|
|||||||||
|
Thread Tools | Search this Thread |
September 9th, 2009, 06:12 PM | #1 |
Regular Crew
Join Date: May 2006
Location: Scranton, PA.
Posts: 91
|
Vertical smear in fireworks display
I'm in the middle of doing a review of the HM100 and I'm sure most are aware of the vertical smear issue. It's very prevalent in a fireworks display as well as the flare associated with it. It's almost unusable for fireworks displays. The burst is exaggerated white circles with extreme vertical extensions. Here are some examples.
__________________
"There are 10 types of people in the world: Those that understand binary, and those that don't!" |
September 9th, 2009, 08:23 PM | #2 |
Major Player
Join Date: Nov 2006
Location: Brisbane, California
Posts: 530
|
This is a pretty common occurrence with CCDs, unfortunately, it happens on my HM100 as well as my other CCD camcorders. I assume it's possible for a CCD not to exhibit this problem as some higher end camcorders don't seem to have this problem.
|
September 9th, 2009, 10:10 PM | #3 |
Major Player
Join Date: Sep 2007
Location: Bellingham, WA
Posts: 457
|
I would go beyond just CCD issue, but I would say it's a problem with entire digital technology, DLSR included. Simply this is one thing that can't compete with film- dynamic range.
|
September 10th, 2009, 10:59 AM | #4 |
Inner Circle
Join Date: Dec 2005
Location: New York City
Posts: 2,650
|
The white circles are in the lens, the vertical smears are in the CCDs. Extreme light will cause some form of smear on all chip pick-up devices, CCD or CMOS. It's a reaction to having too much energy from the light and the sensors are unable to distribute it correctly. The larger the chip's physical size the less smear. Expensive HD camera rarely have this problem as the light source would have to be unusually bright to cause the problem. Turning the light sensitivity up will increase the smear.
Quite frankly, while the smears on your stills are bad they are a lot better than what you would have gotten a decade ago.
__________________
William Hohauser - New York City Producer/Edit/Camera/Animation |
September 10th, 2009, 04:48 PM | #5 |
Major Player
Join Date: Nov 2006
Location: Brisbane, California
Posts: 530
|
I find CMOS overdriven has a much less objectionable smear than the vertical bars on CCDs. It is more of an overall bloom than a definite artifact.
|
September 12th, 2009, 07:44 PM | #6 |
Inner Circle
Join Date: Dec 2005
Location: New York City
Posts: 2,650
|
This due to the way the sensors are arranged on CMOS chips, one of the pluses of a CMOS chip and one of the reasons why they are extensively used in still cameras. However, right now, CCDs are better for video camera operation.
__________________
William Hohauser - New York City Producer/Edit/Camera/Animation |
September 12th, 2009, 10:51 PM | #7 |
Regular Crew
Join Date: Apr 2008
Location: Echuca, Victoria, Australiamate
Posts: 179
|
Being somewhat old fashioned, my Sony DSR-570 uses 2/3 inch power HAD chips which have a -123db vertical smear ratio... so I dont suffer these problems.
Its not really the idea of CCDs, its whats DONE to the CCD in the first place that reduces the smear. Mind you, you get what you pay for. Ben |
September 12th, 2009, 10:53 PM | #8 |
Regular Crew
Join Date: Oct 2007
Location: Foothill Ranch, CA
Posts: 166
|
I rented a Panasonic HMC150 with CCD's this July 4th. I used many different manual exposure settings and I got very similar results as you did. I finally got some good shots by lowering the exposure (f6.5, or f8, fast shutter, little or no gain) to get it to look good.
|
September 13th, 2009, 01:28 AM | #9 |
Wrangler
|
The vertical smear on CCD sensors is due to excessive light spilling into successive columns as they are read out. As Ben says, better cameras don't exhibit this problem. Mine certainly doesn't (PDW-F350).
It's not the arrangement of the pixels, but how they're read out of the chip that makes CMOS less prone to vertical smear. Andy's tip is a solid one, in that you should stop down because the fireworks are much brighter than you think. In your screen shots, your brights are very blown out. Not to mention, you'll likely be at a long distance and zooming in will give you a shallower depth of field if you open the iris up. Stopping down helps keep all the fireworks in focus since they won't all be at the same distance from your camera. -gb- |
September 13th, 2009, 03:41 PM | #10 |
New Boot
Join Date: Apr 2007
Location: Heidelberg, Germany
Posts: 7
|
did some test shooting 2 weeks ago...
....like Andy said: lowering the exposure + no gain helped a lot.
|
September 14th, 2009, 09:34 AM | #12 | |
Inner Circle
Join Date: Dec 2005
Location: New York City
Posts: 2,650
|
Quote:
I have seen light smears on older CMOS chips but they tend to be very faint and more of a halo. Also the fact that most CMOS chips grab a frame using a rolling shutter scan progression versus the single instantaneous frame capture of CCDs helps eliminate light spill (but introduces the dreaded "jello").
__________________
William Hohauser - New York City Producer/Edit/Camera/Animation |
|
September 15th, 2009, 05:33 AM | #13 |
Regular Crew
Join Date: Apr 2008
Location: Echuca, Victoria, Australiamate
Posts: 179
|
|
March 22nd, 2011, 03:35 AM | #14 | |
Regular Crew
Join Date: Mar 2004
Location: Upper Pittsgrove, NJ
Posts: 95
|
Re: Vertical smear in fireworks display
Quote:
A CCD has two major components: the photo sensors themselves, which accumulate an electric charge proportional to the light on the device, and the charge-coupled array itself. When in exposure mode, the photo diodes are charging the CCD. When in read-out mode, the CCD acts like a "bucket brigade", a big analog shift register. One cell dumps its charge to the next, and so on, until finally the charge gets to an ADC (analog to digital converter). The smear happens because, with a very bright light source, some charge is still leaking from the sensor to the charge array during read-out. There are a couple of ways around it. One way is to use interline CCDs, In an interline CCD, you actually have two rows of charge coupled memory. Charge is acquired by one cell, then coupled to a parallel cell, and then shifted out. This is also makes for a faster sensor, as the sensor can be acquiring charge and shifting out the previous charge at the same time. But the only way to eliminate smear entirely is to use a physical shutter, rather than an electronic shutter. This is why you don't see smear in photos from the CCD-era of DSLRs... all DSLRs have real shutters. Of course, you can also use a CMOS sensor, which does eliminate smear, but usually introduces other issues, primarily the effect of the usual rolling shutter. And curiously enough, this could also be fixed using a global shutter (electronic, this time), but presumably causes other problems in the CMOS device, like power surges. Yup. The addition of interline devices, and just decades of making CCDs with better isolation between sensor and readout have improved things quite a bit.
__________________
--Dave |
|
| ||||||
|
|