Introduction to the Z90, NX80, and AX700

[Doug Jensen]

It is silly for me to post footage to show how awful something is. Anything can me made to look awful if someone wants to make a point or simply doesn't know what the hell they are doing. I am asking YOU to show me something that proves S-LOG on the Z90 is better than shooting on REC709. Choose any subject you want, any contrast conditions, anything at all, and show us once and for all that S-LOG is better than any PP you want to choose from 1 - 6, even though I don't think any of those PP's truly take advantage of what the camera can do . Prove it.

BTW, I never once said it was bad because it was 8-bit. Shoot 10-bit HD for all I care, it is the camera itself that performs poorly in the S-LOG mode. If you had tested it you would have seen it, just as I have. And if you haven't seen it, then I suggest that you have not taken the time to shoot A vs. B identical footage with a custom profile. I have done all those things with two different Z90's on multiple occassions -- even though you would prefer to call me a liar. This whole conversation is ridiculous. Have you noticed that Alister doesn't post on forums anymore? He and I have discussed his reasons why he has mostly stepped away, and I am starting to think he has the right idea. Why bother? Life is too short to argue about stuff with people who, as far as I know, has never posted a single frame of video so we can see his work and judge his mastery of the subject at hand.

[Cliff Totten]

OK,..so, I'll take some samples. I don't own a Z90, I have an NX80 which should be the same. Although I admittedly have not shot very much yet with my NX80 as I just got it a month ago, I have shot SLog-2 on the same sensor with the RX10-III a little bit. For small carry stuff now, I shoot 10bit VLog at 400Mbp/s on the GH5.

I'l shoot my backyard in the bright sun (possibly Thursday morning) in Slog-2 and Cine 1 or 2 and I'll upload the raw 30 second clips for folks to play with.

Maybe you are right Doug, maybe it WILL be "rubbish" or maybe not.

The problem here is "expectations"....and our definition of what "rubbish" actually is. (I told you about that DP I know that thinks ALL rolling shutter cameras are terrible and that he only shoots global shutter) I think we all know that it's not going to look like SLog-2 from a 14stop camera like an FS7, or FS700 or FS5..etc. I am, in no way trying to say that it's going to be "awesome" as I KNOW what these 1inch-type Exmor RS sensors can do and what they cant. I might even throw the FS700 12bit raw stream to 10bit ProRes HQ as a side by side baseline comparison. Just 30 seconds to keep downloads small.

I will promise you this. If the NX80 looks like "rubbish" (for a cheap $2,000 1 inch-type camera) than I will have the guts to admit this to you and anybody else. I'm NEVER afraid to admit that I'm wrong about anything.

So what do I expect? About a couple of stops more DR and less cloud clipping than Cine 1 or 2. I'm not expecting 14 stops, I'm only expecting 10-11 stops at very best. Not too bad for a small, cheap camera like this. That is my expectation.

Your expectations, as I interpret them, are that NONE of his will be possible with SLog-2 over Cine 1/2. SLog-2 will NOT give a greater dynamic range over Cine 1/2 that is "usable". This is how I read you Doug.

Ok. Well?.. the footage will speak for itself. If you are right, I will freely and happily admit this to you. You have my word.

I will post the direct files for anybody that wants to inspect the test for themselves.

Let's PLEASE go into this with a very NEUTRAL mind set. OK?

CT

P.s. Im not trying to say you are a liar. Thats not right. Im only trying to say you might not have given it a fair chance based on a possible pre-conceived notion. Thats all. :-)

[Cliff Totten]

Hey....I'm 100% open to whatever the files show. People can do their own grade and everybody will be their own judge.

I'm guessing +2 stops more in SLog-2??? Who knows,...maybe Doug is completely right and there is nothing there.

I'll bring it and do it on my lunch break tomorrow.

CT

[Doug Jensen]

Cliff, just so you know in advance, I won't be downloading your files. I have plenty of my own S-LOG vs. PP tests to work with and I don't need yours. Why don't YOU grade your own footage and then upload a split-screen comparison between S-LOG and PP1 (or any other non-S-LOG gamma) to show us all the difference? I'm not going to do your grading for you.

