What is the practical point of having a Helium 8K??

[Björn Benckert]

Oh thats interesting to hear. That goes way deeper than things I know about the Helium. Which lenses would you recommend to play with and understand this better if I may ask?

Lenses optimized for the S35 format instead of the FF format. Master primes is optimized for S35, so is Leica Summilux and sumicrons. Superspeeds and most good cineglas made the last 20 year.

Only just recently there is lens lines coming out that are optimized for FF and above. And it´s easy to see that they very much loose as much light as their image circle gets bigger.

In comparison: Suprime Primes T1.5 where the master primes with their smaller image circle goes T1.2
Same for leica Summilux goes to T1.4 where the more large format lens line from leica, the Talias "only" goes to T2.7.

Those T numbers are not by coincident larger for the lines with more coverage. It´s simply physics. If you want to cover a larger image circle you loose speed.

A wide range / optimo kind of zoom, something that covers the same Field of view as their Academy 27-290mm but with an image circle for monstro. Will always be slower.


But dense photo site arrays are more difficult to make, pixel quality normally suffers. So its not easy making a 8k S35 sensor that hold as high quality per pixel as a less dense array. But in that sense helium is preforming quite stellar.


People tend to compare the differnt sensor using the same lens, normally a lens that cover the larges sensor in the comparison. But to be fair you need to take into account what lenses are available for the given sensor. And for Helium there is a huge amount of very good and fast glas. Glas that does not cover Dragon or Monstro. Like the above mentioned lenses.

Put a Summilux -C 18mm on helium shot it wide open at T1.4

To get the same FOV / DOF on monstro you would need to use a 25mm lens covering monstro and... with is interesting, open it to 2.0...

Well thats something Leica have not yet produced, as mentioned above you trade image circle for speed. The Talia 25mm only does T2.7 as it´s image circle is so big. And you can not use the Summlux-C 25mm on monstro and open it to T2.0 to get the same big focus fall of as the helium on the 18mm as the summilux-c 25mm does not cover monstro for natural reasons, as it´s image circle optimized for S35...

Here is a lens calculation image of the above example. Monstro to the left helium to the right. Both cameras pretty much delivers the exact same image in terms of field of view, DOF etc. But again such leica lens is yet to be made.

[/url]MONSTRO left HELUM RIGHT by Björn Benckert, on Flickr[/IMG]

So all in all it´s a lot of talk about large sensors and the great focus falloff that they provide... if you use the best lenses available for the given format you use, you will notice thats not really the case. Actually its very much the contrary.

 

[Jon Dishler]

I had to make the hard choice of upgrading Epic W cameras to either keep the Helium sensor or go Gemini, both the same 5k upgrade cost. Or one of each. In the end I chose to keep the Helium, and my "new" Weapon class Helium cameras have just returned. I am very happy with this decision, like the 8k option at 60-75 fps, always liked the image, love how they downscale to 4k or 2k and don't mind lighting a bit, they are still very sensitive sensors. Most of the stuff I have seen posted for Gemini is very low light with virtually no lighting added, and it looks like ... not so great. I started in photography and with my Nikon cameras I can shoot very high clean ISO, without a flash or lighting, but never do as the colors just do not look the same. I can shoot the Helium at 2k 4444 at 60 fps and with a PhilmLUT burned into the image get very nice images right out of the camera ready to edit. The little bit of noise that was in the shadows in R3D is gone once downsampled if I shoot 8k and then deliver at 4k or 2k alternatively. One thing that amazes me which is rarely discussed, is how perfectly the images match when shot at 24 fps and 75 fps with just a little adjustment to the aperture settings but nothing else. I just love these cameras, so pleased with the upgrade!!! Happy 4th everybody.

 

[Aaron Lochert]

So all in all it´s a lot of talk about large sensors and the great focus falloff that they provide... if you use the best lenses available for the given format you use, you will notice thats not really the case. Actually its very much the contrary.

What about the Supreme Primes against Master Primes? Surely T1.5 on Monstro is shallower than T1.3 on Helium for the same FoV, no? You'd need like a T1 of an equivalent focal length on Helium for that kind of DoF.

 

[paulcurtis]

I've still not seen that many side by sides. I did seriously consider the Gemini route but decided to stick with Helium for most of the reasons about.

I thought that Gemini came from left field quite a bit. It was a custom sensor design that Red decided to offer publicly. There was little fan fare (compared to normal) and so i think the interest caught them by surprise a bit. I think people have latched onto 5K out of concern from dealing with 8K. But what i feel a lot of folk have missed is that no one is delivering 8K, delivery is 4K and the whole point of 8 to 4 was such a good downsize. But you could sharpen the 8K then downsize and you get a nicer image. Or denoise at 8K then downsize. Reframe then downsize and so on. You can do the same with Gemini down to 2K though which to be honest most deliveries are.

