FF VG900 hands-on !

MazingerZ said:
I would only buy this camera for one reason -- the ability to adapt full-frame lenses without cropping, including the Leica/Zeiss/Voightlander M-mount lenses which are much smaller than SLR lenses, with the knowledge that it can suffer from aliasing/moire artifacts in certain (but not all) situations. I would only use the digital zoom feature in situations where there's very minimal or no aliasing/artifacts.
Click to expand...

Some M-mount lenses have issues with smearing and color shift when you get wider than 35mm on APS-c Nex cameras, I'd like to see tests because the FF sensor will likely have more issues. Olympus OM, Minolta Rokkor and Canon FD don't suffer from any of those issues - if I were going VG I would pick one of those lines and get a full set for what a couple M-mount lenses would cost. Just my .02 ...
 
Just to clarify, I did not say that I would buy expensive M-mount lenses to use with the VG.
I already have several old and current M-mount lenses from shooting with Leica M film and digital cameras.
And I also have a few Nikon-mount and Canon FD lenses, so I already have the lenses.

Here's is a quick sample JPG:
http://www.palsomedia.com/vg900/

Leica M 24mm 1.4
ISO 800, 1/40 sec, f/2.0

There's some vignetting and blurring in the extreme corners, but it's good enough for my intended use of this particular lens -- low light or shooting wide open where the extreme corners are not critical.
The rest of my M lenses are longer than 24mm, so they shouldn't have any issue.
 
^^^^ MazingerZ, that looks really good, the corners/edges are pretty bad, but no worse than many other WA lenses at f2.0. Thanks for posting.
 
MazingerZ said:
From testing the VG900, I think the camera performs a 2-step sampling/downconversion in video mode.

(1) First it creates a high resolution downsampled image (smaller than the full 6000 pixels, but larger than 1920 HD).
But the image can contain bad aliasing/artifacts in some situations, for example straight lines/edges.
Which suggests that it's not reading all the pixels in video mode.
This is the image that you see in the viewfinder.

(2) During recording, it performs another downconversion of the image from step (1) to HD (1920x1080).
When you use the digital zoom with the zoom rocker, or use an APS-C lens, it takes a crop of the image from step (1).
Using the digital zoom to with the zoom rocker 1.5X on a full frame lens is exactly the same as the APS-C crop mode.
With the zoom rocker, you can do a smooth zoom anywhere from 1.0X to 2.0x on a full frame lens, or 1.5X to 3.0x effective on an APS-C lens.
Because it's taking a crop of from an already downsampled image (which itself already contains artifacts),
the resolution will suffer as you zoom (or if you use an APS-C lens), and the aliasing or any artifacts are only magnified.

Note that this different from the tele-crop mode of the Panasonic GH2, which is a fixed 2.6 crop of the of the sensor area, and only the pixels
within that area are used, and this area happens to have 1920 pixels so there's no downconversion needed.

Also this is different from the Canon and Panasonic GH cameras -- in those cameras, when you zoom in on the image BEFORE recording, you see the pixel by pixel image before any downconversion, so the image is very sharp and you can see fine details with no aliasing.
And when you hit record, that's when it downconverts the image, with each camera doing the conversion in a different way:
5D2- suffers from aliasing
GH2- sharp image with minimal aliasing
5D3- image looks soft but no aliasing, although it may be OK after sharpening in post

With the VG900, you can see in the EVF exactly what will be recorded and if it will have any aliasing or artifacts, before and during recording -- what you see is what you get.
Curiously, in photo mode, what you see in the viewfinder is similar to what you see in video mode, with aliasing/moire artifacts. But of course when you hit the photo button the actual RAW/JPG still image uses all the 24M pixels without the artifacts.


I would only buy this camera for one reason -- the ability to adapt full-frame lenses without cropping, including the Leica/Zeiss/Voightlander M-mount lenses which are much smaller than SLR lenses, with the knowledge that it can suffer from aliasing/moire artifacts in certain (but not all) situations. I would only use the digital zoom feature in situations where there's very minimal or no aliasing/artifacts.

And I will bring the small GH2 (which has minimal or no aliasing) as backup.
And I can use the same lenses on both cameras -- I'll just need to bring one extra lens to cover the wide angle because of the 2X crop on the GH2, such as the small Olympus 12mm f/2 micro 4/3 lens.
Click to expand...

can you post a video showing this? im most interested how much image degredation there is when using full frame lenses. it would seem from your notes that using the 1.0-2.0x zoom won't degrade the image any more than using an aps-c lens with no digital zooming. is that correct?
 
As I said, using the digital zoom to 1.5X on a full frame lens is exactly the same as the APS-C crop mode,
so the image will degrade exactly the same as using an APS-C lens, assuming both lenses have enough resolution.
Zooming in all the way to 2.0X will degrade the image further.

