HVX200 BVW400A BVW600 Mini Shoot-Out [no pics/no footage]

I wonder how the XDCAM will fair. I've already got my HVX on order and woudln't mind picking up an XD if it lives up to what it is on paper.
 
Thanks NBC,
This was no doubt a torture test.

I'm sure we all would of agreed the 2/3" cams were going to smoke the
HVX in the noise department.
 
just hink what a new 600 costed in its day. they still make great images and if the HVX only gets near it in SD I am happy.
 
NBCshooter said:
When lit well, the HVX had extremely beautiful colorimetry, and a very naturalistic characteristic curve. In side-by-side comparisons, the HVX' overall image quality (aside from resolution) seemed very comparable to the BVW600's colorimetry.
Click to expand...
And about the PDW-F330? I'm thinking to buy one but I've doubts regarding a comparision in terms of film-like color properties.

What do you think NBCshooter: HVX rules? Or the Sony's noiseless synthetic colors wins?
 
NBCshooter said:
Illuminated at 20 footcandles (need to confirm this with another light meter--may have been miscalibrated), with all cameras irised ar their arbitrary "starting point" apertures above, the HVX appeared significantly noisier than the 2/3" cameras. I opened each camera's iris by one full stop. Still, the HVX was noticably noiser than the other cameras. I opened up one more full stop. The noise in the HVX picture virtually disappeared.
Click to expand...

Thanks NBCshooter, the BVW400a is a camera I can relate to, so it's very impressive to think that a little $6,000.00 "Toy" camera is nearly as sensitive and produces a sharper image to boot. Wow. I'm going to be trying to convince a client to ditch the 400a on an upcoming greenscreen shoot and embrace progressive and P2, so this was timely.

I'm only confused about your noise test... when you said you opened up two stops on the HVX and the noise went away, what did that mean to the picture? Wasn't it very over-exposed at that point? Or were you not starting from a nominal grey card exposure for that set-up? Or maybe you were adjusting shutter speed to compensate? I guess I don't know what you mean by "arbitrary starting point".

Thanks, great work!

Jim Arthurs
 
NBCshooter said:
When I turned the iris ring on the HVX one more time to f/2.8, the exposure was then more "normal," and the noise became virtually non-existant, or at least very comparable to the 2/3" cameras.
Click to expand...

Thanks, that completely clears what you did! Underexposing the HVX is a bad thing. But, this brings up another question for me, and that is "why?".

Look at it this way; let's say you're at your first test settings, where you were at f8 with the HVX and everything was very dark. Without adjusting the camera, what would happen to that exact same setup if you pounded enough extra light into the scene to bring it up to a proper exposure for that fstop? Would the noise go away, I wonder? Seems like it would, based on your tests and what others have said.

This is puzzling to me and my film background. If I underexpose film, and make no attempt to bring up the levels at the transfer or in printing, the grain stays the same as it would when properly exposed. You just have a dark image. Of course if you "print up" or attempt to normalize you'll introduce grain, but that's not what I'm talking about.

Am I crazy here? Am I missing something basic about ccd imagery? Does every shot have to have an area of "normal" exposure to drive the noise demons away? What happens to the areas of the image that are 2-3 stops below middle grey? Are they always filled with noise, or just when the overall image has no proper points of exposure? Help!!!

Jim Arthurs
 
Jim Arthurs said:
Thanks, that completely clears what you did! Underexposing the HVX is a bad thing. But, this brings up another question for me, and that is "why?".

Look at it this way; let's say you're at your first test settings, where you were at f8 with the HVX and everything was very dark. Without adjusting the camera, what would happen to that exact same setup if you pounded enough extra light into the scene to bring it up to a proper exposure for that fstop? Would the noise go away, I wonder? Seems like it would, based on your tests and what others have said.

This is puzzling to me and my film background. If I underexpose film, and make no attempt to bring up the levels at the transfer or in printing, the grain stays the same as it would when properly exposed. You just have a dark image. Of course if you "print up" or attempt to normalize you'll introduce grain, but that's not what I'm talking about.