With the DSMC2 update and weapon class speeds then the reason is even less so.

But of course unconscious bias means that now i decided to keep Helium i'm clearly swayed into that camp!

cheers
Paul

 

[Björn Benckert]

What about the Supreme Primes against Master Primes? Surely T1.5 on Monstro is shallower than T1.3 on Helium for the same FoV, no? You'd need like a T1 of an equivalent focal length on Helium for that kind of DoF.

Nope, It´s not as "bad" as for the Talias vs Summilix-C but if you look at this image see that a Supreme 65mm got pretty much the same FOV and dof on monstro as a MP 50mm on Helium. Difference is, Monstro is hit with less light per pixel.

I assume the Talias have way bigger image circle than Suprimes and that causing the the difference. Also MP´s are T1.3 trough the set and likely most of the are actually faster. Arri just made sure they where tack sharp wide open and that all lenses open as much. So the "error" in this math is alot due to the actuall coverage of the lenses / how far out on the image circle they dare to use of the given lens and also the max speed the manufacturer think holds up. In the screen dump below Supreme/Monstro to the left is set to T1.6 as the calculator does not have 1. but still the falloff is close enough the see the match.

But yes, MP on helium is a nice equalent to Supremes on Monstro. And the stop difference is there. HeliumMP would get more light per pixel than Monstro Surpime, while creating the same image.

MonstroSupreme65mm vs HeliumMP50mm by Björn Benckert, on Flickr

I been discusing this at length with DP´s when testing the new Alexa LF´s. in comparison to the regular alexas all you gain is more pixels at the cost of light. And the funny thing is, most of those guys using the LF says resolution does not matter they use it for the look... well they dont really like to find out that they get the exact same look using mini and MP´s minus the reslution.

I can only guess Michael bay and fincher and those guys had the above in mind when building their custom cameras around helium. I think Jared would have gave them a monstro each if the begged for it.

But I dont really live as I preach I love monstro and would not trade it for two heliums. Like the color rendition better and also I like to be able to use the rim of the lenses. Think it´s cool to have such large sensor getting some vinjetting at 8k or cropping down to 6k to get a cleaner image. And also better vertical coverage for ana´s and shooting overcranked. And I rarley need prestene 8k images, only when doing stills or such then I can make sure we use lenses that hold up.

 

[Phil Holland]

Actually to that point if you are using that specific tool the way you are you should be doing a comparison between Helium and Monstro at f/1.2 and f/1.4 to make up for the T1.3 and T1.5 variation in this case. Which also alludes to just some of the differences:

And yes, to make up for the DOF difference between something like T1.5 on Monstro 8K VV versus Helium 8K S35 you would be looking at glass around T1 if you are hoping for an optical match it would likely be faster than T1.

The more relevant cut/comparison for me is the 5K versus 8K on Monstro, S35 versus VV essentially and the optical match manifests in about a 1.62-ish stop difference depending on the lens. Which would require glass that's clearly faster than T1.

Not really a big thing unless you are truly optically matching different formats, but just some of the differences between VV and S35. This becomes something more relevant when stopping down and general lens speed in relationship to image quality.

Note in this case I have a precise 68.5mm on the VV lens to match the 50mm S35 FOV to make up for that difference.

 

[Björn Benckert]

Thats probably correct Phil but the Talias does not do 1.4 nor the Suprimes.

And my guess is the Master primes and summilux-c, as sharp as they are corner to corner even wide open, have not utilized their image circle to the full extend and could be build to open up more, T1 for some focal lengths for sure, they make them to match the slowest in the lineup. Where as something that covers as much as Talias doing T1 is pretty much impossible. So even if the numbers above is not showing a totaly striaght formula the physics behind it all still apply.

Sure at some focal lengths that is easy to make large and fast, like a 50mm you can put a noctilux on the helium and match it very well with a T1.2 68mm on monstro. And the differnece in light / per photosite still applyes, Helium get more exposure in such comparison. So what Im trying to say is if focus falloff is wanted it´s not for certain that the race is won with a bigger sensor. Especially if you also want good light transmision.


So to me the point of having a helium 8k is that it´s possible to use fast S35 and even academy sized glass and still get 8k resolution. Which is quite remarkable.


The difference can be tested in a similar maner. Take a lens line covering Monstro, get a booster and use the same lenses on Helium. If your booster is a 1.33 booster you tecnically get the same image in helium, plus more light per pixel. Sure there is some light loss going trough the booster but less than the gain. Simply If the lens is super fast covering monstro. It will be super fast + and give the same image on helium, minus the artifacts the additional glas brings.