I uploaded these files in the same location http://www.palsomedia.com/vg900/

Test 1 - Digital zoom versus Optical zoom
--------------------------------------------
00009.MTS - 1.0 to 2.0X digital zoom
As I said earlier, if there are aliasing or any artifacts, it will only magnify those artifacts when you apply the digital zoom.

00010.MTS - To simulate an optical zoom, I positioned the camera closer to the subject in order to get around the same 2X magnification.


Test 2 - Detail when zooming
-----------------------------------------
DSC0081.JPG
- Sample JPG image -- note how the text is detailed and clear.

00001.MTS
- For comparison, the downscaled video image has lost the detail (even before applying the digital zoom).
And when you zoom in, it's only enlarging the already downscaled video image. and you barely see any more detail when zoomed in,
even though the lens can resolve more detail (as shown in the JPG).


Lens: Zeiss ZF (Nikon-mount) Makro 100mm F2
 
Hey a question for anyone using aps-c lenses on this camera. I only have one such lens, the standard Sony NEX 18-55 kit lens. While I find aliasing to be ok on full frame lenses with this camera, with this aps-c lens I find aliasing to be somewhat more severe. Is that to be expected when using all aps-c lenses on this video camera, or is it more because the Sony NEX 18-55 lens is just average quality?
 
As I already mentioned twice in my previous posts, using an APS-C lens or using the digital zoom will simply crop (i.e. magnify) the image, and if the image has aliasing to begin with, it will magnify the aliasing. So the answer is yes, it will be more severe on an APS-C lens.
And I posted a sample (00009.MTS) which illustrates this in my previous posts, zooming in (i.e. cropping) all the way to 2.0X.
 
Well maz based on your findings, it has me more interested in the vg30 again, but none of them in the wild. if it's the same sensor as the 5n, it could be lovely. though there was one fella who reviewed the shoulder cam version which should be the same sensor as the vg30 who thought it was only as sensitive as a pmw-200, which would be rather unfortunate. not to say that the pmw200 is bad, but it's sensor is of course a fraction of the size of aps-c.
 
I just received my 35mm F1.4 Bower with Sony A mount today and plan on trying it with my Steadicam Merlin and using the LA-EA3 adapters tripod mount since the VG900 tripod mount is too far back with the heavy 35mm F1.4 way out front.
8194434390_32c4f1073a_b.jpg
 
longcipher said:
Read it and weep. http://www.eoshd.com/content/9214/the-eoshd-blackmagic-cinema-camera-shootout
Click to expand...

Yawn. Same tired narrative. No credibility because once again he's trying to push his 'GH2 is superior to everything below the FS100' agenda, except it really isn't all that great.

With EOSHD everything always compared to the "superior detail" of the "baby Alexa" GH2 while constantly ignoring the terrible DR, huge amount of posterization in highlights and shadows that are complete bunk. Then there's the usual Canon bashing and Sony hate for everything that's not a cheap Nex or his FS. He trashed the VG20 and then heaped tons of praise on the Nex5n, yet the image each camera renders is identical.

Its really remarkable how many issues he has with cameras he doesn't like when others don't have the same problems. Its almost uncanny how he finds faults others can't replicate with properly exposed and properly focused shots. Almost unbelievable. My A99 is nowhere near as soft as the VG shot posted, but I don't have access to pre-prod GH3's and I don't have an axe to grind with Canon, Nikon and Sony et al.
 
longcipher said:
Read it and weep. http://www.eoshd.com/content/9214/the-eoshd-blackmagic-cinema-camera-shootout
Click to expand...

I'm still trying to figure out how he managed to take a such a bad picture with the vg-900. I can't say that I'm able to replicate that bad of a look on mine. Maybe it's the lens he used? I use a Sigma 50mm f1.4 on mine. Aside from that, I think some of these guys need to understand that there are some out there like me that don't need the added complexity of a cinema cam and don't need and/or can't use raw support, yet at the same time need very good low light performance and very shallow dof. I did my first video shoot yesterday with the vg-900 for a new website I launched and it worked great, very simple and fast to use and I could take shots in low light with the full dof control I wanted. If I shot in raw then I'd have ~7tb/week of data to somehow manage on both magnetic and optical storage archives, it's just not a workable format for me, and being able to setup the shots fast saves me mucho dinero as I'm paying the models per hour. I can understand and appreciate that some people need a full bore cinema cam recording raw for their business needs, but for mine the vg-900 so far works well. My only complaint with it so far is that they should support full auto focus on a-mount lenses, that would let me use the same cam on another website as well.
 
photochris said:
Yawn. Same tired narrative. No credibility because once again he's trying to push his 'GH2 is superior to everything below the FS100' agenda, except it really isn't all that great.

With EOSHD everything always compared to the "superior detail" of the "baby Alexa" GH2 while constantly ignoring the terrible DR, huge amount of posterization in highlights and shadows that are complete bunk. Then there's the usual Canon bashing and Sony hate for everything that's not a cheap Nex or his FS. He trashed the VG20 and then heaped tons of praise on the Nex5n, yet the image each camera renders is identical.

