it's D20 and D21 cameras out as well. A lot of productions had adopted a hybrid approach where they would use digital for night scenes or visual effects. To say that digital had barely made a dent in 2011 is totally false. It was just the beginning of digital taking the majority of productions.

 

[Karim D. Ghantous]

Lol that's not really how it works. DR isn't about the noise. It's about the signal.

Yes, it is how it works. ARRI measures DR in the shadows with a S:N of 2:1. Sheet film has more DR than a 35mm frame, despite the fact that the DR per unit area is the same.

 

[Björn Benckert]

I agree.

The grain texture of the Alexa is quite different in each channel and finer in the green channel than in blue and red (vs film where the texture is in my experience the same per-channel but the amount is quite different). It has a much nicer texture than some other cameras but this seems like marketing talk.

Film works the same / has also different noise levels in the different channels. But for film you have different weighted film stocks. Some is made for tungsten, some is made for daylight etc. Daylight film stock has quite the same ratios of noise between blue, red and green as most Cmos sensors has. Digital cameras that compress the signal the same in all three channels of course get more artefacts in the more noisy channels as the more noise the more difficult to compress, uncompressed digital cameras are not suffering the same just as film is not compressed so the film noise when looked at in the telecine or projected without compression involved in the pipline normally come across as organic with perfect randomising etc.

 

[Matt W.]

Film works the same / has also different noise levels in the different channels. But for film you have different weighted film stocks. Some is made for tungsten, some is made for daylight etc. Daylight film stock has quite the same ratios of noise between blue, red and green as most Cmos sensors has. Digital cameras that compress the signal the same in all three channels of course get more artefacts in the more noisy channels as the more noise the more difficult to compress, uncompressed digital cameras are not suffering the same just as film is not compressed so the film noise when looked at in the telecine or projected without compression involved in the pipline normally come across as organic with perfect randomising etc.

That's not what I wrote. What I wrote is, intensity aside, the size of grain is larger in the blue and red channels than the green channel in the Alexa, consistent with a Bayer pattern array with more green photosites than red or blue.

In my experience, film grain varies in intensity, as you wrote, but each channel has an even size to the grain. I recently worked on a a job that was 4K S35 200T and another that was 4K S16 500T. 95% of my work is with Alexa and I own an Alexa and shoot raw and have to match grain for work too so I am familiar with it.

If you add a digital grain "overlay" the size is even per channel, like film, unless you specify otherwise. But the Alexa grain is not even in size per channel.

I'm just agreeing with you that this is marketing double-talk from Arri. I don't think the original Alexa has "texture management" so much as it's just a noisy camera in the shadows in a way that happens to look good. Dragon is also a noisy camera in the shadows in a way that looks good.

I am really curious about "texture management" and if it happens before debayer or not. On the other hand, I would never use it and I would just apply grain in post so it's sort of irrelevant to me personally.

 

[rand thompson]

Arri Alexa 35 Sensor Modes

bY OFFSHOOT RENTALS

ARTICLE
Arri Alexa 35 Sensor Modes - Offshoot Rentals

 

[Zack Birlew]

While I think this will be a nice S35 4.6K camera, I think the competition in the 8K to 12K segment is just more appealing and practical. Dynamic range as we have it is already really good and any more at the end of the day is just more ability to slightly bring down the sky and bright windows to balance out with the images as most of us are artificially lighting our scenes to craft our images as it is.

The low light aspect is also a little underwhelming as most cameras now can rock ISO 6400 and with the 8K+ resolutions allowing even cleaner images with minimal resolution loss after noise reduction than 4K or 6K are capable of and all while having the ultimate option to downscale to 4K or 6K if need be for even more visible noise reduction. Not to mention, several of these are also Full Frame sensors as well.

Unless ARRI has an 8K+ option coming out to accompany this new higher dynamic range camera, I can’t see this competing very long with the likes of the RED V-Raptor, Sony Venice 2, and new mid-range video and cinema cameras from Canon, Sony, Panasonic, and Blackmagic Design coming down the line within the next several months to a year away. I agree, there is a “Wow” factor when you see an ARRI Alexa image on the production monitor on set but when you look at 8K televisions being publicly available and looking ahead at what’s next in tech, that 4.6K resolution is going to be a big limitation.

 

[Karim D. Ghantous]

While I think this will be a nice S35 4.6K camera, I think the competition in the 8K to 12K segment is just more appealing and practical.

I slightly disagree. I have no time for the 2.8K sensors (I'd choose the long out-dated Epic MX over those), but the new one is actually something that has a compelling feature. It won't give you regret for using it. Okay so 4.6K isn't quite enough for true IMAX but apart from that, it's actually really good. The V-Raptor wins a lot of categories (and won't lose out too much on DR) but there is actually a good reason to choose the Alexa 35. I say this as a guy who loves Red and their cameras.

Lower budget projects will get more value out of older Red cameras like the Dragon 6K, so it's not like the Alexa 35 is on everyone's wish list. But it is a damned good camera.

I wonder why Red never used dual gain architecture? Or maybe they will eventually. Either way, they already win every metric asides from DR. But DR matters a lot.

P.S. The Alexa 35 makes the Alexa 65 redundant. Fact check: true.

 

[dan shay]

That's not true either, the first Alexa film was Anonymous in 2011 by which point digital had already gotten a pretty solid momentum in the industry following the Panavision Genesis, the Red camera and Arri had it's D20 and D21 cameras out as well. A lot of productions had adopted a hybrid approach where they would use digital for night scenes or visual effects. To say that digital had barely made a dent in 2011 is totally false. It was just the beginning of digital taking the majority of productions.

The Alexa was released in 2010. The design and planning for the camera took place in the years before that. Arri's D20 and D21 didn't have much success. You are reading history backwards as if people knew the rapid transition from film to digital in the movie industry would take place just a few years later. They didn't know. The Alexa was initially targeted for TV and commercial production, with cameras only having internal Prores recording for quick turnaround and were limited to 16:9 2880x1620 for a 1.5 down sample to HD. It was a couple years until Arri released cameras specifically targeted for the film industry(The first 4:3 Camera was the Arri Studio[2012] and internal RAW and Open Gate came with the second generation XT models)

And if you don' believe me here's Arri's perspective from Frans Kraus in 2015 "However, even we were surprised by the speed of the switch from film to digital. We thought that TV and lower-budget feature production would go digital first, and that high-end movies would take another ten years before switching away from celluloid."