Atomos Ninja value when used with FS-100?

R

Ryan Lightbourn

Well-known member
*note* - I'm not saying the Atomos Ninja is useless...this strictly pertains to use with the FS-100

I've been doing extensive testing all day with the Atomos Ninja & FS-100. All of my side-by-side comparisons are showing almost no difference whatsoever between the ProRes (HQ) on the Ninja & the FS-100's native AVCHD. The AVCHD actually may look a little better.

I've tried various frame rates, exposures, resolutions & settings, but it makes no difference.

I've attached an example (where I purposely underexposed the shot, then color corrected to bring out detail):

http://vimeo.com/26183207
password: dvxuser

Here is the original file:

Right click to save file:
http://ryanlightbourn.com/uploads/dvxuser/ninja_test.mov

P.S. The FS-100 is a f-ing killer cam.
 
Last edited:
I was debating about this very thing. Some colleagues told me I should get an external recorder, which my contention is that it would be a complete waste of time/money/effort. I mean, the HDMI is uncompressed which is great...but with an external recorder we're just recording to ANOTHER compressed codec (like Prores on the Ninja). Doesn't make a lick of sense to me. I guess people can get into the whole "ProRes is better than AVCHD" and vice versa...but meh compressed is compressed, right?

Great test...glad you did it. Now I won't worry too much about going after an external recorder! :)

EDIT: Then there's this test, discussed briefly on another thread. http://www.dvxuser.com/V6/showthread.php?251432-AVCHD-in-cam-vs.-Prores-HQ-capture-via-HDMI

You know if you REALLY wanted to see which is better, try to shoot in the flattest, decontrasted picture profile possible with a smooth gradient lighting like a wall or something...then push the contrast waaaaaay up in post and see which holds out better. In the comparisons I've seen so far the footage is way too flat.
 
Last edited:
have you done any extreme color correction to test the difference? try to shoot a scene with a wrong white balance and then correct it in post and see which one holds better
or start throwing in some extreme RGB curves and see how it holds up?
 
speedracerlo said:
have you done any extreme color correction to test the difference? try to shoot a scene with a wrong white balance and then correct it in post and see which one holds better
or start throwing in some extreme RGB curves and see how it holds up?
Click to expand...

I did mess with saturation levels a little bit & saw similar results to the over/underexposure tests (negligible differences), but haven't tested white balance corrections. I assume that will show the same results and if not; that one plus wouldn't be worth the setup hassles & inability to play back footage.

I'm sending the Atomos back to B&H in exchange for another Zeiss prime!
 
Where you will see the codec's impact clearly is when you apply secondaries when gradeing.. and yes i can see a significant diffrence between ProRez@220Mbs and h.264@25mbs when using color to define regions to grade

Your mileage may vary, i'm waiting for the right recorder to appear.. mainly one that grabs 444@ 23.98, and that will likely be a ProRez444 codec used.

d
 
Have you ever tried to chroma-key with an MPEG-4 AVC/H.264 4:2:0 chroma sampling format like the AVCHD recording format of this NXCAM camera?

I bet you'll love doing so on a better codec like Prores 422 HQ.
 
I thought another purpose of using a recorder was to ensure artifact free footage when you stress the codec. A lot of motion, both camera and subject. in an environment with a lot of detail. Shooting a still shot of a motionless dog in a low detail environment isn't really stressing the codec.
 
I would suggest performing the test again with motion. That tends be where artifacts appear, and I would be very interested in seeing the results.
 
I recorded uncompressed (RGB out) to my Blackmagic Decklink. 50i 1/50 shutter. At the same time I recorded to the card.
You see almost no difference untill you magnify to 200%.

But when you apply secondarys, that´s where it makes a difference. Tonns of room in the uncompressed material, not so much in AVCHD.

 
Well, no external recorder is going to give you a magic camera. But it will give you this. Workflow, no transcoding, 422 color over 420, on some recorders instant backups (camera card and recorder). All those little pixels, with motion, why not record in a codec you're going to edit with? So it's not useless. Do you HAVE to have it? No. Can it be advantageous? Yes. But it won't make your magic.
It's nice to pop a card or hard drive in and just start editing.
 
I've owned a Nanoflash (4:2:2, 8-bit, Mpeg2)since they were released a couple of years ago. During that time, I've used it, primarily, with my EX1. Detailed comparison testing between the SxS card captures and the Nanoflash captures have revealed some interesting results. I believe the initial conclusion that there is an insignificant difference in captured video streams is visually correct. Unless projected on a very large screen, the casual differences are unnoticeable. Having said this, I will agree that the real advantages don't become apparent until chroma-keying or color grading. Comparison of footage with high motion in the frame also has some very interesting results. At the lower bitrates(100Mbps or less) it's arguable that the Nanoflash captures are better than the SxS captures. However, at the higher bitrates, and especially at 220Mbps, the hi motion areas show much less blocking and artifacts than the SxS captures. I believe that the compression algorithms are very stressed by hi motion content and that by increasing the bitrate, the compression algorithmns have more data to work with and less artifacting.

I've used the Nanoflash with a Sony NEX-VG10 and a Canon 7D. In both cases, I've found the Nanoflash provided no gain, whatsoever. In these cases, the video signal as fed to the HDMI output port, has significant moire that further stresses the mpeg2 codec in the nanoflash. The AVCHD native codec in these devices seems to handle the line skipping artifacts much better than the mpeg codec in the Nanoflash.

