D800

  • Thread starter Thread starter Guest
  • Start date Start date
t.p. said:
if someone here owns the two (D4 and the D800 ), PLEASE PLEASE , im begging you do a proper side by side comparison so we can stop guessing how they perform
Click to expand...


The video quality between the D4 & D800 is important to my day job, so because there is little information out there right now, I've decided to pick up both bodies (will be getting them this evening). I have a few ideas of side-by-side tests that I want to perform, but if there are specific things that people are wanting to see, now's the time to let me know. However, a few caveats:

1) Once I run my video tests, I plan to sell the camera body I don't want. This means that I want to keep both bodies as pristine as possible, and will not be taking them to the city, to the beaches, to the mountains, etc. I'll be shooting in my house and the surrounding neighborhood only.
2) I do not own an external video recording device, so both will be recorded onto the internal CF cards.
3) I will not be taking any still photos, as I want to keep the shutter actuations on both cameras at 0 until I decide which I want to keep.
4) I plan to sell the body I don't want starting next Monday, so do not have time for extensive and drawnout tests.

Bottom line, I'm taking a risk in buying both cameras to test, but I feel the video quality question is one too important not to chance the risk (for my own business). I want to get in and get out, do quick video tests in full-frame mode at 1080p/24, do a few tests that you all might want to see, and then quickly sell the body I don't want to keep in order to get it off the ole credit card.

Please send me a PM with requests, and I'll do my best to accommodate them.
 
Last edited:
I've started work on a version of Flaat designed for Nikon DSLRs, to see if I can get all those lovely 14 stops of DR to shine in video mode too
for that, I need to get a bunch of very specific (but very easy to shoot) RAW stills taken with the D800 (starting with a D7000 would work too)
if you have a D800 or a D7000, are interested in this, and have some time to spare, please write to similaar.feedback at gmail.com
 
@Geronimo:

That reply from NPS is very interesting. I have noticed that Nikons advertising campaigns have pushed the D800 more as a multimedia camera than they have with the D4.
 
Last edited:
bufo55 said:
WOW! A clear and unequivocal answer from Nikon! Thanks for sharing, even though you were probably not supposed to post that in a public forum.
Click to expand...

Quite right. I'll remove the reply, and promise to relay the findings from my side-by-side tests this weekend.
 
Last edited:
While flat picture profiles are interesting, I'm more intrigued by what might be done with the uncompressed data stream. Thom Hogan noted he recorded the uncompressed data stream and it gobbled up storage space - much more than the same data wrapped and compressed into a ProRes 4:2:2 file. He didn't mention particulars, like what compression level of ProRes he used or how much bigger the uncompressed files were.
Now I know the uncompressed HDMI data stream is in no way RAW, since it has been debayered. But since there is so much data there, I'm wondering if there is a way to preserve that data to give greater color depth and latitude for grading in post. In short, would it be worthwhile for Atomos, AJA, Sound Devices or someone else to develop a proprietary compression routine to record a flat (S-log) file in ProRes using the uncompressed data from the HDMI. By worthwhile, I mean that the resulting file would have greater latitude for grading than, say a flat file produced using preset picture controls, possibly even extending the bit depth beyond 8.
This may sound ridiculous to someone who knows how these things are done, but I'm just intrigued by the idea that there might be more more useful data available in the uncompressed HDMI stream.
 
Geronimo, I'm imagine that this is obvious to you but I just wanted to mention that Nikon does not recommend the "E" version for video. So just in case you were tempted to spend an extra $1 grand, save the dough ;).
 
t.p. said:
im sorry but its not clear to me , it will only be clear that D4 is indeed softer when a proper side by side comparison will be made , ( the only thing i can say i noticed is no alaising no moire and no line skipping in the d4 video ive seen so far , which i can see all three in d800 (look here with this amateur video test http://www.youtube.com/watch?v=ey8el-YcvgA&feature=related , alaising and line skipping is especially apparent )
in those three D4 video which i have no idea what crop factor have been used , it doesnt look at all that soft to me -
http://vimeo.com/35639074
http://vimeo.com/38752046
http://vimeo.com/38738219
so i'm still waiting for a decent comparison , hope it will come soon
Click to expand...


t.p., IMHO that D800 test looks very suspect. There is blatant mosquito noise around the leaves when they go out of focus. That's not line skipping; that's a codec breaking up.

To my eye, FWIW, the all the D4 clips look soft (when compared to the D800). The first D4 clip you linked to has aliasing on the green shutters in the begining. The second has it on the fence around the tree. The third has such shallow DoF that it's just not useful for judging much.

