Epic-W compression ratio

[Nicholas Carlton]

It's been said previously that 5:1 is the "sweet spot" for compression, however, the Epic-W only offers 7:1 compression @ 8K. Just curious if this increase in resolution and the smaller pixels mean we're getting a comparable end result (to 5:1 compression @ 6K)?

Is 7:1 compression @ 8K ideal for shooting green screen?

 

[Phil Holland]

"ideal" typically is the "lowest possible compression ratio".

However, since we're talking Helium, it records REDCODE RAW a bit more efficiently and at a higher quality compared to say Dragon due to the lower noise floor and a few other things. So 7:1 on Helium and 7:1 on Dragon isn't exaclty comparing apples to apples.

 

[Nicholas Carlton]

Thanks, Phil. That's what I thought!

Would you say 8K 7:1 is sufficient for chroma key work?

Also if you were mastering in 4K, would 8K (7:1) downscaled give you a comparable / better result than 6K (5:1) Dragon downscaled?

 

[David Schuurman]

Hey Phil,
do you happen to know the rough comparison of compression ratios would be? for instance, shooting 10:1 on helium is equivalent to shooting 8:1 on Dragon and 6:1 on M?

 

[Phil Holland]

I'll have to write something up. It's not exactly a cut and dry sort of thing because there's a variety of factors involved. Subject matter included really.

8K 7:1 on Helium should be enough for a key Nicholas. In the case of your comparison of scaling down Dragon 6K 5:1 versus Helium 8K 7:1, yes Helium will be higher quality "technically" due to the additional resolution in this case.

 

[Mark George]

Hey Nicholas, FYI with a 512 GB or 1 TB SSD, you can get to 8K @ 6:1 compression on your Epic-W. Those larger capacity cards are faster and allow for slightly better compression ratios.

 

[Noah Yuan-Vogel]

Frankly I think unless people are delivering in 8K now, they should probably rethink and revise their previous preconceptions about compression ratios. Theoretically higher compression ratios on 8K should be as good or better than lower ones they were used to in 4-6K when they result in the same or higher bitrate. Assuming the same output and viewing specs which is likely the case.

 

[Brendan H. Banks]

Frankly I think unless people are delivering in 8K now, they should probably rethink and revise their previous preconceptions about compression ratios. Theoretically higher compression ratios on 8K should be as good or better than lower ones they were used to in 4-6K when they result in the same or higher bitrate. Assuming the same output and viewing specs which is likely the case.

Finding this to be true. I shoot 7k and 8k @ 12:1 for a good number of projects. The quality is fantastic but if you find yourself pushing your ISO, I'd consider a step down in compression as well.

Also forgot to switch my compression for a few shots on a recent piece when i dropped down to 5/6k and boy was it noticeable. Very salvageable with some NR...but a real learning experience. Won't be making that mistake again.

 

[Cristian Bernal]

Finding this to be true. I shoot 7k and 8k @ 12:1 for a good number of projects. The quality is fantastic but if you find yourself pushing your ISO, I'd consider a step down in compression as well.

Also forgot to switch my compression for a few shots on a recent piece when i dropped down to 5/6k and boy was it noticeable. Very salvageable with some NR...but a real learning experience. Won't be making that mistake again.

Just to clarify, what was the mistake you made here? Forgot to switch back to a lower compression ration when going back up to 7K and 8K? you were shooting a higher compression at 5/6K?

 

[Brendan H. Banks]

Just to clarify, what was the mistake you made here? Forgot to switch back to a lower compression ration when going back up to 7K and 8K? you were shooting a higher compression at 5/6K?

I was shooting 12:1 at 5k (for higher FPS) when I should have dropped that down to like at least 8:1. So my 5k stuff is noticeably noiser. Note that I was using 1600 base ISO which is also a contributing factor, but you can see the compression noise because of the 12:1 when at 5k.

 

[Noah Yuan-Vogel]

Just to clarify, what was the mistake you made here? Forgot to switch back to a lower compression ration when going back up to 7K and 8K? you were shooting a higher compression at 5/6K?

I believe he was saying he left it at 12:1 when shooting 5k/6k and it was a problem vs 12:1 in 8k which looks good. Shooting 8K essentially doubles your pixelcount/bitrate versus 5-6k so it wouldn't be surprising if 12:1 6K looked more like 20:1 8K and 12:1 5k looked worse than 22:1 8k...

edit: Brendan beat me to it

 

[Nick Morrison]

I believe he was saying he left it at 12:1 when shooting 5k/6k and it was a problem vs 12:1 in 8k which looks good. Shooting 8K essentially doubles your pixelcount/bitrate versus 5-6k so it wouldn't be surprising if 12:1 6K looked more like 20:1 8K and 12:1 5k looked worse than 22:1 8k...

edit: Brendan beat me to it

Exactly.

My understanding is 8K is twice the data of 6K. So 8K 12:1 in theory should be same data as 6K 6:1....and I'm assuming has similar noise profile as well.

 

[Mike Finegan]

It is about information content. So 4K of information captured at 8K needs less detail preserved than 4K of information captured at 6K etc. to get same information content in the output 4K.
But there is also Nyquist and sampling - and since 8K perfectly meets nyquist criteria should have the cleanest 4K if interpolated.

Probably you could actually measure that if you had some texture test chart.

 

[Noah Yuan-Vogel]

Exactly.

My understanding is 8K is twice the data of 6K. So 8K 12:1 in theory should be same data as 6K 6:1....and I'm assuming has similar noise profile as well.

1.778 times the data to be exact

And 8K is 2.56 times the pixelcount/datarate of 5K.

So plan accordingly...

 

[Luca Volpe]

PURPLE???

PURPLE???

Very strange thing happen this morning...I imported an 8k R3D in FCPX and this is what I see...now I suppose because I used a 120GB mini mag? Or is just a corrupted file??