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 :)

Soft, Noisy and lots of Artifacts

David Mullen ASC said:
The EX1 footage has a lot of sharpening happening in it, plus the blacks are blacker, which is all going in increase the illusion of sharpness.

Here is a test I did comparing the EX3 to the Red One, all reduced to HD and then cropped. The sharpness settings in the EX3 were turned down:....

Obviously if I had finished in 4K the difference between the Red One and the EX3 would have been more marked, obvious, but dumbed-down to HD the two weren't as far off. However, the Red image has the advantage of being detailed without having artificial sharpness applied, and therefore looks more natural and film-like, less electronic.
Click to expand...

David,

I know the EX1 has sharpness added and has its own levels etc all done in camera.. I guess my comparison was just demonstrating that the 3K image Im shooting with the Scarlet (seems to me) should be sharper than the 1080p image Im getting from the EX1, especially after the 3K image is scaled down to 1080p. I know the R3D's arent tack sharp as someone mentioned earlier but Im having to sharpen the image just to bring it into focus. The Canon 550D chart you shot looks identical to what I am getting with the Scarlet and seeing on my 1080p monitor.
 
One thing: The lens mount of the 18-55 IS is plastic, I believe. Try putting it on an SLR and see if you get similar images. Perhaps the mount is just slightly misaligned somewhere....
It seems you've done everything else correctly, and the lens would appear to be the weak point.
 
Well, I wouldn't expect the EX1 to do badly at 1080P because it's a pro 3-sensor camera, and without matching degrees of edge enhancement it's hard to judge the difference. You should reshoot the EX1 stuff with all the sharpness turned down for starters so you don't have that distraction. Or add a lot of sharpening to the Scarlet footage.

Trying to see through the mismatch in sharpening, I don't much difference in actual detail recorded though I agree that the Scarlet shots looks incorrectly soft for some reason. But you wouldn't expect a lot of difference if scaling everything down to the lower common denominator, you should try the other direction. And turn off the sharpening in the EX1. And try two different lenses on the Scarlet. And bring in a second Scarlet and try the two different lenses on that.

You need to begin to track down whether it is the lens or the specific Scarlet camera in order to know whether the results are atypical. To do that you need a second lens and a second Scarlet.
 
David,

to me the Red Image looks like also having quite an amount of USM applied, look at the white outlines around the text on the post-it.
RED did rise my expectations high when claiming to be able to do a still ad campaign, coming from a medium format photographic background all I have seen from my Scarlet looks kind of soft, too.
Don´t get me wrong I really like the Scarlet but I have to admit I was feeling exactly like the OP when I took a first look at the footage.

Cheers, Ulf
 
Last edited:
You're wasting your time with lenses. It's not a lens problem. It's the compression. I almost guarantee it.

Adjusting back focus is also a waste of your time. Even if the backfocus *WAS* off, it wouldn't matter since it would only prevent the markings on your lens from actually being the correct distances. That's why still glass goes beyond infinity... in case the back-focus is off you can still focus beyond infinity. If you were getting a tilt-shift focus effect I would say there was a problem with back-focus that needed servicing. If I saw lots of CA everywhere I would say that the lens is crappy. The problem is in the capture. That's very obviously over-compressed wavelet compression.

Try 6:1 4k. If it still looks like that, send an r3d to red because something is wrong with your camera. This isn't optical. Maybe that resolution has a glitch in the redcode encoder.
 
Gavin Greenwalt said:
You're wasting your time with lenses. It's not a lens problem. It's the compression. I almost guarantee it.

Adjusting back focus is also a waste of your time. Even if the backfocus *WAS* off, it wouldn't matter since it would only prevent the markings on your lens from actually being the correct distances. That's why still glass goes beyond infinity... in case the back-focus is off you can still focus beyond infinity. If you were getting a tilt-shift focus effect I would say there was a problem with back-focus that needed servicing. If I saw lots of CA everywhere I would say that the lens is crappy. The problem is in the capture. That's very obviously over-compressed wavelet compression.

Try 6:1 4k. If it still looks like that, send an r3d to red because something is wrong with your camera. This isn't optical. Maybe that resolution has a glitch in the redcode encoder.
Click to expand...

Is 8:1 considered highly compressed?? I thought 8:1 was a pretty standard choice for a compression ratio.. best quality/disk space ratio. I think I've tried lower compression but I'll give it another shot and try again. I do see heavy artifacts in the shadows on a day shot which is the other issue I mentioned in the OP.
 
I agree with Gavin. I used to sell wavelet based digital video recorders and your frames have a similar compression look to the troubled areas.
 
Testing low compression ratio's:
http://66.147.244.90/~laymenmi/red/Wide_6_to_1.jpg
http://66.147.244.90/~laymenmi/red/55mm_6_to_1.jpg
http://66.147.244.90/~laymenmi/red/wide_3_to_1.jpg

I wouldnt surprise me at all if there was some bad compression thing happening. Seems to me the most likely problem as I've noticed it from the first shot I took with this camera. Wouldnt surprise me either if the compression issue is related to the softness. I've tested this camera with 3 different lenses that I have access to, tried a back focus adjustment and everything else that can be done save a side-by-side test which I'll do tomorrow.
 
@Michael - Send that last .R3D and a log file over to your Bomb Squad Rep. I agree with Gavin and Jacek. If you start sharpening into the file you begin seeing things that aren't exactly normal.