Its really remarkable how many issues he has with cameras he doesn't like when others don't have the same problems. Its almost uncanny how he finds faults others can't replicate with properly exposed and properly focused shots. Almost unbelievable. My A99 is nowhere near as soft as the VG shot posted, but I don't have access to pre-prod GH3's and I don't have an axe to grind with Canon, Nikon and Sony et al.
Click to expand...

Do you really think he has an axe to grind? He loves the FS100 and the GH2/3, I agree. But he obviously loved the Canon enough to name his site EOSHD. Again he is a huge fan (and owner) of an FS100, I see no reason he would trash the VG900 since he already owns a Sony. Perhaps he has a flawed camera.

Perhaps Joker can upload some footage so we can see what good footage looks like. The fact of the matter is that Sony have done a terrible job convincing potential customers to pony up cash for the VG900.

All I want to see is 4-5 minutes of video that looks filmic. I'd also like to see footage captured to a Hyperdeck Shuttle via 4:2:2 HDMI. I could shoot all these tests for Sony in an afternoon. You'd think they'd have this footage available since they want to sell hardware.

And now that they don't have this stuff available, blogs like EOSHD are writing their narrative. That is not good marketing.
 
joker454 said:
I'm still trying to figure out how he managed to take a such a bad picture with the vg-900. I can't say that I'm able to replicate that bad of a look on mine. Maybe it's the lens he used? I use a Sigma 50mm f1.4 on mine. Aside from that, I think some of these guys need to understand that there are some out there like me that don't need the added complexity of a cinema cam and don't need and/or can't use raw support, yet at the same time need very good low light performance and very shallow dof. I did my first video shoot yesterday with the vg-900 for a new website I launched and it worked great, very simple and fast to use and I could take shots in low light with the full dof control I wanted. If I shot in raw then I'd have ~7tb/week of data to somehow manage on both magnetic and optical storage archives, it's just not a workable format for me, and being able to setup the shots fast saves me mucho dinero as I'm paying the models per hour. I can understand and appreciate that some people need a full bore cinema cam recording raw for their business needs, but for mine the vg-900 so far works well. My only complaint with it so far is that they should support full auto focus on a-mount lenses, that would let me use the same cam on another website as well.
Click to expand...

[TABLE="class: forumline, width: 100%"]
[TR]
[TD="width: 100%"][TABLE="width: 100%"]
[TR]
[TD="width: 80%"]He used Sony Alpha-Fullframe SAL 2470 E-Mount Adapter VG900
One terrible video of the sea with chromatic aberration 0:18
http://vimeo.com/52006567

info from here Sony Alpha-Fullframe SAL 2470 E-Mount
http://forum.slashcam.de/weiter-vt104143.html?postdays=0&postorder=asc&&start=35[/TD]
[/TR]
[/TABLE]
[/TD]
[/TR]
[/TABLE]
 
Last edited:
My GH2 with 50mm F1.4 FA Pentax @ F4 on Dec 16 2010 vs VG900 today with same lens in same location.
 
Last edited:
I just saw those on Vimeo and commented. What are your thoughts on the VG900?

Can you provide some raw footage to download so I can see something without web compression? I'd love to see skin tones and a wide shot with a lot of detail.
 
longcipher said:
Do you really think he has an axe to grind?
Click to expand...

Yes. He endlessly whines about what he can't get in a $1000 camera, no 10-bit, no 4k, low bit rates and so on - any camera that doesn't meet that non-existent standard gets bashed or relegated to "can't match the GH2." Read a few of his posts and get beyond the endless GH2 love. He says virtually the same thing in every post, nothing can match the GH2's detail, Canon is mushy, blah, blah, blah. He has a very narrow view of what he likes - and that's fine because its his little blog - but to make blanket statements about camera X being garbage and constantly trashing them in "comparisons" when others are not getting the same results just reeks of an agenda. Same goes for his VG900/A99 bashing, that shot he posted is not in focus or his lens is whacked because I get stunning sharpness out of my 24-70z, IQ and detail is pretty remarkable. IMHO he intentionally makes other cameras look bad to pimp the GH2 and now the Black Magic camera. His "tests" aren't worth the bandwidth they occupy.
 
I have seen a99 footage that is stunning, so I can't argue with you there. Perhaps the VG900 shot was user error -- it almost has to be to look that bad. If anyone reading this is close to Philadelphia, I'd love to meet up and check out the VG900.
 
It (the EOSHD test) is not user error.

I just uploaded 2 new files straight from the camera:
http://www.palsomedia.com/vg900
DSC00101_boat_bridge.JPG
00031_boat_bridge.MTS
The bad aliasing is the same whether it's 24p, 60i, or 60p.
I included the full JPG just to show that it's not an issue with the lens or focus.

The other files in the same location are from my previous posts.
 
You must log in or register to reply here.
Back
Top