Is CF30 really 1/2 resolution?? at DVinfo.net

Brian Broz · January 20th, 2005, 01:53 PM

Hi,
I've been testing/shooting sample footage with an FX-1 viewing on a Sony LUMA HD monitor (via HD analog component). I have used the DVX100A and PD170 a lot so I have a pretty good idea what the different camera's advantages/weaknesses are.

Form the tests I have shot so far, I really can't see a 50% loss in resolution that people are claiming (in CF30 mode)...viewing in HD, not SD/DV! Yes CF24 seems pretty useless IMHO. For example, when shooting with a PD150/170 you see a noticeable softness/loss of resolution in 1/30 shutter mode...and the introduction of stairstepping and "rough" edges. To me, CF30 is basically indistinquishable from the standard 60i mode.

With the FX-1 I really can't see the an appreciable loss of res in CF30 (viewing on a $3000.00 Sony pro monitor). On relatively wide shots I don't get any stair-stepping and can still read incredibly fine details/print that of course I can see in 60i/std interlace mode. Again, this is with a locked-off shot.

Does anyone else have any experience or care to comment? It's possible I need to have my eyes checked :-) but to me the "loss" in resolution is negligable.

Hoping for some input everyone,

Thanks,

Brian Broz
www.llsr.com

Chris Hurd · January 20th, 2005, 02:26 PM

Howdy from Texas,

Well, I have a comment, Brian.

<< I really can't see a 50% loss in resolution that people are claiming (in CF30 mode) viewing in HD, not SD/DV! ...to me the "loss" in resolution is negligable. >>

Regarding resolution loss in Sony's CineFrame mode (or Panasonic's standard definition Frame mode found in Canon and Panasonic SD camcorders), my comment is "so what." I firmly believe that the proof is in what you see. If the the loss in resolution appears to be negligable, then it's no big deal. Too many people tend to get hung up on the number of pixels they're dealing with, and it seems a large portion of these folks have never actually seen the results; they're just going by the numbers.

Are *you* happy with what you see -- and more importantly, is your intended audience (your business clients) happy with what they see. The answer is most often yes, it looks great. So what if there is a loss of res. Can you see it? Does it make a huge difference in image quality so much as to make it unusable? Can you create viable, productive material with sufficient quality? Almost all the time, the answer is yes. I personally know scores of people who shoot in nothing but Frame mode all the time with Panasonic and Canon camcorders -- the real number is probably a significant percentage of all owners of these camcorders -- who are more than pleased with the results they're getting.

Resolution loss in CF mode? So what -- the important thing is, are you happy with the image. That's all that matters. Some folks tend to be too consumed with numbers to realize how good they've got it with these pseudo-progressive modes.

Brian Broz · January 20th, 2005, 02:52 PM

Well said Chris,

I agree completely...However, I would like to know simply for peace of mind:-)

For example if a project is going to be projected onto a large screen, or heavily tweaked/color-corrected in post...it may affect my decision of what mode to shoot in.
Basically I think I need to get a hold of a resoltion chart and see for my self. Personally I don't care if I can see a loss in res zoomed at 200 or 400% of the frame, simply at 1440x1080 res.

I am really surprised to read all the trashing and nit-picking of the FX1/Z-1. Of course it's not a Varicam or HDcam...but it's groundbreaking for the money (again viewed in HD). All the BetaSp and Digibeta shooters (that have also shot mini-dv) have seen our test footage on the Luma monitor were both surprised and impressed with the image.

Well, does anyone know where I could find a download-able resolution chart?

Many thanks,

Brian Broz

Barry Green · January 20th, 2005, 04:22 PM

First, what Chris says is exactly true -- if you like what you see, so what?

But since you ask: there is a loss, but it's nowhere near 50% as compared to the 60i mode.

As tested by a resolution chart, 60i delivers about 775 lines, CF30 delivers about 575.

Brian Broz · January 21st, 2005, 01:46 AM

Hi Barry,
I have read many of your posts regarding the DVX and appreciate all the time you put on these boards.
I have a question (which feels like a dumb question) regarding the resolution. How does the resolution of HDV become 775 lines? That doesn't sound like much...especially considering the images I've seen.
I'm not questioning your findings, I'm curious because most DVCAM or DVCPRO (2/3") camera heads shoot 800-850 lines before the signal is recorded. So higher resolution than hidef...doesn't seem right, or am I missing something obvious?
Many thanks for any response,

Brian
HDV Fan

Toke Lahti · January 21st, 2005, 03:13 AM

I'm also little a bit confused about these resolutions.
How can interlaced resolution be so high as 775 lines?
If there are details less than 2 lines high, aren't they having interlace flickering (25/30Hz)?
If details shorter than 2 lines are low pass filtered, wouldn't the resolution me more like 540?