Am I crazy here? Am I missing something basic about ccd imagery? Does every shot have to have an area of "normal" exposure to drive the noise demons away? What happens to the areas of the image that are 2-3 stops below middle grey? Are they always filled with noise, or just when the overall image has no proper points of exposure? Help!!!

Jim Arthurs
Click to expand...
Jim:

Believe me, I appreciate your confusion. This has been turning out ot be a very interesting exercise. I've been asking myself the EXACT same questions. And no, your conclusion is NOT true (thank god). It seems that when given a low contrast scene, the signal processing electronics in any CCD or CMOS based imaging system tends to generate more noise. I'm not an engineer, I don't know the reasons for this.

For my SciFest short, I shot a VERY high contrast, low key scene with my very noisey DVX100 (non-A). In that scene, more than 95% of the frame was in shadow or in near-black IRE values. But the subjects were heavily backlit and the scene had a lot of contrast. Guess what? NO NOISE! The CCD and the DVX's signal processing electronics saw a "decent" IRE spike SOMEWHERE in the CCD and figured, "Hmmm, this DP has quite a lot of light on one SMALL piece of my CCD, therefore, I won't act so noisy on his near-black parts of the CCD."
 
NBCshooter said:
It seems that when given a low contrast scene, the signal processing electronics in any CCD or CMOS based imaging system tends to generate more noise.
Click to expand...
I'll give you an example. I'll take my DVX100 or VX2000 and shoot blue sky in the daytime. I'm at f/11 with ND on, so light level is NOT the issue. But the contrast in the scene is very low. Guess what? LOTS of noise!
 
I might sound stupid, but is there a ISO setting on the camera ?
I know on my Canon 20D noise come from high ISO when in low light. So could it be the CCD ISO is changed by the processor to different levels ?

It the only thing that come to mind. High ISO increase the NOISE.

It pretty much the only way I get visible noise, because if my scene is dark, well it only dark at low iso, no noise.

But yes the Processor could be the one increasing the brightness... But I would think it might be more the chip telling the CCD to change it ISO level.

Out of context : It could be nice to have a 'RAW' mode like in digital camera, where it record without color correction but save in the metatag this info, then when opening in a editor, the editor apply the filter ;) Maybe in the next version
 
mochouinard said:
I might sound stupid, but is there a ISO setting on the camera ?
I know on my Canon 20D noise come from high ISO when in low light. So could it be the CCD ISO is changed by the processor to different levels ?

It the only thing that come to mind. High ISO increase the NOISE.

It pretty much the only way I get visible noise, because if my scene is dark, well it only dark at low iso, no noise.

But yes the Processor could be the one increasing the brightness... But I would think it might be more the chip telling the CCD to change it ISO level.

Out of context : It could be nice to have a 'RAW' mode like in digital camera, where it record without color correction but save in the metatag this info, then when opening in a editor, the editor apply the filter ;) Maybe in the next version
Click to expand...
My wild-azz, totally uninformed guess would be that the signal processing electronics account for this behavior. The actual response characteristics of the CCD imager may also play a role, but I know nothing about CCD design with respect to this subject. Video cameras don't have "ISO" settings, they have "gain" settings. The response curve behavior is beyond the control of the external gain setting on the camera, and is something internal to the camera's processing electronics, and possibly the imager.

The "raw" mode equivalent of a video camera would be similar to an SDI out, which some pro cameras offer as a board option. With this option, you could record 10-bit, 4:2:2 uncompressed video to a disk array from a camera like a DSR450. The Sony HDW-F950 is capable of recording out 4:4:4, 10-bit, 24psf, 1080i HD video out to a disk array--that's about as close to "raw" as you get in video.
 
NBCshooter said:
Jim:

Believe me, I appreciate your confusion. This has been turning out ot be a very interesting exercise. I've been asking myself the EXACT same questions. And no, your conclusion is NOT true (thank god). It seems that when given a low contrast scene, the signal processing electronics in any CCD or CMOS based imaging system tends to generate more noise. I'm not an engineer, I don't know the reasons for this.

