Question on Kx or K7 video

osv · 02-27-2010, 06:59 PM

you have to look at the full-size moving image gif, in the link i posted beneath the red-x'ed image, to see how bad it really is.

as i understand it, the point here is that the vdslr is not recording and downsampling off of the entire sensor, it's only recording and downsampling off of a portion of the sensor... the full sensor size is used, but a bunch of it is totally skipped for video use.

compare that to a camcorder sensor, where no portion of the sensor is skipped... the result can be upwards of 33% higher resolution than a vdslr, on a camcorder that costs only a few hundred dollars.

that's why vdslr technology is so immature... it's a rather half-assed compromise between optimization for stills, and technology limitations for video use.

jogiba · 02-27-2010, 07:29 PM

Originally posted by osv

you have to look at the full-size moving image gif, in the link i posted beneath the red-x'ed image, to see how bad it really is.

as i understand it, the point here is that the vdslr is not recording and downsampling off of the entire sensor, it's only recording and downsampling off of a portion of the sensor... the full sensor size is used, but a bunch of it is totally skipped for video use.

compare that to a camcorder sensor, where no portion of the sensor is skipped... the result can be upwards of 33% higher resolution than a vdslr, on a camcorder that costs only a few hundred dollars.

that's why vdslr technology is so immature... it's a rather half-assed compromise between optimization for stills, and technology limitations for video use.

Wow, a camcorder for a few hundred dollars has 33% higher resolution than a vdslr ! Maybe they should have gone to Walmart and picked up a cheap camcorder to make Uncle Jack.

[yt]http://www.youtube.com/watch?v=Vez7qSrZfpc[/yt]

[yt]http://www.youtube.com/watch?v=tL86qxCasoI[/yt]

16N64 · 02-28-2010, 11:24 AM

Originally posted by osv

as i understand it, the point here is that the vdslr is not recording and downsampling off of the entire sensor, it's only recording and downsampling off of a portion of the sensor... the full sensor size is used, but a bunch of it is totally skipped for video use.

You made up this assumption all by yourself or is it based on facts?

osv · 02-28-2010, 12:43 PM

Originally posted by 16N64

You made up this assumption all by yourself or is it based on facts?

why do i have to babysit your ignorance of modern technology?

Re: Update considering Canon's claim of 6 pixel binning: Open Talk Forum: Digital Photography Review

video from DSLRs: every third line used? - Luminous Landscape Forum

16N64 · 02-28-2010, 01:47 PM

Originally posted by osv

why do i have to babysit your ignorance of modern technology?

Re: Update considering Canon's claim of 6 pixel binning: Open Talk Forum: Digital Photography Review

video from DSLRs: every third line used? - Luminous Landscape Forum

Sorry, no facts in these links, just guessing.

Averaging rectangular portions of the sensor surface could cause some aliasing, as it is not a very good digital anti-aliasing filter for decimation, but I don't want to participate in further guessing here.

Stefan

osv · 02-28-2010, 02:47 PM

Originally posted by 16N64

Sorry, no facts in these links, just guessing.

sorry, but you aren't paying attention.

as was pointed out in those links, a canon employee has already claimed that their vdslrs use binning to make up for the inability of the camera to read all of the sensor data in a timely manner.

which proves my point entirely.

16N64 · 03-03-2010, 02:17 PM

Originally posted by osv

sorry, but you aren't paying attention.

as was pointed out in those links, a canon employee has already claimed that their vdslrs use binning to make up for the inability of the camera to read all of the sensor data in a timely manner.

which proves my point entirely.

Pixel binning usually means to combine pixels before the ADC, which can have some advantage over averaging in the digital domain, one of them some decreased digital bandwidth and less burden on the signal processing. But there is no information on how this binning is done or what happen digitally afterwards, so there is no information on which pixels and if any are skipped for video.

Stefan

osv · 03-03-2010, 03:34 PM

Originally posted by 16N64

But there is no information on how this binning is done or what happen digitally afterwards, so there is no information on which pixels and if any are skipped for video.

horrible aliasing and low resolution is what happens digitally afterwards.

i suppose that you have a pet theory for why a big dslr sensor provides lower resolution than a small camcorder sensor does, since it can't possibly be pixel skipping or binning.

Coby · 03-03-2010, 10:39 PM

The K-7 and K-x clearly aren't using all the pixels available on the sensor - there was a post on the DPReview Pentax DSLR Forum about the inability of the current chip to process the full data at an acceptable framerate. Google would probably locate it.

Anyway, interestingly, my small camercorder never manages to produce this kinda effect (which I have to say is groovy!): Lumolab K-x - Falk Lumo's Photos

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
KX Video Transfer Question	Zulu209	Troubleshooting and Beginner Help	2	10-14-2010 03:30 PM
Pentax K-X video question	Nflguy33	Video Recording and Processing	6	04-14-2010 02:34 PM
question K7 vs. Kx video	Mystic	Video Recording and Processing	7	01-04-2010 09:14 PM
k7 video question, a/v not in sync on pc	allill	Pentax DSLR Discussion	7	12-24-2009 05:30 PM
Pentax K7 Video Question	Spock	Pentax DSLR Discussion	3	08-10-2009 05:37 AM

02-27-2010, 06:59 PM	#16
osv Veteran Member Join Date: Mar 2007 Location: So Cal Photos: Gallery Posts: 2,080	you have to look at the full-size moving image gif, in the link i posted beneath the red-x'ed image, to see how bad it really is. as i understand it, the point here is that the vdslr is not recording and downsampling off of the entire sensor, it's only recording and downsampling off of a portion of the sensor... the full sensor size is used, but a bunch of it is totally skipped for video use. compare that to a camcorder sensor, where no portion of the sensor is skipped... the result can be upwards of 33% higher resolution than a vdslr, on a camcorder that costs only a few hundred dollars. that's why vdslr technology is so immature... it's a rather half-assed compromise between optimization for stills, and technology limitations for video use.

