Official: HDMI 1080 IS INTERLACED ONLY

I just saw the video below with the footage from the Ninja and internal codec. I don't know if everything is set properly, but the hdmi signal is very contrasty with clipped highlight and crushed shadows compared to the internal Codec. I hope Jeromy or others have an idea what is happening because I am getting a d800 and was thinking of getting a Ninja.

 
WOW! What a slam on the Ninja! UPDATE: Keep reading...

I guess it's time for a AJA Ki Pro Mini (with firmware 3.0) vs. Atomos Ninja (with firmware 2.2) vs. Sound Devices vs. BlackMagic Hyperdeck Shuttle 2 showdown.

Any others which belong on this list?
This sort of bake-off really has nothing to do with the cameras or camcorders, the comparison is between recorders and video quality and compression codecs. Perhaps this has already been done, and would belong in some other forum than HDSLRs. Perhaps in HDSLR - General / Other ?

UPDATE: As following posts in this thread make very clear, the fstoppers folks had incorrectly configured the D800, and the Ninja is blameless.

-- Bob Elkind
 
Last edited:
Brian@202020 said:
As soon as my D800 arrives I'll post footage from my Hyperdeck Shuttle 2.
Click to expand...

Hi Brien,

Can you tell me what codec the Shuttle 2 uses when saving uncompressed QT files. Just to clarify, is it 10bit 4:2:2 or some other flavor.

Steve.
 
Last edited:
eteam said:
WOW! What a slam on the Ninja!
Click to expand...

I may have prematurely yanked this chain. It seems quite likely that the D800 output is set for full (1-254) data coding range, whereas the Ninja is designed to understand 16 as black and 235 as peak white. Hence the clipping and crushing.

Consider the menu selection described on page 257 of the D800 User Manual.

Also: "Full" range is not a permitted option for 8-bit 4:2:2 YCbCr HDMI video signals. The HDMI specification requires respecting the "limited" coding range, which is entirely consistent with universally accepted SMPTE and ITU signal standards.

-- Bob Elkind
 
Stephenh said:
Hi Brien,

Can you tell me what codec the Shuttle 2 uses when saving uncompressed QT files. Just to clarify, is it 10bit 4:2:2 or some other flavor.

Steve.
Click to expand...

The Hyperdeck Shuttle 2 currently records:
10 bit uncompressed 422
10 bit DNxHD 422 (.mov and .MXF)


That said the D800 only outputs 8 bit uncompressed 422, so even tho the codec options are 10 bit, 2 bits will be zero's.

Also the SSD's I have are not fast enough to record uncompressed, so I'll only be able to test DNxHD.
 
eteam said:
Any others which belong on this list?
-- Bob Elkind
Click to expand...

Yeah, I would love to test my D800 with the Convergent Design nanoFlash. It uses Sony's 422 XDCAM codec at pretty high data rates. Which for me, editing in Premiere, is a better option than ProRes.
 
Brian@202020 said:
The Hyperdeck Shuttle 2 currently records:
10 bit uncompressed 422
10 bit DNxHD 422 (.mov and .MXF)


That said the D800 only outputs 8 bit uncompressed 422, so even tho the codec options are 10 bit, 2 bits will be zero's.

Also the SSD's I have are not fast enough to record uncompressed, so I'll only be able to test DNxHD.
Click to expand...

Thanks, I have read you need a high end SSD like the OCZ Vertex 3. Can't decide this or Ninja which looks like can't handle the full DR of the D800.
 
Brian@202020 said:
It can handle it if you change the settings in the camera like Bob said.
Click to expand...

Thanks Brian, I would be really interested to know how the DNxHD codec handles the wide DR capture of the D800. I am like you waiting on my D800 in about 3-4 weeks. One thing that Nikon Australia is offering is a 2 year warranty as standard. Are Nikon else where offering the same. I believe it is trying to stem the import of (grey market) units from Hong Kong/Asia area.
 
eteam said:
I may have prematurely yanked this chain. It seems quite likely that the D800 output is set for full (1-254) data coding range, whereas the Ninja is designed to understand 16 as black and 235 as peak white. Hence the clipping and crushing.
Click to expand...

The D800 mis-configuration has been confirmed by fstoppers in the vimeo comments. After correcting the output range setting and re-shooting, levels look consistent when comparing "clean" HDMI (via Ninja) to h.264 video recorded in the D800. An updated fstoppers video report is in process.

-- Bob Elkind
 
Haha, YES! My gizzard told me to trust the folks at Atomos above our f-stoppin friends. I'll take the D800 to NAB today and have a go with some external recorders. On another note, spent some time with a custom antialising filter for D800 video moire issues with very positive results:

fennworld.blogspot.com/2012/04/d800-redeemed-moire-aint-no-more-eh.html

Combined with clean HDMI out, this one'll be a winner!
 
Dave still hasn't announced a release date, but is working hard at it. Potentially a new design. I used the test filter he got me inside a plastic frame, supported only by pressure of the sides and mirror from the top. I think he might be thinking along those lines. No vignetting, sharp image edge to edge, tested from 16mm to 200mm. Very promising.
 
Pyriphlegethon said:
Dave still hasn't announced a release date, but is working hard at it. Potentially a new design. I used the test filter he got me inside a plastic frame, supported only by pressure of the sides and mirror from the top. I think he might be thinking along those lines. No vignetting, sharp image edge to edge, tested from 16mm to 200mm. Very promising.
Click to expand...

Do you think they will be able to keep up with the demand? Aliasing and Moiré are the only things holding this camera back. If they can solve that, the demand for the filter will be almost as high as the demand for the D800 itself, and Nikon is having trouble keeping up from my understanding.
 
Brian@202020 said:
Do you think they will be able to keep up with the demand? Aliasing and Moiré are the only things holding this camera back. If they can solve that, the demand for the filter will be almost as high as the demand for the D800 itself, and Nikon is having trouble keeping up from my understanding.
Click to expand...

If millions are spent for meeting anticipated demand, and the demand is slow to ramp up, this could sink an otherwise going concern and people's livelihoods.
The effectiveness of -- and market demand for -- the solution is by no means proven, yet.

-- Bob Elkind
 
Pyriphlegethon said:
Bob, just a quick verification. Did that [Black Magic card] EDID information prove useful?
Click to expand...

I finally got around to deconstructing the Black Magic EDID code. If anyone is interested, attached is the byte-by-byte disassembly. Some of the bit fields are only summarised, but all of the missing detail information is readily available online.

-- Bob Elkind
 

Attachments

You must log in or register to reply here.
Back
Top