Scarlet-W Bad Compression in HFR, Worse Than Epic-MX, Any Weapon Examples Out There?

[C. Phillips]

Hello all,

I received my Scarlet-W last week and used it right away. The first thing I noticed was how bad the compression was in the higher frame rates, like 120 fps and beyond. Those frames rates were a big influence in my getting the camera, but when I saw how 120 and up started having those blocky artifacts like you see in low quality mp4 videos, I am on the borderline of considering it unusable footage. I own a Scarlet-MX and have used the Epic-MX on many occasions and never before have I seen these kinds of artifacts.

I would like to know if any of you kind souls out there with any of the other Weapon models, Magnesium and up have any test footage at 120, 150, 200, and 250 fps clips so I can compare. I know Youtube and Vimeo will kind of degrade it a little, but I kid you not, I noticed the difference immediately when shooting with my standard 5.0 RED LCD. I will upload some samples myself for you to compare.

Thank you!

Sincerely,

C. Phillips

 

[Brandon J.F.]

What were you filming?

 

[Christoffer Glans]

Hello all,

I received my Scarlet-W last week and used it right away. The first thing I noticed was how bad the compression was in the higher frame rates, like 120 fps and beyond. Those frames rates were a big influence in my getting the camera, but when I saw how 120 and up started having those blocky artifacts like you see in low quality mp4 videos, I am on the borderline of considering it unusable footage. I own a Scarlet-MX and have used the Epic-MX on many occasions and never before have I seen these kinds of artifacts.

I would like to know if any of you kind souls out there with any of the other Weapon models, Magnesium and up have any test footage at 120, 150, 200, and 250 fps clips so I can compare. I know Youtube and Vimeo will kind of degrade it a little, but I kid you not, I noticed the difference immediately when shooting with my standard 5.0 RED LCD. I will upload some samples myself for you to compare.

Thank you!

Sincerely,

C. Phillips

What were you filming and what was the exact specifications for that shot?

 

[Demetri Zuev]

Wow, this sounds concerning...

Here's a test video from Brian Nguyen that features both 120 and 240fps. I don't see blocky compression at all at 120fps. 240 could've been a bit sharper, but no sharpening was applied in post, these are straight from camera

5600K, ISO 800, Standard OLPF
23.98 footage was shot in 5k, 120 fps in 4k 14:1 compression, and the 240fps footage was shot in 2k (the noisy and flickering footage).
No post noise reduction or sharpening was done.

https://vimeo.com/160315302

He also has 120fps R3D available for download in his blog post here - http://www.thebriannguyen.com/2016/03/red-scarlet-w-first-footage-dragons-and-puppies/

 

[Levi Arnold]

I'd like to know more or see some of this footage. I'm planning on getting a Scarlet-W to go with my Epic Dragon but not if it has these issues. Could you pop a short clip on Dropbox for us to look at?

 

[Zeb B]

If the background is low in fine detail and static with a couple moving elements then a camera compression can perform no artifact footage. Point it at a dynamic scene where every pixel is changing from frame to frame with fine detail and you might crash the codec. A dog on a sandy beach is easy - lots of leaves or fields of wheat in the wind with a cloudless sky in the background . . . not so much. Learn the limits of your gear by testing it to failure and reviewing the results before a paid gig

 

[Jon Betz]

It would be great to see some slow motion underwater footage. For me that's the ultimate codec test.

 

[Nick Morrison]

Scarlet-W has 140mb/s boards. Weapon Mag has 220mb/s and Carbon 300mb/s.

The compression will always be worse on Scarlet-W at HFR. Thats just math.

But again, as others have noted, WHAT you are shooting is just as important as WHAT your compression is. A complicted background (ie trees, foliage, water) is always hader to compress than a gently out of focus shot.

Again, theres a reason Carbon costs $50K, Mag costs $30K and Scarlet-W costs $10k.

But used wisely, they can all look (essentlly) the same.

 

[Matthew Quinn]

I rarely use frame rates over 50-60 fps, 120 would be the top of the mark for most of my projects anything more would warrant a phantom rental or something

 

[Kris Tait]

It's worth noting that Brian Nyugen (the gent with the Scarlet-W footage with the dog on the beach) also wrote on his blog post about the camera that he wouldn't use anything over 120fps as it got too compressed, soft and blocky. I thought it sound borderline scandalous that the newest sensors/cameras couldn't produced commercially usable images past 120 fps. Now we have a second owner stating the same. What does this say about the Raven? Pretty concerning atleast for me, where 240fps is required for a good 60-70% of my jobs. Please keep us updated guys...
thanks

 

[Kris Tait]

Brians quote: "Slow-motion and 2K 240 fps - Is it usable? ....Obviously a significant amount of the video was in slow motion, it was the primary feature I wanted to test in this video....It's very clear which are the 2K 240fps shots. They are noisy, soft, and have dancing artefacts (1:05 in the video). The question is, is it usable?.....To me, 2K is not a viable resolution to shoot when you're shooting RAW. At 2K (which is the only resolution that allows 240fps), you're effectively shooting smaller than Super 16..."

