Ki Pro vs Nanoflash

[Mike Marriage]

Originally Posted by Dan Keaton

Dear Mike,
To be clear, no nanoFlash footage appears to have been used in your testing,

Yes, sorry Dan, I should have made that clearer. I actually did the test a few weeks ago for a different purpose but thought it was relevant here too.

Like I said, I'd like to do the same with the Nano @ 50 and 100Mbps vs Prores if you have a test unit in the UK? I played around with extreme grading on that test but I'd like to experiment more to try and establish any difference between the 8 bit and 10 bit.

[Dan Keaton]

Dear Mike,

Someone near you may have a nanoFlash.

We may be able to arrange a unit for testing.

Please contact Symbiosis, our European distributor.

Symbiosis | Authorised Apple, Adobe Training | Find Us

[Steve Kalle]

Hey Mike,

In your test, I opened both in Photoshop CS5 and expanded to 400% using the progressive images. For image quality, I noticed only slightly better detail in some areas such as the wind muff. However, I noticed what appears to be noise in the Pro Res image and there seems to be a fair amount of it. Another reason I believe there is a fair amount of noise is because I used the Noise Reduction filter in PS and was able to remove most of it which made both images (35Mb & Pro Res HQ) look 99% identical at 400%.

I predict that 50Mb XDCAM will be equivalent to Pro Res HQ but the Pro Res will have more noise.

I believe that a test with the EX1/EX1r/EX3 will show a better image with its native 35Mb versus Pro Res due to the cameras having more noise than the 350.

Many people much smarter than me, including Alister Chapman, have declared that recording 10bit from the EX cameras has no benefit, and wastes roughly 20% of recorded bandwidth.

Don't get me wrong, I really really like the price of the Aja and its 10bit recording, but many including me have been under the impression that recording 10bit automatically makes the recorded image better - but it doesn't with noisier cameras including my EX1s.

For me, I want both but neither support hot-swap and that is a deal breaker for me as I shoot many events. Furthermore, because the Aja has less than 25% of the recording length per card, swapping cards will happen much more often, which makes hot-swapping even more important.

However, if the Aja can record the embedded SDI audio AND the 2 XLR inputs for 4 separate channels, that alone makes it much much more appealing to me as I can avoid using my Mackie mixer in addition to recording 4 channels separately (my Mackie 1220i has 12 inputs but outputs to only 2 XLR).

One last thought: the Aja is good news because it finally introduces competition to the nanoFlash, and what does competition do....it forces companies to build better products. Now, maybe the nanoFlash 2 might arrive sooner than later.

On a side note, when will an AVC-Intra portable recorder arrive? To me, this would be the ultimate recorder as 100Mb AVC-I 10bit should be better quality than 100Mb MPEG2 and Pro Res HQ (220Mb). However, if it required P2, then I wouldn't even consider buying it.

[Rafael Amador]

Hi Steve,
Recording 10b gives us 1024 levels of bright instead of the 256 that offers 8b recording.
Aside of this, and with all my respects, this test has not any validity.
You can not judge 10 YUV stuff through 8b RGB screen captures.
About the AVC-Intra: Just another mistake of codec.
This will be really eaten by any Prores recorder.
rafael

[Peter Moretti]

I've talked w/ one of the designers of Sony's F35 who has done extensive codec testing. His personal feeling is that the codecs are the "low quality pass filters" of digital video.

He was decidedly unimpressed with 10-bit ProRes 4:2:2, and felt it did not compare to the other ten bit codecs like HDCAM-SR, Cineform or DNxHD. "I know it's a ten bit codec, but I find it difficult to think of it as such." Now the ProRes 4:4:4:4 codec is another story.

[Steve Phillipps]

I don't know about the other but SR is double or quadruple the data rate (440 or 880 mb/s) of even ProRes HQ, so it's not really a fair test. Maybe it's fairer to compare it to AVC Intra 100.
Steve

[Dan Keaton]

Dear Friends,

Rafael is absolutely correct when comparing 10-Bit (1024 levels) to 8-Bit (256 levels).

But, this is per color.

8-Bit offers around 16.7 million colors.

10-Bit offers around a billion colors.

I mention this because sometimes the impression is that 8-Bit has only 256 colors.