 

[Michael Tiemann]

Actually to that point if you are using that specific tool the way you are you should be doing a comparison between Helium and Monstro at f/1.2 and f/1.4 to make up for the T1.3 and T1.5 variation in this case. Which also alludes to just some of the differences:

[...]

And yes, to make up for the DOF difference between something like T1.5 on Monstro 8K VV versus Helium 8K S35 you would be looking at glass around T1 if you are hoping for an optical match it would likely be faster than T1.

Yes! DOF is related to f-stop, which can be inferred (somewhat) from T-stop. Light hitting the sensor is related to T-stop, which can be inferred (somewhat) from f-stop. I see that these two terms are being frequently and erroneously interchanged. Let's try to mind our P's and Q's!

The more relevant cut/comparison for me is the 5K versus 8K on Monstro, S35 versus VV essentially and the optical match manifests in about a 1.62-ish stop difference depending on the lens. Which would require glass that's clearly faster than T1.

Speaking of P's and Q's...here's another opportunity for confusion. Is a "pint" 16 ounces or 20 ounces? Is Helium S35 8K or not? Is Monstro 5K S35 or not? Obviously we are all way, way, way in the weeds here. But since the title of this thread concerns Helium at 8K (a 20 ounce pint if you will), then we should compare against Monstro at 6K, not 5K.

 

[Björn Benckert]

Yes! DOF is related to f-stop, which can be inferred (somewhat) from T-stop. Light hitting the sensor is related to T-stop, which can be inferred (somewhat) from f-stop. I see that these two terms are being frequently and erroneously interchanged. Let's try to mind our P's and Q's!



Speaking of P's and Q's...here's another opportunity for confusion. Is a "pint" 16 ounces or 20 ounces? Is Helium S35 8K or not? Is Monstro 5K S35 or not? Obviously we are all way, way, way in the weeds here. But since the title of this thread concerns Helium at 8K (a 20 ounce pint if you will), then we should compare against Monstro at 6K, not 5K.

You are right I write Tstop and mean Fstop.

Yes a bit confusing when relating sensor size to film formats. "Crop factor" is a real confuser I think. "this sensor has a negative cropfactor of -0.3...." etc is often heard. Instead of just stating the mesurments for the given sensor. It gets even more difficult as the aspect ratios of the sensors varies.




But to keep it simple a smaller sensor with the same amout of pixels can reach the same FOV and DOF as a larger sensor + get more light transmitted per pixel. That is an advantage. An advantage Helium got on monstro. But then also, mostro has other advantaged over helium, call it "better pixels" if you will. All in all, for me Monstro for most aplications trump helium. But I totally see why some DP´s think differnet on that point. All depending of what you want to achive. Red covered their bases very well with these two sensors I think.

 

[Hrvoje Simic]

Helium. Small image circle compared to dragon and Monstro, still 8k resolution.

There is correlations between speed, FOV and DOF. Heliums smaller circle simply gives you more light per pixel than for example monstro or Dragon when the same DOF and FOV is achieved if the lens in question is fitted for the sensor.

Its the same correlation that makes the Supreme Primes slower than the Master primes.

So for practical purposes, you can for example get a full 8k image out of your big optimo zooms with helium. Making a zoom with the same FOV range and speed for Monstro is seriously difficult, not to say impossible. You simply loose one of the 3 above. Like, if you enlarge the image circle and you loose speed.

He's saying if absolute exposure level is important, the T1.3 Master Primes are brighter than T1.5 of Supreme Primes by about a 1/3 stop.

That said, it ignores that Monstro is cleaner than Helium. That ~1/3 stop discrepancy can be compensated with ISO on Monstro and you'll probably still have a cleaner image than Helium. So now you're back to the same exposure level, same pixel count, a cleaner image, and still shallower DoF with the Monstro.

Got it. Thanks guys.

If 8K on S35 is a priority, there is no competition there. Helium carries a lot of advantages.

That also has a price though, and lower pixel count with larger pixel size has advantages as well. Glad to see Red offered multiple routes.

 

[AndreasOberg]

For us we decided to go with Helium mostly because of lenses. We want to use Canon cn20x50 zoom which has a whopping range of 50-1500mm. However it only covers 5.5-6k on Monstro and on Gemini about 4.5k. Helium it covers about 7.5k (sometimes 8k)

4.5k just felt too low compared to Helium 7.5k when you want a 4k output. After using the Helium it really impresses me how sharp the you can make. You get an extremely sharp 4k output which I could never get with the MX at 5.1k. At 4.5k it would be worse again.

Monstro is about 1.3 stop better when it comes to noise and seems to have a bit better colors. For us it is just not worth it beause how much it limits our lens choice.
Pulling focus at 8k is frustrating though when we only have a 1080p output. UHD output would make a big difference.
/Andreas