I do hope to see side by side tests.
 
bufo55 said:
I'm more intrigued by what might be done with the uncompressed data stream.Thom Hogan noted he recorded the uncompressed data stream and it gobbled up storage space - much more than the same data wrapped and compressed into a ProRes 4:2:2 file.
Now I know the uncompressed HDMI data stream is in no way RAW, since it has been debayered. But since there is so much data there, I'm wondering if there is a way to preserve that data to give greater color depth and latitude for grading in post. In short, would it be worthwhile for Atomos, AJA, Sound Devices or someone else to develop a proprietary compression routine to record a flat (S-log) file in ProRes using the uncompressed data from the HDMI.
Click to expand...

- You want great color depth for greater [color] grading? You need 10bit uncompressed that's the bottom line.
- Of course uncompressed is going to be huge. hasn't anyone been paying attention? This is why Apple wrote ProRes. It super efficient even if it's not the full color gamut at 4:2:2
After all, if you want to edit RAW you need to switch to RED. Why is everyone trying so hard to duplicate RED Cinema on a shoestring? it's like saying... If Volkswagen just became friendlier with their Porsche division, we should be able to get smoki'n performance out of the next VW. Just give it up.
- "In short, would it be worthwhile for Atomos, AJA, Sound Devices or someone else to develop a proprietary compression routine to record a flat (S-log) file in ProRes using the uncompressed data from the HDMI". Well then it would not be ProRes but I guess that's what a lot of coders are furiously working on in heads-down rooms all over the planet and when they're done it will likely be called RAW-C (for compressed) the purest form of RAW ;)
- device makers like Atomos, AJA, Sound Devices rarely develop proprietary CODECs. Not a good value for the money (ask RED how much their platform cost to develop).
These CODEC activities are usually the result of consortia of HW/SW + Stakeholders hammering it out for a combined roll out, because too proprietary reminds everyone of 'Beta vs VHS' cases in history.
 
macgregor said:
And where do you live that you can buy the 2 cameras? Because pretty much everywhere they are impossible to get.
Click to expand...

Call your local stores. I live in Seattle, and have been calling our Tall's Camera Store every day for the past couple weeks. They have 1 D4, and a handful of D800's, coming in today. I've got my name on the D4 and one of the D800's.


That said, I received an email reply from Nikon this afternoon from a question I posed to them last week about the best video quality between the D4 and D800 for a filmmaker. I am prohibited from passing along the contents of their reply, but I am now nervous about picking up both cameras on my own dime to perform side-by-side tests.

If the tests conclude as I suspect they will, I don't really want to be stuck trying to sell a $6600 camera on craigslist or ebay.
 
I'm still really confused by the D800 samples we're seeing. Most of the time (at least with unchallenging scenes), it looks great - and some shots like those brick/roof ones Pappas posted above look perfectly normal, but then out of nowhere other shots seem to exhibit wildly out of control moire (worse than anything I've seen from the 5DII).

I wish I could form a clearer picture of what the camera's video limits actually are.
 
Grug said:
I'm still really confused by the D800 samples we're seeing. Most of the time (at least with unchallenging scenes), it looks great - and some shots like those brick/roof ones Pappas posted above look perfectly normal, but then out of nowhere other shots seem to exhibit wildly out of control moire (worse than anything I've seen from the 5DII).

I wish I could form a clearer picture of what the camera's video limits actually are.
Click to expand...

I'm not sure who said it but "between the idea and the execution there falls a shadow"
It's impossible to control all the variables between capture and posting to youtube. Only once we have an empirical and controlled test can we be sure what's what. That's why Phil Askey is on my FAV list. he does it correctly.
 
Grug said:
I'm still really confused by the D800 samples we're seeing. Most of the time (at least with unchallenging scenes), it looks great - and some shots like those brick/roof ones Pappas posted above look perfectly normal, but then out of nowhere other shots seem to exhibit wildly out of control moire (worse than anything I've seen from the 5DII).

I wish I could form a clearer picture of what the camera's video limits actually are.
Click to expand...


I agree, that has me completely confused too. The detail like the wire that's on the chimney should be aliased as it would on a 5DMII. However it doesn't. The detail on the label of the bottle should break down as we have seen on other DSLR's. However it's not. Then we see some D800 clips that look all messed up in certain areas. Could it be camera settings that are different from those that shot the clips? Could detail be dialed down on those that seem to exhibit less issues?


Pappas
http://pappasarts.wordpress.com/
http://twitter.com/pappasarts
 
You must log in or register to reply here.
Back
Top