What compression are you using?

[James Brundige]

I'd appreciate some feedback on compression settings for Epic. Looking for the sweet spot between quality and data size.

For instance, has anyone compared 12:1 to 8:1. I don't have 4K monitoring, so my tests are not conclusive.

I attach a spreadsheet I've started to understand Redmag storage capacity, if that's helpful.

 

[Justin Kirchhoff]

Everyone claims they do anywhere between 5:1 to 8:1. That's what I shoot at, except for the high speed settings of course.

 

[Johnny Friday]

i'm 10:1 on most shoots...i've tested 7:1 up to 12:1 and clearly don't notice any difference, but for sake of keeping compression "in range" or not too far out of range...that's where i settled.

 

[joe hedge]

6:1 is as high as you can go with HDR, I believe, so I generally just leave it there for the shoot

 

[Demetri Zuev]

+1 for the question

I'm trying to decide which REDmags to get and a "sweet spot" compression is obviously a huge factor

 

[Zeb B]

Has anyone posted any test footage showing compression ratios from 6:1 up to 12:1?

 

[Martin Stevens]

Has anyone posted any test footage showing compression ratios from 6:1 up to 12:1?

+1 please

 

[Zeb B]

I'm sure that RED has conducted tests but can't find the post for it. Anyone have a link?

 

[Peter Strietmann]

Here is some relevant information from a post I made a couple of weeks ago...

http://www.reduser.net/forum/showthread.php?65382-Epic-compression-comparison&highlight=

 

[conrad gaunt]

My redcode testing on the r1 showed me that , except in forests .. redcode 36 is probably, mostly, just capturing noise floor and im a big believer in rc28, which had a pleasingly softer look.

I'm hoping this holds true for the epic (ie lower data rates are perfectly usable)

 

[James Brundige]

Here is some relevant information from a post I made a couple of weeks ago...

http://www.reduser.net/forum/showthread.php?65382-Epic-compression-comparison&highlight=

Thanks, just what I was looking for. To summarize what I see there: 12:1 - for the web, 10:1 is near Redcode 28 , and 8:1 like Redcode 36. Spiderman shot 5:1

Good reference points.

 

[James Brundige]

Any more opinions on this? Shall I start a poll?

 

[Jason Comparetto]

+1 +1 +1

 

[Zeb B]

Not so much a poll - would love some actual scientific test footage running up the compression ration from 3 - 12:1 on something like these sets:

1. Flesh tones in studio lighting
2. Flesh tones in low light (candle and mixed light night scene)
3. Harsh sunshine high dynamic range landscapes
4. Fine detail like wires, bricks, cloth patterns etc (zoomed in to see detail)
5. Test colors and patterns in controlled lighting

Anyone done something like this yet? Would love to see the point of diminishing returns on compression. Seems to be a bunch of opinions on this subject without much science.

 

[Peter Cote]

Not so much a poll - would love some actual scientific test footage running up the compression ration from 3 - 12:1 on something like these sets:

1. Flesh tones in studio lighting
2. Flesh tones in low light (candle and mixed light night scene)
3. Harsh sunshine high dynamic range landscapes
4. Fine detail like wires, bricks, cloth patterns etc (zoomed in to see detail)
5. Test colors and patterns in controlled lighting

Anyone done something like this yet? Would love to see the point of diminishing returns on compression. Seems to be a bunch of opinions on this subject without much science.

+1 to all of the above.

 

[James Brundige]

Bump this thread. Any new opinions?

 

[Shane Kelly]

I shot 9:1 on a harley davidson commercial last week. My other current project is more documentary in style and I shoot 11:1. I never really saw a lot of difference between redcode 28 and 36 on the redone. I'll do some tests someday soon but currently too busy shooting.

 

[Björn Benckert]

I think it's safe to say that if you do not shoot in harsh lighting conditions or is doing green screen / blue screen work you might as well stay on rather high compression rates such as 12:1 or so. If you shot critical chroma screen shots then you should go as low as it goes for the compression on those shots. The same goes for difficult to compress scenery like a crisp forest with leafs blowing in the wind with a slow camera movement or such that lets the eye catch the detail but gives the compression a hard time.

So I think it's ion a shot by shot basis. Also off course the final resolution matters. But for chroma screen work I would give as much data as possible no matter what. For fixed camera shots with little movement you can go a lot higher than 12:1...
.

 

[Tehben Dean]

There was another thread on this recently: http://www.reduser.net/forum/showthread.php?70981-4k-3-1-vs-6-1 where Kemalettin uploaded a couple shots at different compression settings. 3:1- 6:1- 9:1- 12:1- 15:1- 18:1

For his first shot: http://s3.amazonaws.com/Kemalettin/test1.zip
I noticed that the noise particle size grew slightly from 3:1 to 9:1 and the compression became noticeable in the shadows starting at 12:1. Viewing at full res/100%.
I bumped ISO up to 4000 and contrast slider all the way up for these grabs.

Screen grabs below.

3:1

6:1

9:1

12:1

15:1

18:1

 

[James Brundige]

I've just done some extreme pixel peeping on some test footage. I CAN tell the difference between 8:1 and 12:1, but you have to look really close. Think I'll split the diff.