Comparisons: S-log, different recording formats

[Steve Kimmel]

I made a mistake so this is a repost.

I purchased the Vortex Media external recorders set of sequences and would like to share a few thoughts on my findings. I cleared this with Doug Jensen, but see item #1 below.

A few caveats first:

1) This is an outstanding set of files and I would encourage anyone who is trying to decide on external recorders for their F3 to consider purchasing them. There is no way I could have done all of this work myself (let alone get access to all of the recorders). I can’t say enough about the quality and thoroughness of these tests.

2) These are all just my impressions based on what I see at full resolution and based on a lot of playing with the files.

3) With 126 different shots, there are literally thousands of comparisons one could make (over 7,000 pairs and over 10 million groups of 4). So, I have limited it to a few. I’m not going to be able to do more as I’ve already spent a ton of time on this and I have to get out and do some shooting! (See #1 above)

4) I’m only reviewing the female model files. There are a bunch of others scenes as well, but my main interest is skin tones so I didn’t try to sort through the green screen ones, nor did I do the ChromaDuMonde or table top ones (the latter and the green screens include fine, fast motion but, to tell you the truth, I’m not very good at seeing subtleties in this). (See #1 above)

6) Uncompressed files: there are Gemini uncompressed files in the set but I haven’t even tried to view them. The files are huge and I have almost no disc space left. I am also not looking for uncompressed recorders.

6) These are really tests of recording formats, not the recorders themselves. For each recorder, different formats are used. As Vortex explains it: “in theory, the same settings on two different recorders should provide the exact same results.” They note that future tests may test this theory.

Here are the files types by recorder (there are even more that you can try – see #1 above)

SxS: 35 mbps, 4:2:0, 8 bit, XDCAM EX
Nanoflash: 140 mbps, 4:2:2, 8-bit, XDCAM HD422 (I-frame)
KiPro Mini: 117 mbps, 4:2:2, 10-bit, ProRes 422
Pix 240: 176 mbps, 4:2:2, 10-bit, ProRes HQ
Cinedeck ProRes: 264 mbps, 4:4:4, 10-bit, ProRes
Cinedeck Cineform: 256 mbps, 4:4:4, 10-bit, Cineform

With all that, here goes. I’ll split this up into a few posts.

1) Default picture setting, no SLog.
My impressions: Compared with SxS, the recorders do a better job of capturing tonality overall. The Nano has perhaps the blockiest shadows, followed by the Pix and KiPro (tied) and then the Cinedeck (I personally prefer the Cineform files by a bit).

Next up will be review of S-Log files

[Steve Kimmel]

OK, so I did something wrong. Attached files but they're not showing up (only links to files). What did I do wrong?

[Nate Weaver]

You're putting Doug's work up for free?

[Steve Kimmel]

I made a mistake -- I deleted attachments. Didn't realize they would go up like that!

Thanks.

[Steve Kimmel]

Next installment:

2.) S-log. I applied a standard color correction from Magic Bullet Looks (mostly just lift/gain/gamma and some saturation). First, S-log to my eye makes a huge difference, even when recorded to native 420 SxS cards. With that said, you do gain with external recorders. The Cinedeck files to me look the best – far better skin tones, definition in the blacks. In motion, you can see a lot of noise in the yellow square and some in blue in all except the Nano (maybe, as explained by Alister Chapman, this is really due to less detail in the Nano rather than less noise?) and the Cinedeck. The Cindeck files look a bit flat with this standard correction. Seems like it takes a lot more work to get them to have more pop (anyone know what that might be?)

[Alister Chapman]

Ah, the question of noise in codecs again! I am wondering if the Sony Mpeg encoder chip as used by Convergent Design in the NanoFlash and used in the EX1 is using a degree of noise reduction that is always present even at the much higher bit rates (100Mb/s +). It would need some more sophisticated test equipment than I own to look into this properly. Noise reduction is an important factor in encoder design.

[Steve Kimmel]

Originally Posted by Alister Chapman

Ah, the question of noise in codecs again! I am wondering if the Sony Mpeg encoder chip as used by Convergent Design in the NanoFlash and used in the EX1 is using a degree of noise reduction that is always present even at the much higher bit rates (100Mb/s +). It would need some more sophisticated test equipment than I own to look into this properly. Noise reduction is an important factor in encoder design.

Thanks Alister. I'd be curious in anything else you can find out about this.