Zacuto Camera Comparison Test

Film is awesome! (of course...)

-I tought that the HPX170 came out really well compared to film.
-The HPX3000 would probably be my second choice after film. (too sharp in the test, but could be bump down...)
-The RED one looked "bloomy", almost like it had some kind of filter (why?)
-Strangely, I thought that the Elite looked as good or better then the Ultimate (?)
-All those adapters remove too much DOF...
-5DMKII... What the hell?! Way to "punchy". Something's wrong with the contrast?
-D90... Simply too soft.

Would love to see a Cinemek in your next test.

2 Questions :
-Why didn't you push the HPX,HVX to 1080p?
-Did you convert the 5DMKII to 24fps? How?

Thanks for the test and all the troubles that you must have went trough!
 
I understand shooting the round table in 720/24pn, but why the HVX/HPX shootout footage?

e
 
TheRealMe said:
Film is awesome! (of course...)

-I tought that the HPX170 came out really well compared to film.
-The HPX3000 would probably be my second choice after film. (too sharp in the test, but could be bump down...)
-The RED one looked "bloomy", almost like it had some kind of filter (why?)
-Strangely, I thought that the Elite looked as good or better then the Ultimate (?)
-All those adapters remove too much DOF...
-5DMKII... What the hell?! Way to "punchy". Something's wrong with the contrast?
-D90... Simply too soft.
Click to expand...

170 is a sweet camera. We were disputing which scene files to use on the 3000 if I remember correctly. It has such a wide range of color though and good dynamic range! RED well there were some problems with exposure on the RED as has been previously discussed, but RED is a great camera. The contrast was set on the stock settings. It can be changed to -2 on the MK2. D90, well it's a prosumer product so what do you expect.


TheRealMe said:
Would love to see a Cinemek in your next test.
Click to expand...

That can be arranged if you put word word in for me over there. I'm aiming for late January to start my test including the Blade, Ultimate, Extreme, M2, Brevis, and mini35.

TheRealMe said:
2 Questions :
-Why didn't you push the HPX,HVX to 1080p?
-Did you convert the 5DMKII to 24fps? How?
Click to expand...

I don't quite remember why that was. Everything in the vimeo video is down-rezzed to 720 anyway though. I'll keep this in mind down the road though for future tests on my own and with Zacuto. 5D2 footage was converted at the post production house. Ask those wizards!
 
Although I don't work directly for Zacuto, you can see from the video that I'm friends with all those red-bleeding boys. So if anyone is interested in knowing more about their products feel free to PM me. I might even be able to help you get a little discount! :)
 
joe 1008 said:
Any mail? I will do so seriously. :)
Click to expand...

HAHA. I love how this shootout has evolved. Being the producer, I was more concerned with the message of the discussion, the quality of the musicians, and the organization of the days prior, during, and after production. I feel like I've answered a million questions already... SO KEEP EM COMING!
 
I think a lot of us no budget filmmakers would love to see a shootout between the D90 and other cameras in that price bracket like the HF11/HG20 and particularly the DVX. It would be great to see how the D90 stands up against an SD camera like the DVX that has had it's fair share of transfers to 35mm. hint hint. The poor D90 didn't have a chance, where was the super 8 comparison? ;P
 
Wow, the Nikon looked like sh#t. Blocky and soft.

I'm pretty sure I saw something better looking out there on vimeo... Maybe it's the same problem as RED, a daylight balanced sensor and a noisy blue channel - only worst since the camera is compressing the heck of it.

Let's see if they can get it a little better with their rumored firmware upgrade...

I'm still not convinced on the 35mm adapter footage. Those things can look great for a second only to let you down the mintue something changes (lenses backfocus, camera misalignment, whatever). I just can't take equipment on a shoot that has historically given me an extra problem to solve...

And Red, shame on you... :D Just kidding, the red footage could be sharpened quite a bit, remember they give us zero sharpness footage off camera...

I think you didn't do the Canon justice. It looks better in your RED vs 5D excerpt.

Thanks for the effort, and honest discussion. I love to hear from a product manufacturer that it doesn't matter how much money you put on your gear. This kind of approach to filmaking is guaranteed to render some great products!
 
I exported stills for all the cameras from the 306 MB QT file, just to compare what's happening at 720P resolution. The 5D-MKII settings definitely need some work.

...I tried dropping the 5D clip into Vegas and used the Levels plug-in to adjust the "Input End" to 0.800, which is something I would normally do with any footage I'm working on, and it brought the 5D clip back to life. ( see images below )

Original 5D Still from QT file
5D_Original.jpg


Tweaked 5D Still from QT file ( Sony Vegas "Levels" plug-in )
5D_Tweaked.jpg


...One thing to look for with the different camera clips is how the candle flames are recorded. In the 35mm clip the candle flames have a fairly sharp edge, with color and tone in the flame. With most of the DOF adapters you can see obvious chromatic abberation in the flames, which is softening the image a fair bit. The Redrock clip looks like the worst in this test. Notice that the candle flames in the 5D clip are sharp, but have little color or tone, indicating that the camera settings were still too contrasty for the scene being recorded. ( the highlight on the woman's cheek has little detail, another indication of too much contrast )

If the 5D exposure had been lowered and the fill lighting made stronger, you could probably come much closer to the look of the 35mm clip using the 5D.

My Take on the Cameras:

35mm : Looks great, almost as if the lighting was designed for this camera.
HPX3000 : Looks great. Sharp and able to handle the contrast with ease.
RED One : Surprisingly soft. Color abberation in the candle flames. ( allignment ? )

HPX170 : Nice look, less detailed than the HX-A1 or EX3.
XH-A1 : Nice look, sharper than the HPX170, but not the EX3. Slight color abberation in the candle flames.
EX3 : Nice look, sharper than the HPX170 footage.