On a side note and something I've observed now for a while; shooting with still glass wider than 24mm is a mixed bag of nuts when it comes to Scarlet and Epic. If you're looking for a general rule of thumb avoid the more "compact" lenses if possible. To get the best images out of Red cameras you want lenses that are designed more telecentricly to reduce the degree of the angle of incidence of the light rays hitting the imaging plane.

It's not talked about much here, but it's 100% the truth.
 
Phil Holland said:
@Michael - Send that last .R3D and a log file over to your Bomb Squad Rep. I agree with Gavin and Jacek. If you start sharpening into the file you begin seeing things that aren't exactly normal.

On a side note and something I've observed now for a while; shooting with still glass wider than 24mm is a mixed bag of nuts when it comes to Scarlet and Epic. If you're looking for a general rule of thumb avoid the more "compact" lenses if possible. To get the best images out of Red cameras you want lenses that are designed more telecentricly to reduce the degree of the angle of incidence of the light rays hitting the imaging plane.

It's not talked about much here, but it's 100% the truth.
Click to expand...
That sounds like an interesting topic!

Which Canon mount wide angle lenses do you feel meet those qualification when used with the Scarlet/Epic?
 
Phil Holland said:
@Michael - Send that last .R3D and a log file over to your Bomb Squad Rep. I agree with Gavin and Jacek. If you start sharpening into the file you begin seeing things that aren't exactly normal.

On a side note and something I've observed now for a while; shooting with still glass wider than 24mm is a mixed bag of nuts when it comes to Scarlet and Epic. If you're looking for a general rule of thumb avoid the more "compact" lenses if possible. To get the best images out of Red cameras you want lenses that are designed more telecentricly to reduce the degree of the angle of incidence of the light rays hitting the imaging plane.

It's not talked about much here, but it's 100% the truth.
Click to expand...
I don't think so Phil- all lenses for DSLRs have to clear the mirror so they are retrofocal and, by definition, nearly telecentric. The FL hasn't got much to do with it.
The size of the rear element is mostly related to the speed of the lens not the telecentricity. Telecentricity will severly affect the corner illumination but not the performance. The example here are the Leica M lenses that are not retrofocal (much), not telecentric but are great performers. The only aspect they suffer is corner illumination. Leica rectified that by tilting the off axis micro lenses on their camera sensors...
 
I think it does, but you certainly have more lens knowledge than I do.

After examining material shot on S35 and Red and stills from DSLRs it made me hunt for something that could be creating the variation in optical quality on certain lenses. So we set up a test with some MPs, still lenses, and shorter cinema glass in 2010. Something tells me that the distance of the OLPF and the gap between the sensor comes into play. Jarred also stated that more telecentric designs perform better as well.
 
Of course they do. However Arri standard primes, super speeds are absolutely not telecentric and, other than the corner illumination issues(vignetting)- everybody seems to like their performance...
 
I should also add we didn't shoot lasers through the damn things to really know what was up, but once we settled on the best optical performer we made a lens selection for the project.
 
I don't know if I'd say everybody, but I agree they perform well. At the same time though, not as "well" as others. That's getting into subjective territory though.

Back on point though. Michael, email your rep with that file and/or open a support ticket. There's something funky in your soup.
 
Phil Holland said:
I don't know if I'd say everybody, but I agree they perform well. At the same time though, not as "well" as others. That's getting into subjective territory though.

Back on point though. Michael, email your rep with that file and/or open a support ticket. There's something funky in your soup.
Click to expand...

Did that Friday. No news yet.
 
I still think you need to shoot a test of your lens and a reference lens on your Scarlet and the same tests on a second Scarlet to figure out whether it's the lens or the camera. Otherwise there are too many variables involved to track the problem.
 
As much as compression plays a role in you "forrest" shot, it shouldn't effect your shots of a focus chart. RED ONE shoots a max of 7:5:1 compression and most of my shots have been around 9:1 or Redcode 36. Compression is vital in high detail scenes but a simple chart shouldn't need that low of compression to get a sharp image.

Compression is definitely playing a role in your exteriors but not your charts.
 
Michael McCaffrey said:
Testing low compression ratio's:
http://66.147.244.90/~laymenmi/red/Wide_6_to_1.jpg
http://66.147.244.90/~laymenmi/red/55mm_6_to_1.jpg
http://66.147.244.90/~laymenmi/red/wide_3_to_1.jpg

I wouldnt surprise me at all if there was some bad compression thing happening. Seems to me the most likely problem as I've noticed it from the first shot I took with this camera. Wouldnt surprise me either if the compression issue is related to the softness. I've tested this camera with 3 different lenses that I have access to, tried a back focus adjustment and everything else that can be done save a side-by-side test which I'll do tomorrow.
Click to expand...


For a 2.8k image these look much more in line with what I would expect. Just add a little unsharp mask after your grade. That being said there still seems to be some odd optical properties or compression artifacts or something above the roof for instance in the 55mm almost like it's catching a reflection. You aren't shooting through a mattebox are you?
 
Gavin Greenwalt said:
For a 2.8k image these look much more in line with what I would expect. Just add a little unsharp mask after your grade. That being said there still seems to be some odd optical properties or compression artifacts or something above the roof for instance in the 55mm almost like it's catching a reflection. You aren't shooting through a mattebox are you?
Click to expand...

Yes I am shooting through a mattebox but there is no glass/filters in it.
 
You must log in or register to reply here.
Back
Top