For my SciFest short, I shot a VERY high contrast, low key scene with my very noisey DVX100 (non-A). In that scene, more than 95% of the frame was in shadow or in near-black IRE values. But the subjects were heavily backlit and the scene had a lot of contrast. Guess what? NO NOISE! The CCD and the DVX's signal processing electronics saw a "decent" IRE spike SOMEWHERE in the CCD and figured, "Hmmm, this DP has quite a lot of light on one SMALL piece of my CCD, therefore, I won't act so noisy on his near-black parts of the CCD."
Click to expand...

Yea thats totally true. Barry mentioned this a few days ago, and i never even noticed it until he says so. Then i went out, shot some stuff in the dark with a few areas Properly lit ( like above 50 or 60 ire, noone says you need 99IRE t ge the chips to run "quiet") and suddenly the image just Pops! its really interesting.

ITs like finding a new tool to use for making low light stuff look good.
 
Just to clarify the statements I made previously regarding the sensitivity of modern 2/3" cameras, I happened to see my friend at Inside Edition the other day. I asked him how he compared the sensitivity of their old BVW400As with their new PDW530 XDCAMs. He did exactly as I did. He told me he set up both cameras on the 18% gray part of a chip chart. He irised both cameras at 50% IRE, and the PDW530's iris was THREE STOPS more closed than the 400A's!

I forgot to ask him about 24P mode (I assume you lose one stop in 24P on the 530). So at the very least, that would mean that a PDW530-class camera (cameras using the same chip block and similar processing electronics) such as the DSR450, would be about two stops more sensitive in 24P mode than an HVX200.
 
Interesting tests.
It would have been interesting to compare to the industrial Sony's - the D30 & 500 series as well as the 1/2" 300 series. I remember when the D30 came out it had substantially more noise than the BVW400 which it pretty much supplanted in common use. You could shoot at 9db on the 400 and never see the grain, but 9db on the D30 was golf balls. So the 400 & 600 were notably quiet cameras esp the 600.

I too wonder about the noise vs exposure thing. Haven't tested my HVX much yet, but it looked like what you've described on stuff I've looked at.
Possibly this is a monitor issue as well. I feel like I've always seen this on video to some extent and even in film we always used to say get solid exposure somewhere in the frame too avoid the impression of swimming noise, but it doise lok more pronounced on the HVX.
Wonder if this is a codec issue with DVCPro codec.

One thing about the HVX in low light progressive is that you can drop the shutter speed to 1/24, which gives you an extra stop with only minimal blur. I actually prefer 1/30 to 1/48 which also gves you more stop.
 
Although film sensitometry is quite different from CCDs, and the response curve of their related signal processing electronics, I remember when Kodak first introduced their 5294 color negative stock in the early 1980's, that a DP I read about in American Cinematographer recommended that, if he didn't throw at least 5-10 footcandles of light in the shadows [when shooting with 5294], that that they would go "milky." He said something to the effect that, throwing a few footcandles of light in the shadows helped it go jet black.
 
x2mag68 said:
Yea thats totally true. Barry mentioned this a few days ago, and i never even noticed it until he says so. Then i went out, shot some stuff in the dark with a few areas Properly lit ( like above 50 or 60 ire, noone says you need 99IRE t ge the chips to run "quiet") and suddenly the image just Pops! its really interesting.

ITs like finding a new tool to use for making low light stuff look good.
Click to expand...
Well, that's comforting to know that an informed guy like Barry had also come to similar conclusions. Could you post a link to the thread where he mentions this? Since you also have confirmed this observation, now at least I know, I'm not "imagining" this phenomenon! Thanks for the info!
 
oh jeez i just came across it. I think it was in a HVX200 thread. I cant even remeber where isaw it.
 
You must log in or register to reply here.
Back
Top