[Rafael Amador]

That's right dan, but our eyes are much more sensible to the changes in bright than in color.
The best of a 10b codec is the 10 Luma. That gives real definition. For the Chroma 8b are enough.
With all the possible shortcomings, Prores is clearly an step forward.
I don't care if there are better codecs are not (will always bee), but that 10b acquisition is at the reach of every pocket. I'm working on this since 27 years and for me this is a dream comes true.
rafael

[Paul Cronin]

Just to put in my 2 cents.

I own the Nano and it has not failed me in some very tough shooting conditions. Also I have found Dan and Mike to be first class when I have been in need of support.

But with all of my testing I find XDCAM 422 100Mb/s about the same as I-Frame 220Mb/s. But either way at 50Mb/s XDCAM 422 I have never had a customer even flinch. This is a broadcast standard with a lot of my clients and I do not find a need to go higher, unless we are shooting VERY fast motion close.

Would it be nice to shoot 10 bit, Yes. But is it needed at this time, No not for me and I bet most of the people on this forum. But should we not check out the Ki Pro, No by all means we should check it out and stay informed with the fast changing industry. But be careful to just chase new Tech gear unless you have the funds and are not concerned with the bottom line. For me for now the Nano helps me make money and that keeps it in the work bag ready to go.

I think the lower bit rate 420 cameras need to catch the Nano not the other way around.

[Steve Kalle]

I'll talk to my guy at my local Abel Cine and see when he gets the Aja in so we can do a Nano vs Aja test with their EX1r and 350. I assume it should be easy since I can take SDI into Nano and its output into Aja.

[Tim Kolb]

Originally Posted by Steve Kalle

I'll talk to my guy at my local Abel Cine and see when he gets the Aja in so we can do a Nano vs Aja test with their EX1r and 350. I assume it should be easy since I can take SDI into Nano and its output into Aja.

Well, first I'd wonder about the loopthrough on the Nano... With the internal structure of the device, will it pass full 10 bit as output from the camera?

It would best be tested with a clean HDSDI DA, with neither handling the signal in line before the other.

Some of this has been discussed and done to death in my opinion. Neither manufacturer seems anxious to try to dump on the other...I talk to, and work with both companies and have worked with both devices and still believe that they each have a sweet spot.

As far as some of the comparisons that are flying around regarding codecs, I'm in favor of embargoing comments on image quality for codecs you have not worked with personally. Some of the opinions being repeated here are so broad and vague that they serve no purpose.

These passing remarks about "better" or "worse" don't note whether it's in context of acquisition (and for that matter, what camera? Does anyone here believe that the lastword on HDcamSR quality can be determined by piping an EX1 into it and comparing the image to the internal SxS?), or editing, effects, color correction, archive... Runnability with a certain NLE? Transportability in a mixed OS environment?

There are so many factors to consider... Run time in the field, file compatibility with your chosen workflow (which stretches wider than just the NLE app these days), type of device it's packaged in for field use and the datarate for storage and post...and time investment in conversion along the way if it becomes necessary.

Not to mention asking about simpy how one plans to display the images and the various biases different devices can have...as well as who will be doing the critical viewing and what they're focused on... Somone with a film background and someone with a television background can look at the exact same characteristic and declare it "good" or "bad". I've witnessed it happen. Even restricting the test to two cameras from a single manufacturer as sources limits the actual information one can gain from such a test.

I'd add an F23 to the test to test a really clean signal...and I'd use at least four different high end monitors to view the result...and make sure you have at least one 10 bit LCD in there and you know you are actually feeding 10 bit all the way to the display.

...but overall I'd still say that this topic has been more or less beaten to death.

[Rafael Amador]

Hi Tom,
I agree with you very much.
There is no point to make a general comparisons between the NANO with the Mini Ki-Pro.
There are tools with two different purposes: one to record in 8b 422 and the other to record in 10n 422.
Is to the user to know if its needs one or the other other.
I've been recording Prores from my EX-1 through my AJA ioHD, for ore than two years.
Is great.
I use my NANO to record on the field. Is SUPER.
I've been very excited with the arrival of the Ki-pro, because I think is a huge leap forward for all of us.
But I don't gonna run to buy one. The NANO gives me what I need. Is perfect for my job.
The day I will need the extra quality that 10b can add, I will consider to get one.
Best,
rafael