DRAGON-or-HELIUM

[John Marchant]

John are you seeing any DR changes in Helium vs Dragon? Is the native ISO indeed higher?

I'm reserving judgement 'til I shoot a Xyla chart.

I did a basic DR test racking exposure in camera by stops on a split card lit with a one-stop difference. Main point of note was that RG4 doesn't seem well suited to representing Helium's available DR, which is not surprising. In RLF I see 14 stops at 800ISO that way, and 15-16 at higher ISO (1600-3200) - That's pretty much parity with Dragon.

Not sufficiently scientific to publish though. Anecdotally I don't feel as though the camera suffers limited DR in any sense, and highlight rolloff at sensible ISOs is really pleasing.

 

[Phil Holland]

Would be cool if you had an option to debayer and subsample down the 8k to 4k inside of the camera. Then you could film fullframe in both 4 and 8k. I mean 8k down to 4k should be easier to sample down compared to say 7k to 4k.

You can do this if you are happy with an Intermediate Codec like ProRes or Avid DNxHR in 4K while dual recording with REDCODE RAW at 8K.

The Intermediate Codecs are demosaic'd/debayered and scaled to 4K or 2K.

 

[luigivaltulini]

I'm reserving judgement 'til I shoot a Xyla chart.

I did a basic DR test racking exposure in camera by stops on a split card lit with a one-stop difference. Main point of note was that RG4 doesn't seem well suited to representing Helium's available DR, which is not surprising. In RLF I see 14 stops at 800ISO that way, and 15-16 at higher ISO (1600-3200) - That's pretty much parity with Dragon.

Not sufficiently scientific to publish though. Anecdotally I don't feel as though the camera suffers limited DR in any sense, and highlight rolloff at sensible ISOs is really pleasing.

John will see the same .. for now Dragon has about 1 stop in more latitude under normal conditions.
I have not tried to shoot at ISO higher than those for this comparison.
True, we have to wait the new science of color to judge.
I'm looking forward to this soon, so I can decide whether to upgrade.

Keep us aware if you have news eheh

 

[Tommaso Alvisi]

I'm reserving judgement 'til I shoot a Xyla chart.

I did a basic DR test racking exposure in camera by stops on a split card lit with a one-stop difference. Main point of note was that RG4 doesn't seem well suited to representing Helium's available DR, which is not surprising. In RLF I see 14 stops at 800ISO that way, and 15-16 at higher ISO (1600-3200) - That's pretty much parity with Dragon.

Not sufficiently scientific to publish though. Anecdotally I don't feel as though the camera suffers limited DR in any sense, and highlight rolloff at sensible ISOs is really pleasing.

Thanks John!

So do you think native ISO is around 2000iso?

OT Please reply to my PM!

 

[Jake Wilganowski]

I'm waiting for the color science and official release after tweaks, because, I dug through the helium and dragon r3d files in redcine, and I'm not seeing anything crazy awesome yet. I definitely see the color difference, but I had them matched in about 30 sec of tweaking, so that's not a huge deal to me. I think we should wait for the official tweaks before judging because in my opinion I'm not seeing the benefits yet, comparing at 1:1, I'm sure 8k downrezzes cleaner simply because it's bigger, but not a huge difference in noise yet in my opinion at 1:1, and seems the highlight retention is worse, at least until we get some official color science

 

[Derek Carlson]

If native ISO for Helium is 2000, that's incredible. Base ISO for slog-2 on the a7s is 1600. This may just be RED's response to "we don't make low light cameras".

 

[Brad Grimm]

Well, we know they are reworking the image processing pipeline along with new color science, so hopefully we gain that extra Highlight retention (deal breaker). I really don't want to be stuck shooting at iso 1600 or higher all the time... guess we will wait and see, I know many of us are heading to VV, so the real question is do we want the Helium intermediate...

 

[Akin A]

Looking forward to multiple testers putting out their results this time around, so we can compare before concluding.
I remember word on the street was that DSMC1 Dragon LLO/STH difference was 1-1.5 stops, but my own tests showed .66 stops...

The more people who test, the better. If we can get some sort of standard for our testing, that would be nice for consistency.
I don't plan to upgrade anytime soon, so count me out this round, lol.

 

[John Marchant]

I'm waiting for the color science and official release after tweaks, because, I dug through the helium and dragon r3d files in redcine, and I'm not seeing anything crazy awesome yet. I definitely see the color difference, but I had them matched in about 30 sec of tweaking, so that's not a huge deal to me. I think we should wait for the official tweaks before judging because in my opinion I'm not seeing the benefits yet, comparing at 1:1, I'm sure 8k downrezzes cleaner simply because it's bigger, but not a huge difference in noise yet in my opinion at 1:1, and seems the highlight retention is worse, at least until we get some official color science

The way I see it, the pixels are individually as clean as dragon, and the colour science is unarguably better - but then I have some good data on that which I'll be able to share soon. Not every scene challenges colour reproduction.

And then on top of that, there are 1.8x as many photosites on the sensor versus Dragon 6K.

