OT: Camera Recommendations

Cliff Etzel wrote on 12/7/2006, 1:12 PM
I am in negotiations right now with a company to begin producing a series of demonstration video's for both web and DVD distribution. They are in the dive industry and since that is an area of focus for me, it is right up my alley.

The question I now need to ask what to do about cameras.

IT seems that with the new crop of SON Y camera's coming out, it seems the pricing on the current crop of cameras is coming down - since I will be shooting more than just this kind of stuff, I want a camera that can give me very good quality but still be in a small form factor - Even though there are light sensitivity issues with it from what I have read - the camera that has my interest is the A1U - it is similar in size to my TRV950 and gives the HD format I will want to begin working with this next year.

How high is the quality of the footage shot with this camera? Is it good enough for possible broadcast/stock footage?

When needed, I have lights to shoot with underwater already so that isn't an issue. How does it compare to the higher end Prosumer cameras (FX1/Z1U)? I look at the cost of the FX7/V1U and almost choke - remember, I need to obtain a housing for this camera and they cost more than the camera typically.

The price point is main contributing factor in my looking at A1U - Please - I need some real pro insights on this decision - 2007 is looking to be my breakout year and I want to do it right.

Cliff

Comments

FuTz wrote on 12/7/2006, 5:48 PM
A pro I've worked with yesterday said he shot recently (underwater with a housing) with the A1U and was very pleased with the results.
MH_Stevens wrote on 12/7/2006, 5:53 PM
I have the FX1 and love it, but if I was in the market again now I would wait for a 1080p offering.
Cliff Etzel wrote on 12/7/2006, 8:27 PM
What is the procedure for conversion of 1080i to progressive mode in Vegas?

As I stated earlier, the criteria is the cost factor - I work very lean in what gear I work with - Almost self contained video production - all gear must be carried by me both on land and in water - the A1U is a perfect size, especially since it fit's so well for my shooting style. I use a figrig when hand holding and have found it works very well. Other shooters I have talked with feel that anything bigger is tiresome after awhile.

If I could afford either the FX7 or V1U cameras, I would go with them, but as stated earlier - I shoot underwater as well - Any of the serious housings for shooting with HD START at around $3200 (You still have to get a port and external monitor for the housing - that can be as much as the housing - if I go A1U, the cost drops more in line with what I can afford next year. Housings for the HC1/A1U begin at $2000 - that extra cost is significant from my budgetary standpoint.

As it is, I still have a soft spot for my TRV950 and even the PDX10, both of which fit my $3100 housing, but both are obsolete from what everyone is saying - I personally would buy a PDX10 and shoot in DVCAM if I had my way and just use Red Giant's InstandHD to uprez to 720p HD, but it isn't perfect and the PDX10/TRV950 are only selling for a few hundred less than the A1U currently.

Any additional thoughts?
Avene wrote on 12/7/2006, 9:54 PM
Firstly, remember that the HC1/A1, and the newer FX7 and V1 cameras all use CMOS chips that suffer from the rolling shutter problem. I never knew what about the rolling shutter issue when I first bought my HC1. If I did, I most probably wouldn't have bought it!

The problem is CMOS chips. Even if they're progressive like the FX7/V1, they don't capture the entire frame at once. Forinstance, say if you do a fast pan accross a wall.. Instead of seeing a sharp 90 degree vetical edge as you should, you may see a 45 degree sloping edge! Because the camera samples the image from top to bottom. So by the time it gets around to the bottom part of the frame everything's in a different position. It's annoying. I've got a whole bunch of footage I'd taken from trains in the UK and Holland. The rolling shutter effect is quite visible, especially when another train goes by, or even a building or lampost. As soon as I can save enough money I'll be buying a Canon XH-A1. Good old reliable CCD chips on those!

Anyway, under normally circumstances it may not be noticible, but just knowing it's there is bad enough! Apart from that I think the HC1/A1U are great for under water photography. I've seen underwater footage taken with a HC1 that looks great.

