Nikon D5300 video samples.

Samuel H said:
GREAT TEST!
* much better DR in the D5300, at least a stop or two better in the highlights, and about the same in the shadows
* the mesh in the pencil can looks very stable, it seems the D5300 will be OK in the aliasing/moire front
Click to expand...

You are seeing the same thing than me. The Nikon's in general are more like one stop better DR than the Canons, but in this case I see at least 2 stop more which put the D5300 at least 13 stop DR !!!!! It would be insane but not impossible because the D4 was measured at 13 stop in the BBC test. I don't have a light meter to measure it but this video share.ols.inode.at/D0314R3IZ8OL2YPT4SZ5NE72Z8UXVI89R00V1BD1 seems to have a lot of DR with the highlight/reflection on the cars windows/rooftop and some shadows under the bridge and the D5300 did a very fine job of capturing it.
 
Here is a quick test, with Neutral all to minus.. original files http://yadi.sk/d/GoupLGm2Dtkrv

D5300, 5dmk3. at 1600/
above 1600 D5300 produces too much noise to compare with 5dmk3, at 1600 dynamic range on 5300 better..
I find 5dmk3 to be softer as well.. Nikon 5200, 5300 are amazing within 800 iso limit, its their bread and butter, both easily beat 5dmk3..
 
It worked for me.
This time I see much better shadows on the D5300 than on the 5D3, but the specular highlight is too tight to notice a difference.

It seems to have better DR than what we had before, it has 1080p60, it is decently sharp, and I still haven't found any serious issues with the image.
The D5200 was my recommendation at its price bracket for the last year, I think this is a very mighty successor.
 
For me the test is not good and it is not the fault of Oleg. The problem is that it is either Nikon which under values it ISO or Canon which overvalues it ISO (which is more likely the case). At least since the Nikon D800/5d3 the Canon is 2/3 iso darker than the Nikon. So ISO 1600 on Nikon is more like ISO 2500 on Canon 5dmark3. You can see a test done by Joe Marine from nofilmschool between Nikon D800 and Canon 5D3 and you can clearly see that the Nikon is much brighter at same ISO https://vimeo.com/40113110 . At least he redid the test when he saw that to equalise the exposure in this second test https://vimeo.com/42381520 .

At least a good reviewer because the rest of the reviewers are just dump, they just take what is written on the box as gospel and I have even seen it in photo reviewing site. Those pseudo expert testing ISO and the shot of Nikon (and other brand) is at lets say 1/100 sec and the Canon one at 1/60 sec. Can people be that foolish or have vested interested.

To come back to the sample above it is really to make any assumption at the ISO performance because the Nikon is clearly much brighter. At first I thought that perhaps it was because of better Dynamic range, but looking at histogram they have more or less same DR, it is just that the Nikon is brighter where you see more into the shadows and the highlight clipping more.

For the DR I think it was to be expected that it would be much closer now for both camera than at ISO 100 because the Canon 5d3 tend to have better DR in high ISO even against the D800 as from ISO 1000. So for this test to be good we would need like the Joe Marine test, ISO 1600 Nikon D5300 and ISO 2000, 2500 and 3200 from the Canon 5d3 to know more or less what is the equivalence in terms of exposure.
 
Daniel, I agree with your point to compare cameras at the same exposure.. I noticed that Canon usually 2/3 stops darker, vs Nikon and Panasonic cameras.
Still D5300 above its 1600 iso produces too much noise, which to me is fine, I never go beyond 1250 with Nikon 1.4, 1.8 glass. any cinematic expression requires lighting more or less..
I will do another test and compare, give original of both 5300, 5200. I find overall D5300 produces less noise than 5200.

just did a test, D5300 at 1250 vs 5dmk3, iso 2000, 2500, at the same exposure.. amazing, at 2500 on 5dmk3 highlights are getting blown away, still not too much detail in shadows)At neutral setting all to minus, D5300 is a full stop brighter! I also see some banding in 5dmk3 footage at higher iso.
http://yadi.sk/d/XkAaSjyLDupLF

Daniel, you've asked about D5300 vs D5200.
I found out that D5300 at iso 800 produces more noise than iso 1250, nothing like that on D5200.
overall D5300 produces less noise than 5200, look at 800, 1250, 1600 iso.
http://yadi.sk/d/-W9s7oihDurQD
 
Last edited:
What's your take on the 1080p/60? Is the image compromised (lower bit rate, different codec)? Wondering how they squeezed that into a "cheaper" body.

Would LOVE LOVE LOVE to know the scoop on that! Especially since you guys are testing with some seriousness... (thanks for that BTW).
 
60 p looks the same more or less, so fat did not have a chance to study the cadence, smoothness, etc..
I looks great so far, one thing to know, there is no exposure indication, no histogram while using manual aperture lenses, have to use light meter, which is more cinematic any way.
 
Yes, that's common in many DSLRs (including the first Canon bodies, back in the day, until ML came to rescue us). If you don't have a light meter, the cheap solution is to take a still and look at its histogram.
 
