Samsung NX1 - which is 4K video and which is the 28MP raw still? Can you tell?

[maxotics]

I couldn't have been wrong with a nicer image ;)

[GMaximus]

OK, B-image is 3840x2162

didn't want to spoil it for you guys.

[MattH]

Okay B is raw and it is a bit better but that's still pretty impressive. No one but video geeks we are here would be able to tell the difference! And the raw images are good enough for billboard-sized prints. I feel like with the advent if 4K the quality of motion is really finally catching up with stills.

Well done samsung. This is the first camera I can see no obvious image issues in. No weird colours, blue and colour blotchs, highlight aliasing, horrible rolling shutter, digital sharpening, this looks closest to the 1DC than anything I've seen and I really like the Samsung colour science.

 

Look again. Rolling shutter is as bad as any camera out there.  33ms readout in UHD.  Its the only thing that truly bothers me about the camera.

[Andrew Reid]

Look again. Rolling shutter is as bad as any camera out there.  33ms readout in UHD.  Its the only thing that truly bothers me about the camera.

 

That rolling shutter on the NX1 is similar to the 1D C is actually quite a strong point.

 

One costs $1599, the other $12,000.

 

One is doing a 28MP readout in those 33ms, the other just a 8MP readout (and with a crop of the sensor).

 

One is doing that on a smaller battery. The other is a clunky beast all round.

 

One is labeled Samsung, one is labeled Canon.

 

One is raved about and hyped to an alarming degree. Guess which.

[TravisKelleher]

My guess is that sample "A" is 4K.
 
In sample "B" it looks to me like some chromatic abbreviation around the numbers.
This leads me to think that the video processing takes care of removing CA on the fly.
The encoder also applying some technique to scale down the resolution to 4K, further reducing the effects of CA.

[Guest 560a4aedcb80685284629074497fdc75]

My guess is that sample "A" is 4K.
 
In sample "B" it looks to me like some chromatic abbreviation around the numbers.
This leads me to think that the video processing takes care of removing CA on the fly.
The encoder also applying some technique to scale down the resolution to 4K, further reducing the effects of CA.

 

Yes I'd definitely go for A being video now too. Particularly after reading the other comments ...

[Guest 560a4aedcb80685284629074497fdc75]

I'm not such a big fan of this approach. Having a trained eye is different from feeling your way through differences, and it's good to know exactly what you're looking for and where you're going to find it. That's why I tried to be quite specific about what I was seeing when I wrote my guess.

 

I wasn't suggesting it be the only method you use, just that you try it. I think this whole 'test' proves that a trained technical eye can sometimes be as much a handicap as a professional necessity.

 

Very few of us make images for other tech-geeks. We are the illusionists trying to wow the audience, not impress other magicians with our technical virtuosity. If we can't step outside ourselves and see our work without analytical eyes, we just get bogged-down in easily quantifiable stuff like resolution and compression artefacts. It can take away from the bigger picture. I wasn't boasting when I said I could see B was RAW clear as day, immediately. I really could. I honestly thought that everyone was going to follow me and say "easy, B is raw." I was genuinely amazed when people started saying A is raw.

 

If you look at the two images and genuinely forget about the analytical stuff and just ask yourself one question: "which image feels most real", I believe that B is the only answer. 

 

Anyway, well done, you were right too!  :) (thank god)

[Guest 560a4aedcb80685284629074497fdc75]

As jcs notes, both images are now 8-bit jpegs, so this talk of 8-bit versus 10-bit isn't entirely convincing.

 

It doesn't work quite like that. I see everything on an 8-bit monitor but I can still usually tell the difference between an 8-bit and XX-bit image. 4K downsampled to 1080 looks "better" than HD native footage. Why not the same with bit-depth? You can't analyse and quantify everything objectively. Sometimes you just have to believe your eyes.

[Andrew Reid]

 

Here's a 1:1 crop, as you can see the Samsung 30mm F2.0 has some chromatic aberration wide open in the raw (bottom image) but the video processor completely gets rid of it, which sharpens up the black text nicely. Best video processor ever in a stills camera.

[MattH]

That rolling shutter on the NX1 is similar to the 1D C is actually quite a strong point.

 

One costs $1599, the other $12,000.

 

One is doing a 28MP readout in those 33ms, the other just a 8MP readout (and with a crop of the sensor).

 

One is doing that on a smaller battery. The other is a clunky beast all round.

 

One is labeled Samsung, one is labeled Canon.

 

One is raved about and hyped to an alarming degree. Guess which.

 

I wasn't really talking about the 1DC as something that expensive would never be on my radar anyway.   I was just pointing out that as far as rolling shutter goes it certainly isn't in the good category.  Its in the category where you have to be particularly mindful of it.  Slow pans with a Tripod or IS lenses being a must in my opinion.  As far as comparing it to other cameras, the GH4 is better (just waiting on that log profile) and the BMPC of course is perfect.  Although it has its own problems.

[GMaximus]

...the video processor completely gets rid of it, which sharpens up the black text nicely.

It also applies the noise reduction!