As for the FX7 and V1, why Sony would include rolling shutter CMOS chips on semi pro cameras such as those is ridiculous. Sure, they may offer better resolution, but would be totally useless to anyway doing visual effects work that requires CGI footage integrated with live action.
Serena wrote on 12/7/2006, 11:47 PM
>>>suffer from the rolling shutter problem<<<

This would seem to be equivalent to a focal plane shutter on a still film camera and, to some extent, the rotating shutter on a cine camera. However with a cine camera the whole frame is exposed while the shutter segment is fully open, so I expect such motion distortions are much less obvious. Clearly you find them very noticable in the video version. How does shutter "speed" change the effect?
Avene wrote on 12/8/2006, 12:24 AM
Good question.. I've never tested that.
farss wrote on 12/8/2006, 1:21 AM
Hm,
interesting issue!

My understanding is that in a CCD device the charge from all pixels is transferred to the charge bucket at the same time, then one by one those buckets are transferred to the A/D converters.

In CMOS each charge from the photodiodes are transfered one by one to the D/A converters, hence not all pixels are read at the same time. No doubt this is the cause of the rolling shutter effect.

Thsi would seem to be similar to what happens in a film camera however as noted the effect maybe mitigated by averaging / motion blur to some extent.

I'd imagine in a CMOS camera faster shutter speeds would make the effect more pronounced?

Whatever it is two thoughts spring to mind.

Is this why the D20 uses a mechanical shutter?

We might have to get used to it, CMOS seems to be the way forward, SI-2K, RED etc.

Or is there a way to reduce the effect by altering the sequence in which the photodiodes are read, say from the bottom and the top towards the middle. I remember this issue coming up on a forum about some camera and a designer saying they'd changed something that reduced / fixed the problem.
Serena wrote on 12/8/2006, 3:54 AM
Bob, the record portion of the cycle has to be completed within the shutter time, so using a higher shutter speed reduces the possible motion that can occur between first and last pixel read out and hence the problem mentioned. However the accompanying lessening of motion blur introduces other problems.
farss wrote on 12/8/2006, 4:57 AM
Here's the bit that I don't get:

The actual sensor is a photodiode. It accumulates charge based on light hitting it. Lets say the shutter time is 1/48 at 24fps. Now I'd assume all the photodiodes need to be 'exposed' for the same period of time. I assume the process of reading them destroys the charge, so as soon as they're read each one start being exposed again.

Or is there some mechanism that I don't know about / understand that starts and stops them being 'exposed'?

I've yet to find a detailed 'how it works' explaination of these things.

Bob.

MH_Stevens wrote on 12/8/2006, 6:33 AM
.............."What is the procedure for conversion of 1080i to progressive mode in Vegas?"

Your 1080i will could end up on DVD as standard DV or in WMV9/DivX/QT etc. at 720p.

Cliff Etzel wrote on 12/8/2006, 6:35 PM
I spent som etime in the dvinfo forums just reading how many people still love the TRV950/PDX10 cameras!

As much as many have jumped on the HDV bandwagon, many still swear by these cameras. But some state that the closest replacement for the PDX10 is the A1U - go figure ;-)

I guess the biggest gripe about the A1U is the fact that the tape loads from the bottom! Who was the Einstein who came up with that design feature??? That precluded using the thing on a tripod unless you have a special spacer for crying out loud!

I guess there is no perfect camera, but I am pretty much relegated to SONY's cameras due to the LANC connector for the electronic housings and the word isn't for sure on the Canon A1 yet for housings either - and only one serious housing company is contemplating a housing for the camera...
Steve Mann wrote on 12/11/2006, 1:09 AM
Your first explanation is pretty accurate:

"My understanding is that in a CCD device the charge from all pixels is transferred to the charge bucket at the same time, then one by one those buckets are transferred to the A/D converters.

In CMOS each charge from the photodiodes are transfered one by one to the D/A converters, hence not all pixels are read at the same time. No doubt this is the cause of the rolling shutter effect."

The CCD basically starts an exposure sequence with an identical charge on each photoreceptor site. Then for the duration of the "shutter" the charge is allowed to discharge. The more light, the faster the CCD photo site discharges. At the end of the "shutter" period, the remaining charge (ironically, an analog value) is dumped into a buffer on the A/D converter wher the processor scans the digitized value for each photosite.

The CMOS detector simply converts the amount of light on the receptor site to an analog voltage that the processor samples sequentially.


Steve Mann