HVX200 Letus Elite : Nice look. A bit soft. Obvious color abberation in the flames.
HVX200 Letus Ultimate : Nice look. Softer than Elite for this setup. Color abberation in the flames.
HVX200 Redrock : Nice look, but quite soft. More color abberation than Letus DOF.

5D MKII : Very impressive for still camera. With tweaking it should be great.
D90 : Very soft ( low resolution ), with lower contrast.

For day to day shooting on my budget I would still go for the HPX170 because I like the look, and absolute sharpness is not a big deal for me. The EX3 is also very nice, but the ergonomics aren't great for my one man show.

If I had a crew and a healthy budget, I would switch to the HPX3000. The great color and long exposure latitude are really nice.

I am defintely buying a 5D MKII. With proper tweaking I think it could be a great small production camera, especially when you have time to play.
 
The lighting was designed for the HVX200 Redrock setup. However, the 35mm film was properly exposed using a light meter to f/4. This was based off of reflective light metering to the female violinists face. This explains the differences in the DOF between RED and 35mm. If we had thought it through, we would have used additional ND filters such that 35mm film would still be properly exposed, but at f/1.4 versis f/4.

It has already been determined that the backfocus on the RED could have been off.

We used the stock settigns on the 5D MKII. The only real tweaking that can be done is lowering the contrast to -2. Our contrast was set at 0. I was hoping for more out of the D90 myself. The fact that it's exposure changes off of lock every time you turn off the movie recording function is rather annoying as well. Anyway, such are first gen products!
 
Not only was the backfocus on the RED possibly off, but the lens that was on it was set to an f/1.4. That's the real reason why the footage is so soft. Any lens that's not a Master Prime will have tons of coma and aberration wide open.

If the RED was at an f/4 instead of the 1.4, it would be tack sharp.
 
I'm looking at the all the specs of the shoot and wondering how did you set the gain on HVX200 to -3db ? I didn't know you could do that... I always thought you could only go "up" on the gain on the HVX200....
 
joe 1008 said:
And there is it again on the strings of the violin: The aliasing that worries me so much...
Click to expand...

The violin strings are essentially white strings against a black background, a pretty extreme situation. Look how the violin bow is being recorded, there's no apparent aliasing there.

My take take on the whole 5DMKII thing is that this is Canon's first "toe in the water" for this type of technology. I expect to see this technology mature over the next couple of years, then maybe we'll have a DSLR video body being released by one of the majors.
 
Because of the recording pattern of the 5D's sensor, thin horizontal lines tend to aliasing. Vertilcal lines are almost not affected. Comparing this still to other footage I would say that the aliasing on the strings is quite typical.

Of course there will always be something better. But I personally am so intrigued by the possibilities of the 5D that I'm looking for workaround for its mayor flaws - and there are MANY of them.
 
all the scaled formats have aliasing btw, including film and red one. i guess the post people either didn't scale it correctly or messed up the field settings. the 720p footage doesn't have it, except the d90 which has the same sampling flaw as the 5d but even then actually less than all the scaled formats. look at the strings near the bridge.

/matt
 
Well, I don't think it's quite fair to compare the D90 based solely on the AE Lock issue. If you compared the Canon HV20 or HV30 in the test you'd have to play around to lock the exposure on those too.

The D90 footage has to be treated in post, that's all there is to it. Without doing the displacement map fix and adding some slight sharpening, of course the D90 doesn't stand a chance. I wish I could post my Film Production One final project that I used the D90 for. That footage smokes this test and really inspired a few of my fellow HVX and DVX-using students to convert over, however I did have to warn them about how the footage needs to be treated first and the workarounds for it.

Regardless of how many years of experience the people who did this test have, if they don't know how to properly use the new tools given to them, especially something as tricky as the D90, how can they be expected to accurately make a fair comparison? Just reading the D90 stickies on DVXUser can show you how to get the best out of it.

I've learned that a really good thing to do with the D90 is to film brighter and darken in post, you can easily do this by adjusting the exposure by holding the lens cap up in front of the lens and trying to find the point where the footage is about a stop or so above where it should be and then locking it. This is different for low light situations, you really do have to play with it then to prevent the scrolling lines. But that's the nature of the camera, you have to play around with it and it isn't as bad as it sounds, once you know what to do and what to watch out for the workflow is very quick. If you just lock things down where they look just okay and expect magic results, you're going to be dissappointed.

EDIT: My reasoning for going brighter is that because the footage is so compressed, I find it's easier to bring the brightness down and preserve shadow detail than it is to try to bump it up from being too dark and just making the blacks more gray. Hopefully Nikon adds ISO adjustment somehow but even then I would select slightly higher for the same reason of preserving shadow detail.

By the way, I too was surprised how well the HD cameras held up against the RED and 35mm film. But I have to think, would a 35mm show like "Heroes" look the same if shot on one of these other HD cameras? In that sense, the test is relative to what you want to do. I want more of a 35mm look but digital and a RED isn't financially feasible and the 5D Mark II doesn't have 24p, so the D90 is the only option. Maybe I wouldn't do a special effects extraveganza on it, but for the non-special effects films I'm doing now, it's fine for that. If I needed special effects then a RED rental would make the most sense.
 
Last edited:
mattsand said:
all the scaled formats have aliasing btw, including film and red one. i guess the post people either didn't scale it correctly or messed up the field settings. the 720p footage doesn't have it, except the d90 which has the same sampling flaw as the 5d but even then actually less than all the scaled formats. look at the strings near the bridge.

/matt
Click to expand...
:thumbsup:
 
You must log in or register to reply here.
Back
Top