02-27-2010, 07:29 PM	#17
jogiba Banned Join Date: Sep 2006 Location: Millstone,NJ Posts: 6,491	Originally posted by osv you have to look at the full-size moving image gif, in the link i posted beneath the red-x'ed image, to see how bad it really is. as i understand it, the point here is that the vdslr is not recording and downsampling off of the entire sensor, it's only recording and downsampling off of a portion of the sensor... the full sensor size is used, but a bunch of it is totally skipped for video use. compare that to a camcorder sensor, where no portion of the sensor is skipped... the result can be upwards of 33% higher resolution than a vdslr, on a camcorder that costs only a few hundred dollars. that's why vdslr technology is so immature... it's a rather half-assed compromise between optimization for stills, and technology limitations for video use. Wow, a camcorder for a few hundred dollars has 33% higher resolution than a vdslr ! Maybe they should have gone to Walmart and picked up a cheap camcorder to make Uncle Jack. [yt]http://www.youtube.com/watch?v=Vez7qSrZfpc[/yt] [yt]http://www.youtube.com/watch?v=tL86qxCasoI[/yt]

02-28-2010, 11:24 AM	#18
16N64 Inactive Account Join Date: Nov 2009 Location: Former capitol of Germany Posts: 63	Originally posted by osv as i understand it, the point here is that the vdslr is not recording and downsampling off of the entire sensor, it's only recording and downsampling off of a portion of the sensor... the full sensor size is used, but a bunch of it is totally skipped for video use. You made up this assumption all by yourself or is it based on facts?

02-28-2010, 12:43 PM	#19
osv Veteran Member Join Date: Mar 2007 Location: So Cal Photos: Gallery Posts: 2,080	Originally posted by 16N64 You made up this assumption all by yourself or is it based on facts? why do i have to babysit your ignorance of modern technology? Re: Update considering Canon's claim of 6 pixel binning: Open Talk Forum: Digital Photography Review video from DSLRs: every third line used? - Luminous Landscape Forum

02-28-2010, 01:47 PM	#20
16N64 Inactive Account Join Date: Nov 2009 Location: Former capitol of Germany Posts: 63	Originally posted by osv why do i have to babysit your ignorance of modern technology? Re: Update considering Canon's claim of 6 pixel binning: Open Talk Forum: Digital Photography Review video from DSLRs: every third line used? - Luminous Landscape Forum Sorry, no facts in these links, just guessing. Averaging rectangular portions of the sensor surface could cause some aliasing, as it is not a very good digital anti-aliasing filter for decimation, but I don't want to participate in further guessing here. Stefan

02-28-2010, 02:47 PM	#21
osv Veteran Member Join Date: Mar 2007 Location: So Cal Photos: Gallery Posts: 2,080	Originally posted by 16N64 Sorry, no facts in these links, just guessing. sorry, but you aren't paying attention. as was pointed out in those links, a canon employee has already claimed that their vdslrs use binning to make up for the inability of the camera to read all of the sensor data in a timely manner. which proves my point entirely.

03-03-2010, 02:17 PM	#22
16N64 Inactive Account Join Date: Nov 2009 Location: Former capitol of Germany Posts: 63	Originally posted by osv sorry, but you aren't paying attention. as was pointed out in those links, a canon employee has already claimed that their vdslrs use binning to make up for the inability of the camera to read all of the sensor data in a timely manner. which proves my point entirely. Pixel binning usually means to combine pixels before the ADC, which can have some advantage over averaging in the digital domain, one of them some decreased digital bandwidth and less burden on the signal processing. But there is no information on how this binning is done or what happen digitally afterwards, so there is no information on which pixels and if any are skipped for video. Stefan

03-03-2010, 03:34 PM	#23
osv Veteran Member Join Date: Mar 2007 Location: So Cal Photos: Gallery Posts: 2,080	Originally posted by 16N64 But there is no information on how this binning is done or what happen digitally afterwards, so there is no information on which pixels and if any are skipped for video. horrible aliasing and low resolution is what happens digitally afterwards. i suppose that you have a pet theory for why a big dslr sensor provides lower resolution than a small camcorder sensor does, since it can't possibly be pixel skipping or binning.

03-03-2010, 10:39 PM	#24
Coby Senior Member Join Date: Dec 2009 Posts: 106	The K-7 and K-x clearly aren't using all the pixels available on the sensor - there was a post on the DPReview Pentax DSLR Forum about the inability of the current chip to process the full data at an acceptable framerate. Google would probably locate it. Anyway, interestingly, my small camercorder never manages to produce this kinda effect (which I have to say is groovy!): Lumolab K-x - Falk Lumo's Photos