However, with that said, I would be more than happy to take a look at a finished split-screen video to see proof of your claims about the fantastic abilities of S-LOG on the Z90. Naturally, I will also accept anything from your NX80. I am really looking forward to seeing this great video that you have promised that backs up your claims.

BTW, I don't understand this obsession you have with dynamic range. I don't see how you can think DR trumps everything else and is all that matters. Nothing could be further from the truth. DR is far, far down on the list of things that make a beautiful image. If you really believe that DR is that important why don't you purchase a dynamic range test chart from DSC and shoot some charts. The camera that shows the most range is the winner!! No other test need be done if that is all you care about.

I wish I could post some of the video we've been shooting this week at the Olympics at the downhill racing events in bright sun and ice hockey to show you how (even in very high contrast conditions) dynamic range is not important. The footage looks absolutely gorgeous and blows away anything the Z90 could ever do even if S-LOG worked perfectly on that little camera. There is just so many more elements of a great picture that you don't seem to understand. I've got our 4K highlights team's three Z450's running on a custom scene file built around Hypergamma 4 and it just looks awesome right out of the cameras. No S-LOG needed. Amazing, really. Tune into NBC and watch some of the "live" coverage from any of the events anywhere at the Olympics. It's all REC709 and almost all HD (not even 4K). There's no log or grading on any of it. Think your NX80 looks like that?

Here's some Z450 footage I shot last year after NAB with regular REC709 settings. Do you think your NX80 (with or without S-LOG) can look this good?

Sony PXW-Z450 4K Demo Reel on Vimeo

[Donald McPherson]

Admins. Will you stop this tit for tat before we lose one or both of our members?

[Ricky Sharp]

Well renting these videos was a wise choice as now I'll ultimately save cash. I'm going to stick with the X70 for a few more years. While the AF improvements would be welcome, there's still too much compromise for 4K for my liking.

Once 10-bit 4:2:2 4K shows up at around 3 to 3.5K USD, I'll move.

Side note... I'm currently trying out your WSYWIG recommended profile Doug, but on the X70. Matching things up the best I could. I think shutting off the auto knee and the black level changes made the most difference. Previously, I was just using a very slight modification to PP4.

[Doug Jensen]

Thanks, Ricky. I'd be interested to hear how well my profile works on the X70 after you've had time to use it more.

[Lou Bruno]

Doug, where are the profiles located?

If included in your Vimeo tutorial, I am not up to that chapter yet. Nice job by the way.

[Doug Jensen]

Hi Lou, chapter 14 is the one you want for profiles. I'm glad you're liking the video so far. Thanks.

[Lou Bruno]

Use Cine 1 as there is NO Cine 4 in the X70. I was able to match my AX-700 and X-70 very well. I would recommend +4 for saturation on the X-70, but that is a matter of personal preference. I kept +2 on the AX-700.

Doug's PP is SPOT ON!!

Originally Posted by Doug Jensen

Thanks, Ricky. I'd be interested to hear how well my profile works on the X70 after you've had time to use it more.

[Cliff Totten]

So it looks like some of our lively debate was deleted. So, I'll just type an easy and neutral "admin friendly" comment because the other stuff must have been too strong. (Sorry if that was the case)

Dynamic range - I believe the NX80 puts out about 10 stops, not the 14 that was stated. I believe that SLog-2 does capture more DR than Cine 1 or 2. I think the files show this with no trouble.

As promised, I'm posting 3 files I shot today. These are only 20 seconds each and are straight off the NX80 card. It's a tough scene with bright sky and dark trees. I exposed all three to the right and took the clouds up high but nothing is clipped. It's a "top to bottom" test just to see everything this sensor can put out.

Without even grading it, it's easy to see that SLog-2 does capture more DR than Cine 1 and 2. Dark trees show detail in the bark and shadowy grass stands out easily. You can choose to expose and grade this any way you want. When shooting, with that extra latitude, you can expose in two ways. You can protect more whites and blow them less or expose to the right and dig deeper down into the shadows. So it will give you a little more in post to work with. (Not allot but a little more) No, I never said this was an Arri Alexa, FS7, F5 or RED or anything of the sort. I know the dynamic range limitations of this sensor full well and all it's other limits too.

