Tested the Atomos Ninja and FS100

The Ninja is working very well for us in PAL too. Progressive footage is sent and recorded as PsF, and some NLEs don't recognize it as such. Just tell the NLE it's progressive and you are set.
 
Personally I think it's the editing/workflow ease of recording to ProRes, that is the Ninja's greatest benefit. The broadcast compliance is the other big benefit.

The camera's AVCHD footage already looks fantastic in most scenarios, so I'm not sure how much better people expected it to become.
 
Barry_Green said:
Glad to hear that's sorted out.

I've been telling people for a while that AVCHD is surprisingly good, and that so many people who think that they "need" an external recorder...
Click to expand...

I was one of those, but I´m converted after doing some tests with the BM Hyperdeck Shuttle.
I would have sworn, that there MUST be a worthy difference between a highly compressed 4:2:0 codec and uncompressed 4:4:4 (or 4:2:2).

But even if you really torture Sonys AVCHD in color grading, there is - if any - only a very subtle difference in noise.
The bummer is, the AVCHD seems to have more organic noise.

I think this is a FS100 specialty. The codec adjusts himself to the db settings and other parameters.

So it looks like all we know about bit depth and color sampling doesn´t somehow apply here.
On paper and in theory, there should be an obvious difference between highly compressed 4:2:0 long GOP codec and uncompressed intra frame 4:4:4 (or 4:2:2).
For some reasons, I´m not able to reproduce that theory in reality.

Drives me nuts. I really want to understand, what´s going on here.

Frank
 
Please someone who owns ninja test green screen with moving subject, blurred out of focus areas, shadow on green screen, waving hair and transparent objects.

No one has posted any examples of this kind yet, so it would be very useful. If there is no difference, then Ninja is not worth anything...
 
Eimulis said:
Please someone who owns ninja test green screen with moving subject, blurred out of focus areas, shadow on green screen, waving hair and transparent objects.
Click to expand...

My good friend Rick Young and I fully intend to do a full-on test with Ninja and FS100 vs native AVCHD, plus an FS100 to F3 side by side (and compare a £5k and a £500 wide angle lens together, and the rest). All we have to do is synchronise diaries so all the kit comes together without affecting our day to day work. Some time between now and January. Of course we do our own tests (which are often so crap you wouldn't want to share with peers) and have great examples from the stuff we shoot but can't publish publicly.

We did this a while back with EX1s, combing dead cats and waving around glasses of squash, but this time we'll include colour markers for tracking hand held cameras and zooms, plus resolving 3d camera position, as our hypothesis is that this will be a better test for the advantages of 422 over native recording.

We're going to do this for our own benefit, we'll share, but all this comes way after actual paid work, family, pro bono stuff and administrivia. It's our playtime!
 
Barry_Green said:
really surprised that you didn't see more of a difference in your keying.
Click to expand...

Mr. Crook noted he used Adobe Ultra to key -- and it should be noted that Ultra (the newest version since CS5) is specifically designed for 4:2:0 footage. Adobe calls the new technology "vector keying" and said it excels with "marginal footage."

So, it's entirely possible that results from Ultra with 4:2:0 are going to be better than the same key run with some traditional keyer like Keylight or Mac whatever.

This note isn't intended to bash AVCHD -- I've been using it since the HMC150 and argued, back when it wasn't fashionable, that it was holding up just fine to DVCPROHD. Rather, I suspect that if someone runs a "test" with Keylight they're going to get different results than a test run with Ultra.
 
I think the tidbit about ultrakeyer the AGMedia mentions is interesting info. I actually recently had a post house almost completely reject some greenscreen footage I shot because they said the codec (avchd) looked awful, and until now I was stumped as to why they would say that because I was pulling a key with Premiere/Ultrakey off the same footage without any problems whatsoever. I pulled a key and sent them a screengrab and stumped them back. Now I guess I know why. But perhaps it's not just the 4:2:0...but the way it works with AVCHD too?
 
Richard Allen Crook said:
I think the tidbit about ultrakeyer the AGMedia mentions is interesting info. I actually recently had a post house almost completely reject some greenscreen footage I shot because they said the codec (avchd) looked awful, and until now I was stumped as to why they would say that because I was pulling a key with Premiere/Ultrakey off the same footage without any problems whatsoever. I pulled a key and sent them a screengrab and stumped them back. Now I guess I know why. But perhaps it's not just the 4:2:0...but the way it works with AVCHD too?
Click to expand...