 

[Christoffer Glans]

One question I have is about what framerates you get with the "sideways" option for MG upgraders. If we are only going to be able to shoot 25 fps or 30 fps in 8K then we loose high speed capabilities in s35mm. Does anyone know what the framerates will be on the Epic-W (which I presume is the Helium sideways upgrade)

 

[Chris Kennedy]

One question I have is about what framerates you get with the "sideways" option for MG upgraders. If we are only going to be able to shoot 25 fps or 30 fps in 8K then we loose high speed capabilities in s35mm. Does anyone know what the framerates will be on the Epic-W (which I presume is the Helium sideways upgrade)

Hopefully all your/our questions will be answered in the announcement next month! Hopefully the epic W is 8k but since it's a 35mmish sensor so I'd assume we'd loose the higher framerates with the higher resolution.
We shall see!

 

[PatrickFaith]

Helium really needs a 18.5 plus sensor option, both for the anamorphic crowd but also the stills guys. Without that there is just too huge a gap to the weapon VV, which seems like it will always be costly with very limited availability. This would also separate red from all the "me too" 4K cameras (i.e. Cannon 700 with only a 15mm sensor height), and put it even to the Alexa for anamorphic shooters.

 

[Antony Newman]

If Helium sensor Yield is comparable to Dragon (for the same area) ...

I wonder how many people would be willing to pay to jump from the VV Dragon queue to a Helium VistaVision sensor queue (16:9 format / max of 40 fps FF : 10,957 x 6,163)

This would reduce the Dragon VV queue down to those want it yesterday (or need Higher Frame rates in FF) and give others who think that Helium pixels represent the future, an opportunity to wait a bit longer to have the best of both worlds.

AJ

 

[Anthony Berenato Jr]

If Helium sensor Yield is comparable to Dragon (for the same area) ...

I wonder how many people would be willing to pay to jump from the VV Dragon queue to a Helium VistaVision sensor queue (16:9 format / max of 40 fps FF : 10,957 x 6,163)

This would reduce the Dragon VV queue down to those want it yesterday (or need Higher Frame rates in FF) and give others who think that Helium pixels represent the future, an opportunity to wait a bit longer to have the best of both worlds.

AJ

Sadly the problem here is power / throughput on the boards. I'm sure it will happen eventually, just not now.

 

[Antony Newman]

Good point.

If the rate at which Pixels being processed & stored increased, we could end up with a portable microwave (or GalaxyNote 7!!)

Rate of pixels :

Helium 8k = 8192 x 4320 x 75 fps = 2,654,208,000 pixels a second (1.896 aspect ratio)
Helium 9k = 8857 x 4988 x 60 fps = 2,653,715,760 pixels a second (16:9 aspect ratio)
Helium 11k = 10,957 x 6163 x 39 fps = 2,633,591,649 pixels a second (16:9 aspect ratio)

Ie this would represent a reduction in processing & storage requirements.

Rate of horizontal lines :

Helium 8k = 4320 x 75 fps = 324,000 lines a second
Helium 11k = 6163 x 39 fps = 240,537 lines a second

Ie this would also represent a reduction in processing (and likely power) requirements.


AJ

 

[Tom Gleeson]

The way I see it, the pixels are individually as clean as dragon, and the colour science is unarguably better - but then I have some good data on that which I'll be able to share soon. Not every scene challenges colour reproduction.

Fom my perspective if the Heluim matches Dragon DR with an added stop of sensitivity the improvement in colour will be worth the ticket. When I tested the Dragon chip in a DSMC2 it matched the Alexa DR. I am happy with that 98% of the time. While more DR wouldn't hurt I would first take superior colour reproduction. When it came to colour Alexa kicked the MX sensor in the nuts then Dragon came along and matched it. I will be very impressed if Heluim can out perform it.

Choosing between VV Dragon and S35 Heluim may end up a tough decision.

On a tangent the pressure is mounting on Arri. While the Mini is selling and renting well if you take a close look the Weapon Dragon easily matches it and the Heluim may surpass it. No one should underestimate Arri and I have high expectations for their next generation sensor which must be around the corner?

 

[luigivaltulini]

Of course, the new science of color should give more substance to Helium and DR.
For everyone is different I prefer to have more latitude in the highlights at the same aperture and not having to lift the ISO and close the iris or add ND.
I like more the idea to throw open the iris without worry of clipping

Is the choice is difficult between Helium and VV ...

I think that you are coming closer to Arri release of a new sensor, but it is not easy now to make a big difference. We will wait this time

Linear gamma

[/URL][/IMG]

 

[luigivaltulini]

The more people who test, the better. If we can get some sort of standard for our testing, that would be nice for consistency.
I don't plan to upgrade anytime soon, so count me out this round, lol.

Yep this is a good point .

 

[Sabovic Adis]

So, Helium clips 2-3 stops earlier than Dragon?! :biggrin:

 

[Kemalettin Sert]

If highlights are Helium's weakness whats the advantage over Dragon? I dont see that much sharpness improvement over dragon.Do you guys see it?