[GMaximus]

Tried to grade A-image as close as it gets (with my eyes and this screen and ACR):

http://we.tl/PGDkRov6vb

Probably someone better than me could get more out of reds and magentas. Or i think they're lost.

[Jaw]

Thanks, very interesting. I just collected my NX1 and is very happy with it. I think the quality is great and unbeatable at that price. However I do notice a few very negative reviews this last two days. One said the focus tracking as very bad and another basically implying it's the worst camera ever for video shoot and couldn't understand how any one will give it a positive review. It's getting curiouser and curiouser.

[Andrew Reid]

Thanks, very interesting. I just collected my NX1 and is very happy with it. I think the quality is great and unbeatable at that price. However I do notice a few very negative reviews this last two days. One said the focus tracking as very bad and another basically implying it's the worst camera ever for video shoot and couldn't understand how any one will give it a positive review. It's getting curiouser and curiouser.

 

Allow me to fill you in on that second one at C5D. It's a seriously biased piece of writing and testing. Johnnie and Seb at that site fell out with me a year ago after they made a deal at NAB with a Berlin cine shop to do an exclusive review of the KineRaw Mini, practically whipping the camera from under me in my own back yard, doing deals behind my back with my own contacts and friends. Regardless when the camera landed in Berlin the shop offered the demo unit to me first anyway and I emailed Johnnie to ask if it was ok to put my results out there. He said yes ok, as long as no video. I forgot about the video and as part of the page of images put a very short clip out there showing some bubbles in slow mo and he lost his rag saying I couldn't be trusted. That was over a year and a half ago.

 

Now we come to the present and the NX1 is a target because I have been advocating it as a decent camera. So the C5D test is basically there to discredit me. And this tweet just before it came out, is another passive aggressive swipe at EOSHD - 

Baffled by some positive Samsung #NX1 reviews out there concerning video. How can anyone with any sense of pro video recommend this camera?

— cinema5D (@c5dnews)

December 3, 2014

 

It's full of flaws anyway. They criticise the EVF at the same time as giving the 7D Mark II top marks for handling... well that doesn't even have one! The EVF on the NX1 is great anyway! They complain about usability when all their complaints are solved merely by setting up the camera properly.

 

Put stills to 16:9 so it matches video framing, use peaking and magnified focus assist, hit record. Not hard is it? The 7D mark II doesn't even have peaking and neither can magnify focus whilst recording yet the NX1 gets a kick in the knees for it and the 7D doesn't. They barely mentioned it.

 

Most critically the workflow they used was totally the wrong one. They had SmartRange turned on and Contrast -10. Double trouble. They say 18 hours to transcode 1 hour of footage, but that simply isn't the case. They didn't use Wondershare to transcode to ProRes and only used the supplied Samsung software to transcode to H.264.

 

On top of this their rolling shutter test has been flawed previously, rating the GH4 in 4K better than the 5D Mark III in 1080p and the A7S in APS-C crop mode. That's not borne out in reality. Likewise their dynamic range test has brought up weird results. They rated the A7S at 14 stops, same as an Arri Amira and the 5D Mark III with CineStyle ahead of the 1D C with Canon LOG!! I'm curious how they get 11.8 stops out of the 5D Mark III's stock video mode when the sensor is only rated for 11.5 stops in raw!!

 

It all just seemed overly critical, more critical than their other reviews for distinctly less impressive cameras. What a shame some personal bias has affected their impartiality as reporters. I hope nobody cancelled their NX1 order as a result.

[Nikkor]

That site is dead, who cares.

[MattH]

That site is dead, who cares.

 

Andrews wrists and fingers care after his last coment and your blunt summary. :)

 

It is obvious that the writer of that cinema5d review was determined not to like the camera from the start.  It seems rushed and they probably rushed the tests as well.  It's sort of telling that they dont show the methodology or even show the images or videos they used to draw their conclusions.  The rolling shutter result is in the right ball park.  I measured it at 33ms in uhd from various clips.  There measurment for the a7s is wrong though.  In full frame the rolling shutter of the a7s is roughly equal to the nx1.

[Promit Roy]

So as long as we're on the subject - what IS the current feeling about the GH4 vs the NX1 as hybrid cameras? (So taking both stills and video into account.) Both are obviously excellent cameras, but are there any key points to tip one way or the other?

[Jaw]

Allow me to fill you in on that second one at C5D. It's a seriously biased piece of writing and testing. Johnnie and Seb at that site fell out with me a year ago after they made a deal at NAB with a Berlin cine shop to do an exclusive review of the KineRaw Mini, practically whipping the camera from under me in my own back yard, doing deals behind my back with my own contacts and friends. Regardless when the camera landed in Berlin the shop offered the demo unit to me first anyway and I emailed Johnnie to ask if it was ok to put my results out there. He said yes ok, as long as no video. I forgot about the video and as part of the page of images put a very short clip out there showing some bubbles in slow mo and he lost his rag saying I couldn't be trusted. That was over a year and a half ago.

