Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Extreme Purple Fringing on Dragon

Dana Rice said:
I was shooting the MX so I wasn't even aware of the issue till I took the footage home to transcode. I've never seen anything like it on MX and I've been shooting that almost daily for the last 5 years or so. In regards to lenses... I had a Sigma 18-35 on the MX... I believe the Dragon had the Canon EF 1.4 on.

Also what about the back of her hair, near the veil? Looks like some weird stuff going on there to.

I'm going to look into the Dragoncolor1 in a moment and will post the result once I do.
Click to expand...

That Dragon frame looks like having heavily overexposed backlight from the window behind the couple, that purple effects in her hair is probably the reflexion/ghosting of the flare. I was on MX before Dragon and especially regarding the highlight roll off I was impressed by the improvement. It´s probably a sum up of different things, if you find out keep us posted!
 
had this happen to me a week ago, there was a c100 on set as well, changed lenses,... from Leica R to Canon 50mm, 24-70... it happened on red dragon and c100. it was a dark hall with small window giving all the light and shooting against the window like that,.. it happened
 
There was definitely no scientific testing going on there with that shoot. It was just a very small paid gig with another operator who happened to have a Scarlet Dragon. I had agreed to transcode footage for the client and was shocked when I saw what that footage looked like in comparison to the MX. Sounds like the lens may very well have been the culprit... that said, I assumed it was the Dragon sensor because I had did a quick search on reduser before I posted and found a couple threads claiming Dragon was much worse in that regards compared to MX. Here's one thread I had found in particular... this one was same lens and settings just different sensors.

http://www.reduser.net/forum/showthread.php?146380-Purple-Fringing-WORSE-on-Dragon-PHOTOS

Also my past experiences with Dragon, it has always looked the same or worse than MX footage when doing 2 cam shoots. I have yet to see Dragon footage next to MX and be wowed by the Dragon or disappointed by the MX, but on the flip side I've been disappointed at least a few times by the Dragon footage compared to the MX. Admittedly, nothing has ever been a scientific test with same lenses, settings, etc. However, at some point you'd think the Dragon footage would've outshined the MX on the 2 cam shoots I've done. One was even a 3 week feature with an Epic Dragon from Panavision with a set of Primos vs my MX with DSLR glass, I transcoded the whole thing, so I saw every shot... they always looked the same or MX seemed to look a little better, especially in the lower light stuff.

Another shoot I did had a dark skinned black girl performing on a dimly lit stage with blue light hitting her... with the MX footage I was able to easily bring her natural skin tone back by using the WB Kelvin and Tint sliders in Red Cine. While the Dragon footage was uncorrectable.... no matter what I did she was dark blue or purple.

Obviously I want to like the Dragon sensor... I have a Scarlet-W on preorder, so I'm about to be invested in it, I'm just a little worried it might not perform as well as my MX. I definitely plan to do some side by side real testing once I get it to make sure I'm not making a mistake. And if I ever shoot with the operator from the purple fringing shoot again, I'll be sure to test his lens on my camera with similar lighting conditions if possible.
 
When this has happened to me in the past if had pretty good luck removing it in post using Resolve. I selected this purple color and area and used various tools to make it less visible, such as desaturating the purples etc.
 
Dana Rice said:
There was definitely no scientific testing going on there with that shoot. It was just a very small paid gig with another operator who happened to have a Scarlet Dragon. I had agreed to transcode footage for the client and was shocked when I saw what that footage looked like in comparison to the MX. Sounds like the lens may very well have been the culprit... that said, I assumed it was the Dragon sensor because I had did a quick search on reduser before I posted and found a couple threads claiming Dragon was much worse in that regards compared to MX. Here's one thread I had found in particular... this one was same lens and settings just different sensors.

http://www.reduser.net/forum/showthread.php?146380-Purple-Fringing-WORSE-on-Dragon-PHOTOS

Also my past experiences with Dragon, it has always looked the same or worse than MX footage when doing 2 cam shoots. I have yet to see Dragon footage next to MX and be wowed by the Dragon or disappointed by the MX, but on the flip side I've been disappointed at least a few times by the Dragon footage compared to the MX. Admittedly, nothing has ever been a scientific test with same lenses, settings, etc. However, at some point you'd think the Dragon footage would've outshined the MX on the 2 cam shoots I've done. One was even a 3 week feature with an Epic Dragon from Panavision with a set of Primos vs my MX with DSLR glass, I transcoded the whole thing, so I saw every shot... they always looked the same or MX seemed to look a little better, especially in the lower light stuff.

Another shoot I did had a dark skinned black girl performing on a dimly lit stage with blue light hitting her... with the MX footage I was able to easily bring her natural skin tone back by using the WB Kelvin and Tint sliders in Red Cine. While the Dragon footage was uncorrectable.... no matter what I did she was dark blue or purple.