I've noticed on the 7100, video in playback (the still before hitting "play") also shows a histo and flashes highlights. I used to use stills but can't recall if earlier bodies gave a histogram for video.

I like it through, since you're seeing the histogram/blowouts for the actual video clip.
 
So they know we want it, but won't put it there yet. It's painful to see them bringing the features so horribly slow...
Shooting stills for this keeps your clips organized, but stills and video can be different depending on your settings (e.g. the old Flaat_13p was useless for video but looked great on stills) so doing it with video will be more accurate.
 
Samuel, have you tried to use more agressive curves, and lower pedestal, to 0.9 instead of 1.0 for example?
I have, interesting results with D5300.. Nikon in this camera changed colorimetry, it became more like Canon, less yellow green.. a lot, a quite high pedestal, black level can be lowered, thus producing less noise. I'd try your Flaat13.
 
OlegKalyan said:
Samuel, have you tried to use more agressive curves, and lower pedestal, to 0.9 instead of 1.0 for example?
I have, interesting results with D5300.. Nikon in this camera changed colorimetry, it became more like Canon, less yellow green.. a lot, a quite high pedestal, black level can be lowered, thus producing less noise. I'd try your Flaat13.
Click to expand...

Hi Oleg, could it be possible to do a 100 ISO side by side between D5200 and D5300 with lowes contrast setting, because you did one between 5d3 and D5300 and the result in terms of DR where very big. I would like to know if I was not too much optimistic, because it surely looked like more than the 1 stop difference the latest Nikon have above the Canon.
 
OlegKalyan said:
Samuel, have you tried to use more agressive curves, and lower pedestal, to 0.9 instead of 1.0 for example?
I have, interesting results with D5300.. Nikon in this camera changed colorimetry, it became more like Canon, less yellow green.. a lot, a quite high pedestal, black level can be lowered, thus producing less noise. I'd try your Flaat13.
Click to expand...

In-camera image processing changed a lot between the D7000 and newer cameras, I wouldn't be surprised if they changed it again with the D5300.
Write to similaar.feedback@gmail.com and I'll send you the old Flaat_13 profile. On last-gen cameras like the D800, it worked great for stills but not for video (test here).
 
Samuel, thank you, will test it out.. interested of how you've created specifiv RGB curve.. was there a test on every point while creating Flaat profile..

Will do the test on iso 100 over the weekend.
 
For creating Flaat, I took a bunch of very special stills that, composed together, become the poor-man's equivalent of a very expensive wedge chart.
Adjust profiles, develop stills, look at results, adjust profiles, develop stills... rinse and repeat.
That's why I ended up with a Flaat_13 profile that seemed to work but actually didn't, btw.


And also, just a technicality: both in Nikon and Canon cameras, the curve you add with a profile is not an RGB curve, but a luma curve.
 
Last edited:
So what are you both saying Oleg and Samuel, does the D5300 has better Dynamic range than the D5200 in video? I would really have confirmation of that. So with Flaat 13 we would get 12+ DR !!!!

If it is the case could you please post us some example Oleg, I would really be grateful :) even only with Flaat 11 and Neutral dial down.
 
OlegKalyan said:
Daniel, I agree with your point to compare cameras at the same exposure.. I noticed that Canon usually 2/3 stops darker, vs Nikon and Panasonic cameras.
Still D5300 above its 1600 iso produces too much noise, which to me is fine, I never go beyond 1250 with Nikon 1.4, 1.8 glass. any cinematic expression requires lighting more or less..
I will do another test and compare, give original of both 5300, 5200. I find overall D5300 produces less noise than 5200.

just did a test, D5300 at 1250 vs 5dmk3, iso 2000, 2500, at the same exposure.. amazing, at 2500 on 5dmk3 highlights are getting blown away, still not too much detail in shadows)At neutral setting all to minus, D5300 is a full stop brighter! I also see some banding in 5dmk3 footage at higher iso.
http://yadi.sk/d/XkAaSjyLDupLF

Daniel, you've asked about D5300 vs D5200.
I found out that D5300 at iso 800 produces more noise than iso 1250, nothing like that on D5200.
overall D5300 produces less noise than 5200, look at 800, 1250, 1600 iso.
http://yadi.sk/d/-W9s7oihDurQD
Click to expand...

Looking at the samples, I think there is a big chance that it is the internal noise reduction engine which is responsible for the ISO 1250 to be cleaner. It seems that it has less detail. One test to be sure is to do some photo in RAW and compare ISO 800 and 1250 without any NR. I remember my Nikon D300 800 iso in jpeg being cleaner than ISO 400, but you could see that there was some NR going on as detail was a bit washed out.

It is baffling how Canon is clearly cheating. let say 1/3 stop difference would have been acceptable, but 2/3 to 1 stop is too much and all those pseudo expert just shut up there mouth!!!!!!
 
You must log in or register to reply here.
Back
Top