Now we come to the present and the NX1 is a target because I have been advocating it as a decent camera. So the C5D test is basically there to discredit me. And this tweet just before it came out, is another passive aggressive swipe at EOSHD -

Baffled by some positive Samsung reviews out there concerning video. How can anyone with any sense of pro video recommend this camera?

— cinema5D (@c5dnews)

Baffled by some positive Samsung reviews out there concerning video. How can anyone with any sense of pro video recommend this camera?

— cinema5D (@c5dnews) December 3, 2014

">December 3, 2014 ">

Baffled by some positive Samsung #NX1 reviews out there concerning video. How can anyone with any sense of pro video recommend this camera?

— cinema5D (@c5dnews)

Baffled by some positive Samsung reviews out there concerning video. How can anyone with any sense of pro video recommend this camera?

— cinema5D (@c5dnews) December 3, 2014

">December 3, 2014

It's full of flaws anyway. They criticise the EVF at the same time as giving the 7D Mark II top marks for handling... well that doesn't even have one! The EVF on the NX1 is great anyway! They complain about usability when all their complaints are solved merely by setting up the camera properly.

Put stills to 16:9 so it matches video framing, use peaking and magnified focus assist, hit record. Not hard is it? The 7D mark II doesn't even have peaking and neither can magnify focus whilst recording yet the NX1 gets a kick in the knees for it and the 7D doesn't. They barely mentioned it.

Most critically the workflow they used was totally the wrong one. They had SmartRange turned on and Contrast -10. Double trouble. They say 18 hours to transcode 1 hour of footage, but that simply isn't the case. They didn't use Wondershare to transcode to ProRes and only used the supplied Samsung software to transcode to H.264.

On top of this their rolling shutter test has been flawed previously, rating the GH4 in 4K better than the 5D Mark III in 1080p and the A7S in APS-C crop mode. That's not borne out in reality. Likewise their dynamic range test has brought up weird results. They rated the A7S at 14 stops, same as an Arri Amira and the 5D Mark III with CineStyle ahead of the 1D C with Canon LOG!! I'm curious how they get 11.8 stops out of the 5D Mark III's stock video mode when the sensor is only rated for 11.5 stops in raw!!

It all just seemed overly critical, more critical than their other reviews for distinctly less impressive cameras. What a shame some personal bias has affected their impartiality as reporters. I hope nobody cancelled their NX1 order as a result.
Well, I bought mine and loving it! Thanks for the background information. I certainly won't be visiting that site any nire. I mean they made it sound as though the NX1 is really the worst camera available for video!

[Promit Roy]

I wasn't suggesting it be the only method you use, just that you try it. I think this whole 'test' proves that a trained technical eye can sometimes be as much a handicap as a professional necessity.

 

Very few of us make images for other tech-geeks. We are the illusionists trying to wow the audience, not impress other magicians with our technical virtuosity. If we can't step outside ourselves and see our work without analytical eyes, we just get bogged-down in easily quantifiable stuff like resolution and compression artefacts. It can take away from the bigger picture.

I work in real time 3D graphics (read: games) and unfortunately people have a habit of dissecting the hell out of screenshots. I mean Halo was excoriated over rendering at 900p instead of 1080p and then doing an upscale. Can you imagine someone complaining that a movie, even a CGI movie, was only delivering 900 vertical lines of resolution when screenshotted? That would be insane. Unfortunately it has led to emphasis of technical parameters over aesthetic quality, and so a lot of these supposedly high quality games look awful. I don't feel that the art direction in many big name games is competent, as well. 

 

More relevant to here, though, is that working in imaging that way attunes you to every single fault - and feature - in an image. I need to be able to see everything, because my job is to get it right and make sure there are no mistakes in the underlying systems which might later crop up in surprising and unwelcome ways. My wife gets annoyed when I stop a game or movie in order to take a look at the technical aspects of how it's put together  :D

[majoraxis]

Between the GH4 and NX1 - I will be purchasing the NX1 with the 16-50S f2.0-f2.8 over the GH4 with 12-35 f2.8 for the following reasons:

 

1. Autofocus with the 16-50 2.0-2.8 seems to work pretty well (in good light).

2. Image stabilization on the Samsung does not seem to cause jitter.

3. Colors are great with not that much grading required. It looks like it can be graded if desired.

4. Depth of field or lack there of. 16mm at f2.0 for the NX1 is about 24.6mm f3.0in FF.  The GH4 in 4k mode at 12mm f2.8 is 26.4mm at f6.16.

5. Shows less noise and the image does not lose its punch at higher ISOs.

6. Slow motion can go slower 120 fps and is said to be as good as 96 fps on the GH4.

7. Photo mode is 28.2 Megapixel vs 16.05 Megapixel - 43% greater resolution or dare I say full frame DSLR resolution.

 

That said if I was shooting video with very bright/high contrast scenes and I had fast manual glass with an ND (and did not need auto focus or image stabilization) I would pick the GH4 as it seems to have a bit more dynamic range at the lowest ISOs and with up and coming log profile it will most likely be able retain more of that dynamic range.

 

If your not going with fast manual glass - the NX1 with the 16-50S is the way to go IMHO.