Obviously I want to like the Dragon sensor... I have a Scarlet-W on preorder, so I'm about to be invested in it, I'm just a little worried it might not perform as well as my MX. I definitely plan to do some side by side real testing once I get it to make sure I'm not making a mistake. And if I ever shoot with the operator from the purple fringing shoot again, I'll be sure to test his lens on my camera with similar lighting conditions if possible.
Click to expand...

I did a dynamic range test between Scarlet MX and Red Raven that day; I keep meaning to cut it together an put it up online, but I just haven't had the time. When I do, I'll try to remember to send you a PM so you can check it out.
 
Shooting Dragon against a hot window (especially with window panes) can often bring this out. I had a similar experience recently where I had to pull an Angenieux wide zoom due to CA and use the GL 18-35 and the 24-290 instead on a two-camera set-up. Fortunately, we caught it immediately when framing the shot, but anytime you're shooting the Dragon toward a hot source with high contrast, check your edges.
 
The OP raises an interesting point though.

So are we saying that this is entirely lens choice. Swap the lenses over then the MX would have shown the same? I get the impression that this is not the case.

Clearly lenses wide open will exhibit various forms of CA. The purple fringing on high contrast areas is a classic.

But purple fringing can also be exacerbated by a given sensor, there was a time when it was the result of leaking energy into adjacent pixels. Now clearly those designs have evolved a lot but this really does suggest that the sensors are performing differently and the MX is doing better.

The Dragon sensor seeing more range? Well they're both clipping as the guy leaves the church. And that's a creative situation that does happen - someone leaves into the bright light - you can't ND the door and you don't want to fill the person in this case.

The glow around the guy leaving on the MX looks great. The Dragon is unusable and even with post work, would prove a tough shot to fix.

I too have a SW on preorder, i can't offer any experience specific to Red though.

It would be ideal to have the two side by side and the same lenses and critically look at the difference - has anyone done this?

I'm not trying to troll or anything, but lots of us shoot into the light and it is a slight concern

cheers
Paul
 
About 3 months ago I had an agency client who insisted on filming the talent in front of a super-hot floor-to-ceiling window, letting it gently blow out, while we keyed the talent with a 2K to keep exposure. He wanted us to film at f1.something. Wide,wide open- so we had to use ND of course on the lens. I had exactly this same problem, and we were not using crap lenses: we were using Zeiss CP2s.

I came back last month for the agency, and shot pickups using another camera (that claims to have 14 or 15 stops DR), same lenses, same setups. Zero issues.

Just a datapoint.
 
I need to point out that there's too many variables in this example to reach any conclusions- the ND, angle of the shot, etc. had something to do about it, because I have shot into the light a LOT and never gotten this weirdness before, and with the SAME Dragon, no less.
 
My thoughts are, to really nail the debate, that a controlled test should be done on the (same) 2 cameras - duplicating the same exposure strategy on both. This should also take into account the OLPF (ie Skin-tone most people use 320), but also, at least, the same lenses on both cameras.

I agree that there are too many variables at the moment.
 
Last edited:
patrickortman said:
About 3 months ago I had an agency client who insisted on filming the talent in front of a super-hot floor-to-ceiling window, letting it gently blow out, while we keyed the talent with a 2K to keep exposure. He wanted us to film at f1.something. Wide,wide open- so we had to use ND of course on the lens. I had exactly this same problem, and we were not using crap lenses: we were using Zeiss CP2s.

I came back last month for the agency, and shot pickups using another camera (that claims to have 14 or 15 stops DR), same lenses, same setups. Zero issues.

Just a datapoint.
Click to expand...

What were the two cameras? Also, CP2s (while not terrible) are not incredibly high quality. They are the same optics as the ZE/ZF lenses, just in larger cine housing. I own the 50mm and 85mm ZF lenses, my 50mm has glowy highlights wide open, and those lenses have been known to exhibit CA when wide open. They can look fantastic but they do have a few flaws that show themselves in challenging situations like this.
 
patrickortman said:
I had exactly this same problem, and we were not using crap lenses: we were using Zeiss CP2s.

I came back last month for the agency, and shot pickups using another camera (that claims to have 14 or 15 stops DR), same lenses, same setups. Zero issues.
Click to expand...

Zeiss CPs have more than average amount of chromatic aberration.

As to your 2nd shoot not having the problem, my experience is how much CA you see is dependent on how many stops difference there is between the two sides of the edge.

For instance, if the window behind is 10 stops brighter than the foreground object, you will see CA pretty badly.

If the window is only 5 stops brighter, it will still look the same to the camera because it's still going into roll-off after Redgamma is applied, but the CA will not be nearly as bad.