Were you sending them MTS files? Premiere and MTS seems to be a recipe for disaster lately in terms of decoding properly on some FS100 and AF100 footage. I've verified with Premiere Pro and After Effects CS5 that native footage from both cameras was interpreted VERY wrong.
 
What do you mean with "very wrong" Rick. I never had a problem as long as I import via Media Browser.

Frank
 
Postmaster said:
What do you mean with "very wrong" Rick. I never had a problem as long as I import via Media Browser.

Frank
Click to expand...

On some clips, and I cannot figure out why, something is getting misinterpreted internally on CS5. It's treating the footage as though it is somehow aliased and ends up adding aliasing to the image by whatever method it is using to remove whatever it thinks it needs to do.

Both times I have seen this, it's been with footage straight from the camera, MTS, right into either Premiere or AE. A thread recently about AF100 greenscreen showed it happening again. I didn't pay close enough attention to if it was 24p or 30p footage only, but it was definitely progressive footage and off of either the AF100 or FS100.

The thing is, it doesn't always do it with every clip. It's consistent with a particular clip.

Back a few years ago, an employee of Adobe mentioned on one of their forums that in that particular case, AVCHD decoding wasn't being done well by CS4 I believe it was.

You'll know when you have it as you will see some HORRIBLE aliasing that you KNOW was not in that clip! :)
 
Rick Burnett said:
On some clips, and I cannot figure out why, something is getting misinterpreted internally on CS5. It's treating the footage as though it is somehow aliased and ends up adding aliasing to the image by whatever method it is using to remove whatever it thinks it needs to do.Both times I have seen this, it's been with footage straight from the camera, MTS, right into either Premiere or AE. A thread recently about AF100 greenscreen showed it happening again. I didn't pay close enough attention to if it was 24p or 30p footage only, but it was definitely progressive footage and off of either the AF100 or FS100.The thing is, it doesn't always do it with every clip. It's consistent with a particular clip.Back a few years ago, an employee of Adobe mentioned on one of their forums that in that particular case, AVCHD decoding wasn't being done well by CS4 I believe it was.You'll know when you have it as you will see some HORRIBLE aliasing that you KNOW was not in that clip! :)
Click to expand...
I misunderstood what you were saying. I thought you were saying that the post house i gave it to was misinterpreting...haha. But I haven't had issues with keying nor the aliasing probs you talk about with Premiere.I had cs5 and it seemed to have issues with avchd, but cs5.5 seemed to clear it up. In fact, isnt it in their update notes that 5.5 improved avchd?
 
Good question, not sure. I only have CS5 at the moment, I bought it so recently that I didn't want to pay for the update yet. In any case, something to keep an eye on if something strange pops up. Believe me, I was VERY surprised when I saw it. Given it was easy to reproduce (using the MTS file that did it from the camera) I've been scared of pulling files right in now.

What's odd, like I said, it hasn't been ALL files from the AF/FS. Maybe it's a setting or some meta data getting the software confused. Strange to say the least.
 
Looking back, the reason I thought they improved avchd performance in cs5.5 wasn't from the update notes, but from a thread I had in the Adobe forums. An adobe employee states in the thread that there have been improvements to how Premiere handles AVCHD in cs5.5. I had some problems with the codec too, though not the same ones you are having Rick, and the problems went away with the update. http://a5.sphotos.ak.fbcdn.net/hpho...18054864_1465877682_33010361_1611339906_n.jpg Is this problem apparent in exported clips? Or just the preview monitor?
 
Last edited:
Rick Burnett said:
Good question, not sure. I only have CS5 at the moment, I bought it so recently that I didn't want to pay for the update yet. In any case, something to keep an eye on if something strange pops up. Believe me, I was VERY surprised when I saw it. Given it was easy to reproduce (using the MTS file that did it from the camera) I've been scared of pulling files right in now.

What's odd, like I said, it hasn't been ALL files from the AF/FS. Maybe it's a setting or some meta data getting the software confused. Strange to say the least.
Click to expand...

I'm SO glad someone else is experiencing this, Rick. All last night I've been trying to solve this problem with no luck. It just wasn't making sense to me...is there a fix for it?
 
bkmvincent said:
I'm SO glad someone else is experiencing this, Rick. All last night I've been trying to solve this problem with no luck. It just wasn't making sense to me...is there a fix for it?
Click to expand...

My fix is to transcode to ProRes first before I get into any Adobe product.
 
You must log in or register to reply here.
Back
Top