I just received my new Blackmagic Hyperdeck Shuttle(Quicktime uncompressed, 4:2:2 10 bit) and am eager to test it out on my EX1, which supposedly has native 10-bit out of the HDSDI port.
 
Last edited:
It seems several of you have seen better results in pushing the external recording grading over the AVCHD. This is what I'm most interested in seeing...like an example comparing the two after an extreme grade to see just how much better the external recording gets handled. Anyone want to do this and post up the results? :)

Even better, post a 5 sec clip of AVCHD AND the external recording so we can grade ourselves and examine more closely. ;)
 
mico said:
Here, take a look a the difference. Its huge at least with the nanoflash. Detail and motion.http://convergent.contentpros.net/Products/nanoFlash/Quality/tabid/1651/Default.aspx
Click to expand...
Well, this is a marketing page which I take with a huge grain of salt, besides its a different camera. I understand what the external recorders are marketed to be able to do...would still love for someone to post examples with the fs100. I'm not sold that its a significant enough difference for me to purchase an external recorder. Man, I wish it was 10 bit HDMI! :)
 
My purpose isn't to push someones product but to show the situations where you might see the benefits if you want to post a test. Since you can dual record with the cam it will be easy to see the differences if you shoot something with motion and detail like in those shots. The fact that the BBC requires higher bit rates and recommend using a recorder with these low bit cams cams and they have no financial interest in recorders there must be a reason why so many people are using them that are visible.
 
AVCHD is really quite good and, in a non-challenging environment, it isn't surprising that you can't see much (if any) difference. We've seen the same situation with the AF100 when comparing it to the Nano and even AVC-Intra. This is the reason I tell folks not to worry about an external recorder at first. Try out the native codec, you may be pleasantly surprised and find it's plenty adequate. However, as said
 
In a prior post, there are definitely areas where an external can exceedthe internal. For keying, the true 4:2:2 is a big advantage. And for some super high detailed scenes the external codec will be more resilient. Try a daylight forest shot with a waterfall, handheld, and you'll see a noticeable difference.

That's why the BBC requires it - 4:2:2 and immunity from codec degradation.

However, in general and for most purposes, I think most folks will have a hard time justifying the cost of any external recorder, over just using the internal AVCHD. Definitely for those for whom money is tight, they should carefully consider whether a recorder should take preference over, say, lights or a good mic kit or something else that may have a bigger impact to their picture.

Except for F3 owners. For them an external gives too much advantage to ignore and the cost is such a small amount compared to the total system.
 
mico said:
My purpose isn't to push someones product but to show the situations where you might see the benefits if you want to post a test. Since you can dual record with the cam it will be easy to see the differences if you shoot something with motion and detail like in those shots. The fact that the BBC requires higher bit rates and recommend using a recorder with these low bit cams cams and they have no financial interest in recorders there must be a reason why so many people are using them that are visible.
Click to expand...
Oh no...I didn't take it as you pushing it, haha. Like I said, I want the external recorders to hold up better, but so far the difference is minimal in terms of holding up in post. But yeah as far as other non-image quality benefits like higher bitrates and ease of use in FCP then yeah, they're great.Another point I'd say is the AVCHD codec compression is VERY efficient, and is surprising to many as to how well it performs. It proves that it's not all about bitrates.
 
I have to disagree. I have done some testing and I can see a difference in the codecs when fast motion comes into play. However, as someone said, when keying for green screen or for secondaries in grading, that's where you see a huge difference. Have a look here http://www.dvxuser.com/V6/showthread.php?253306-AVC-rears-it-s-ugly-head.&p=2369161#post2369161

Other than not seeing enough difference between avc and prores, how did you like the functionality of the ninja? Does it seem like it's built solid and is worthy of real world use? Did you have any experience converting 1080i back to 24p via pulldown? If so, was it a pain or pretty straight forward?
 
cheezweezl said:
Other than not seeing enough difference between avc and prores, how did you like the functionality of the ninja? Does it seem like it's built solid and is worthy of real world use? Did you have any experience converting 1080i back to 24p via pulldown? If so, was it a pain or pretty straight forward?
Click to expand...

I checked out your post & I'll try the same experiment with the Ninja.

The package is great for the price. The functionality & setup (aside from the lack of footage playback) are both very straightforward. It comes w/ two batteries, a hard case, a USB/Firewire caddy, and a few other things. I've seen some AVCHD Vs. Hard Disk/Nano Flash tests that have shown great results...but this one just isn't doing it for me. I'm not sure if it's the FS-100 HDMI out or the Ninja itself.

The pulldown is fairly easy as well....I just run it through Compressor with a 24p reverse telecine ProRes preset & it kills the interlacing.

dustylense said:
Well, no external recorder is going to give you a magic camera. But it will give you this. Workflow, no transcoding, 422 color over 420, on some recorders instant backups (camera card and recorder). All those little pixels, with motion, why not record in a codec you're going to edit with? So it's not useless. Do you HAVE to have it? No. Can it be advantageous? Yes. But it won't make your magic.
It's nice to pop a card or hard drive in and just start editing.
Click to expand...

Unfortunately this is not true if you plan on shooting 24p...it still has to be transcoded with reverse telecine to remove the interlacing.

***edit*** I'll try some handheld tests today & see what happens. The thing is...I've seen locked tripod comparisons of AVCHD Vs. Uncompressed footage where you can clearly see a difference (especially with blocky digital noise). This is definitely not the case with the FS100/Ninja combo.
 
Last edited:
You must log in or register to reply here.
Back
Top