In your case, if it was the same shot, it could be the difference between a sunny day out the window and an overcast day. Both might be blown out, but the measure of light between two halves of the CA edge would be quite different even though they look the same.
 
Andy White said:
My thoughts are, to really nail the debate, that a controlled test should be done on the (same) 2 cameras - duplicating the same exposure strategy on both. This should also take into account the OLPF (ie Skin-tone most people use 320), but also, at least, the same lenses on both cameras.

I agree that there are too many variables at the moment.
Click to expand...

Hey Andy, I did something like that. Check it out: http://www.reduser.net/forum/showth...SE-on-Dragon-PHOTOS&highlight=purple+fringing
 
Hi Guys,

First: I second (third? fourth?) the advice to go to DragonColor1. It helps. Not a fix. But it helps.

Second: Let's hang back a sec on the CP2s, and our overall urge to blame the lenses. Do we really want to say that a RED Dragon sensor'd camera needs lenses that cost over $5K each to make pictures that work? I can understand if we're talking cheap lenses, like a $50 eBay special. But isn't it a bit odd that we're getting this error on Dragon with a wide variety of lenses, from Canon L to Nikon to Sigma still lenses, and CP2 (rehoused ZF2s, sure) and even higher? Maybe, just maybe, it's the Dragon that is the problem, and we need to find a way to make it work without blaming "poor quality $4900 lenses".

No, I don't own the CP2s, I rent lenses. I don't have a horse in this race, outside of loving to shoot with Dragon. And I've had the same issue with an Optimo, So it ain't the lens. It's a combination of the Dragon sensor and this kind of shooting, where people are heavily backlit.

Clients do not care that another camera (Alexa, F55, F5, FS7, C300MII) isn't quite as nifty as a RED. They see the CA, and freak out. It's our job to deliver kickass images that don't have issues. So, how do we fix this problem?
 
how about a DragonColor3? If it is improved with DC1 over DC2 would it not stand that it could be eliminated entirely with a new DC3? It is my understanding that the specific purple color in some CA with some lenses is outside of the color gamut in DC2 and to a lesser extent in DC1. Because of this the purple spikes in color and brightness. CA is a thing...we know it and most understand how to control it....but lets not have a color science that seems to exaggerate it.

My apologies in advance, as I do not know what is involved in whipping (joke) up a DragonColor 1,2, or 3. It must be difficult and time consuming. I'm just hoping that some help is on its way...soon. As Patrick has indicated for himself, I too have found myself explaining to clients what that bit of ugliness is at times.
 
patrickortman said:
Hi Guys,

First: I second (third? fourth?) the advice to go to DragonColor1. It helps. Not a fix. But it helps.

Second: Let's hang back a sec on the CP2s, and our overall urge to blame the lenses. Do we really want to say that a RED Dragon sensor'd camera needs lenses that cost over $5K each to make pictures that work? I can understand if we're talking cheap lenses, like a $50 eBay special. But isn't it a bit odd that we're getting this error on Dragon with a wide variety of lenses, from Canon L to Nikon to Sigma still lenses, and CP2 (rehoused ZF2s, sure) and even higher? Maybe, just maybe, it's the Dragon that is the problem, and we need to find a way to make it work without blaming "poor quality $4900 lenses".

No, I don't own the CP2s, I rent lenses. I don't have a horse in this race, outside of loving to shoot with Dragon. And I've had the same issue with an Optimo, So it ain't the lens. It's a combination of the Dragon sensor and this kind of shooting, where people are heavily backlit.

Clients do not care that another camera (Alexa, F55, F5, FS7, C300MII) isn't quite as nifty as a RED. They see the CA, and freak out. It's our job to deliver kickass images that don't have issues. So, how do we fix this problem?
Click to expand...

AMEN! This quirk happens with lenses at all levels, from $100 to $10,000. Sure, maybe the CA is there, and the Dragon is just "sooo" good at seeing stuff that it exacerbates it. But can we just modify the color matrix so that it doesn't? It's unbelievable to me that this issue has gone 4 years without being addressed. Then again, maybe I shouldn't be surprised since every time it's brought up, users are quick to shift the blame to the lenses...

The bottom line is, if I'm the kind of person who needs to shoot high contrast (esp. backlit) subjects, then I might look elsewhere for a camera, since others seem to be FAR less susceptible to purple fringe. Is Red ok with that?
 
I'm going to make another call for an R3D file posted to this thread. It's important to see the raw data to be able to give proper advice.
Also, no mention of which OLPF was used.

ONe thing is for sure. When you clip all three channels on Dragon or MX you are going to be looking at these kinds of issues. PLus this was a controlled setting where stuff like that could be avoided. I'm not entirely sure you can blame the camera for this. :)

David
 
You must log in or register to reply here.
Back
Top