In my opinion, avoid SLog-3 like the plague. It's way, way too much for this sensor to handle and I'm surprised that SLog-3 was actually given to this camera. You cant drop a 10 stop sensor output into a 16 stop curve. This sensor just cant cover enough distance to make Slog-3 into anything usable. But definitely try this for yourself. A 10stop sensor needs a 10stop log curve. This is why SLog-2 can be used, because it's bent to match each sensor's full output and nothing more. There are no wasted stops.

Importance of dynamic range vs every other aspect of shooting - I never said dynamic range was more important than any or every other aspect of shooting. Obviously, a persons ability pick his subject, to frame, focus and expose well and tell his story is the ultimate goal of any shooter. DR is simply ONE thing on a list of many, many, many things. I never said this was not the case. I have never believed DR was some kind of be all, end all for one second of my life.

I never said or suggested that the NX80 was any spectacular or exceptional performer in any way. I simply said it was very good for what it was built for and it's low price. No, I do not believe this should replace anybody's super35 camera and I do not believe that it should be used for television broadcast. (Another idea that came up that doesn't make sense to me where it came from) I believe it is a very nice little camera that does a lot for it's price. I believe that if you own this camera,...and maybe if it's the only one you own,..you can use it for certain situations in SLog-2 and it can give you more post options if you are willing to do the work on it. (The usual, color grade and noise reduction etc...)

I certainly never said this camera should replace a traditional shoulder ENG camera for broadcast. I also certainly never said SLog-2 should be used for live feeds either. Again,...I'm not sure why this was ever brought into question.

Anyhoo,...here are the three 20 second clips and one of me just describing what I'm doing. This is intended for anybody that is curious to download and test on their own and judge for themselves. If you are planning on buying an NX80/Z90, you can see these untouched files and inspect. Remember, these are all exposed for a maximum dynamic range test so they are on the hot side of the scopes.

I'm off to Walt Disney World for the 3 day weekend and I'll play with it on Tuesday.

No,..I'm not expecting anybody to "grade this for me". Again, really confused as to where that sentiment came from.

For anybody that is curious, check it out and see for yourself and see how this all works for you. (for good or bad)

https://drive.google.com/open?id=1RP...z8dtLiBRg1K0en

Respectfully,

CT

[Doug Jensen]

I will sum up my thoughts with this: In all my testing with the Z90, I have never seen S-LOG provide superior results (even after grading in Resolve for standard REC709 delivery) to what can be done with a custom Picture Profile. I'm pretty certain that I'll never shoot S-LOG with the Z90 on a real shoot. Your mileage may vary.

[Craig Seeman]

Just a quick look 1t Cliff's clips as I see them in FCPX 10.4 with scopes and the S-Log 2 clip quickly graded in CoreMelt Chromatic

[Cliff Totten]

I'm sure most of us know this but for anybody that doesn't, on Cine 1 vs Cine 2 and SLog-2 on the scopes:

SLog-2 actually allows for 109 IRE. It's not broadcast legal. (there is no reason for that anyway)

Let me see if I get this right from memory... 8bit values for broadcast are 0 to 236? This takes you to 100IRE. However, 0-255 is 109 IRE and that is the full 8bit sampling range.

Technically, 8bit depth capability is 0-255 but when we shoot under boradcast legal, we slice anything over 100....so we are using a high of 236 instead of 255 so that's "7.5" bits instead of the full 8 bit ? Hehe.

Cine1 also goes outside broadcast legal to and Cine2 stays inside at 100 IRE clipping.

Did I get that right? ;-)

[Craig Seeman]

I asked Alister Chapman about the use of S-Log 2 on the Z90 in UHD and his response:

"I used S-Log2 because in UHD you only have a very limited data range to play with and S-Log3 only uses 80% of the already limited range while S-Log2 uses 95%. I am happy with the results having graded it in Resolve but there isn’t quite the same dynamic range you can get from a camera like the FS5 or FS7."

Alister's review in which he mentioned,
"I also shot at a couple of locations using S-Log2 to test how that worked (I was shooting in UHD and the camera is 8 bit in UHD. For 8 bit I prefer S-Log2 over S-Log3)."
The Sony PXW-Z90 ? a compact 4K camcorder with auto focus at it's best! | XDCAM-USER.COM