Link to Brian's entire blog post about the camera + problem: http://www.thebriannguyen.com/2016/03/red-scarlet-w-first-footage-dragons-and-puppies/

It's worth noting that he does however praise the 4K 120 fps to the heavens. Though the starter of this thread claims his 120fps at 4K is NOT usable. So where are we I wonder...

 

[Matthew Quinn]

240 fps? what are you shooting? and yea 240 fps would require cropping the sensor to a place where the image looses a lot of its luster. would prob look better if you shoot 120 fps then ramp it down in post a bit for a slower look. RED does high frame rates REALLLLLY well, but to a point.... As I see it, RED are best at 24 fps. at 120 fps RAW Red is still the best around. anything more may require a camera designed for super slow motion capture

 

[Zeb B]

Brians quote: "Slow-motion and 2K 240 fps - Is it usable? ....Obviously a significant amount of the video was in slow motion, it was the primary feature I wanted to test in this video....It's very clear which are the 2K 240fps shots. They are noisy, soft, and have dancing artefacts (1:05 in the video). The question is, is it usable?.....To me, 2K is not a viable resolution to shoot when you're shooting RAW. At 2K (which is the only resolution that allows 240fps), you're effectively shooting smaller than Super 16..."

Link to Brian's entire blog post about the camera + problem: http://www.thebriannguyen.com/2016/03/red-scarlet-w-first-footage-dragons-and-puppies/




It's worth noting that he does however praise the 4K 120 fps to the heavens. Though the starter of this thread claims his 120fps at 4K is NOT usable. So where are we I wonder...

Crap footage to use as any kind of a test. RED cameras deserve professionals posting first looks. Not filthy casuals with home movies ;-)

 

[Kenneth C Merrill]

Crap footage to use as any kind of a test. RED cameras deserve professionals posting first looks. Not filthy casuals with home movies ;-)

And we wonder why people think Reduser is full of assholes...

 

[Daniel Reed]

Someone yell for me? heh j/k

assholes...

Time to put on my serious face. :smiley:

To avoid any perceivable dancing compression noise,
think of the Raven & Scar-W as normal speed cameras w/ a little overcrank potential.
(remember the trusty wind-up bolex?)



48fps at 6:1 at 4.5K FF on Raven should look great - apparently Raven got a stealth upgrade to FF mode in 4.5K
60fps at 7:1 in 4.5K wide on Raven mode may be a reasonable limit for most scenes not shot in a white void.

For Scar-W, in most 5K modes, it may prove wise to consider not pushing higher then 40fps for 6:1~7:1
If going for wide mode, 48fps may be a good limit advisement to maintain 7:1

Comparing the two charts above, it appears Raven gets lower compression at every 4.5K FF speed notch the two cams share.
(unless a typo is afoot)

 

[Kris Tait]

Crap footage to use as any kind of a test. RED cameras deserve professionals posting first looks. Not filthy casuals with home movies ;-)

Not even going to dignify that comment with a response.

 

[Daniel Reed]

I'm pretty sure Zeb was joking around sarcastically... note the smiley wink at the end.

Not even going to dignify that comment with a response.

Crap footage to use as any kind of a test. RED cameras deserve professionals posting first looks. Not filthy casuals with home movies ;-)

:sarcasm:

 

[Kris Tait]

Hey Daniel
Some pretty sound predictions in your post, I'll be subscribing to this. Thanks for posting.

Though I have to admit, coming from 3 years with sony, I'm a little surprised to see that the high frame rates on Raven and Scarlet-W are more like party tricks rather than commercially viable features. I haven't researched it enough yet, but I'm curious why Red uses the sensor windowing technique for high frame rates as opposed to what some of the other systems utilise. I absolutely had to leave Sony because I couldn't deal with the color science anymore, but in all fairness they really have figured out how to get even their lower-end cameras to produce some impressive results in the HFR department, not using sensor-windowing. Going to have to look into this more. If only there was all the Red cam, flavor, color science and depth, with that one feature from sony...HFR. But I still have a hard time believing that anything above 120fps won't be usable on Raven or Scarlet-W. That's just too damn hard to believe. Hmmm...

 

[Demetri Zuev]

Comparing the two charts above, it appears Raven gets lower compression at every 4.5K FF speed notch the two cams share.
(unless a typo is afoot)

If there's no typo that is just wrong...a cheaper camera performing better than the more expensive and feature rich one, both made by the same company.

 

[Brody_B]

Wow, this sounds concerning...

Here's a test video from Brian Nguyen that features both 120 and 240fps. I don't see blocky compression at all at 120fps. 240 could've been a bit sharper, but no sharpening was applied in post, these are straight from camera

I always thought the 240fps was completely unusable in this footage. And the 120fps didn't look that great, but I would use 120 if I had no other option, key words being "no other option".