Barry Green · January 21st, 2005, 03:29 AM

The specs you'll see about a broadcast camera's head resolving 800 or 850 lines or whatever refers first of all to "TV Lines", not scan lines, and secondly, that's talking about the camera head, not the recording format. The camera head may be capable of resolving 850 lines, but by the time it gets recorded on DV/DVCAM/DVCPRO/DigiBeta tape, the absolute maximum possible is 540 lines. That's a limitation of the recording medium, regardless of what the camera head is putting out.

The FX1 can resolve and record somewhere around 850 lines horizontally, and about 775 lines vertically, as measured on a resolution chart.

Now, 775 sounds low, when talking about a format that's supposed to have 1080 pixels, right? Well, that's where Toke's observation comes in, about interlacing and field-combining and low-pass filtering. Steve Mullen postulated that the FX1's maximum vertical resolution should theoretically be 820 lines, as he calculated that a field shouldn't be capable of resolving more than 410 lines. When I observed 775, I figured that's close enough to his 820 to call it "close enough".

Consider that a normal DV camera has ostensibly up to 480 scan lines of resolution, but in observed practice about the most you can get resolved is around 360-380. In progressive-scan mode you can actually get all 480 on a DVX or XL2 in thin-line detail. But in interlaced, those same cameras will max out at about 360 to 380.

So, 775 may not sound as promising as 1080 would make you think, but it's still around double what an SD camera can do.

Concerning the 540/575 lines of CineFrame 30, I measured 575 on a resolution chart. I don't know exactly how the camera accomplishes this; you'd think that if it was a straight de-interlace of the interlaced signal, you'd get around 400 lines (depending on whether my observed 775 or Steve's theoretical 820 were the more accurate number). However, it's clearly much better than 400, it's around 575. That leads me to wonder if it isn't just a straight raw field right off the 1080 CCD, which would give it 540 resolved scan lines. 540's close enough to what I saw (575) to make me think that that's probably what it's doing.

Betsy Moore · January 21st, 2005, 05:35 PM

I agree the proof is in the pudding--but in the case of any frame mode I've ever seen the allegorical pudding is terrible to my eyes. And just in case you think my opinion might arise from a numbers bias, the first time I ever saw a Canon frame mode I assumed it had the same resolution (or higher) before I looked at it, then when I looked at the image I was appalled, I thought maybe I finally needed glasses. Later when I found out it had significantly less resolution and saw how appallingly blurry the chart tests were, it all made sense. My only point is, sometimes the numbers do bear out--and sometimes just because we aren't bugged by a particular flaw doesn't mean others will be as forgiving (whether or not they know the resolution beforehand). Unfortunately, as TVs get bigger and bigger and the casual viewer develops more and more of a demanding eye, that lost resolution could really hurt.

That said, if you love it, go for it!

Toke Lahti · January 22nd, 2005, 05:29 AM

<<<-- Originally posted by Barry Green: Consider that a normal DV camera has ostensibly up to 480 scan lines of resolution, but in observed practice about the most you can get resolved is around 360-380. In progressive-scan mode you can actually get all 480 on a DVX or XL2 in thin-line detail. But in interlaced, those same cameras will max out at about 360 to 380. -->>>

If xl2 has 480 lines with pNTSC and 576 with pPAL, how come I haven't seen any interlace flicker (25/30Hz) in the interlaced crt? Shouldn't there be flicker with details that are one line tall?
Or does all dv-equipment have low pass filter in the da-converter that removes the flicker and lowers the resolution for analog domain.
That way one could get full resolution only out of firewire.

Barry Green · January 22nd, 2005, 03:54 PM

Depends on what setting you put "vertical detail" on. "Normal" bypasses the interlace flicker reduction filter, "Low" employs the filter.

If you shoot 24p or 30p with vertical detail "normal", and there's fine thin horizontal lines in your image, yes you'll see flicker, ranging from mild to overwhelming. That'll happen with any camera that shoots progressive footage but displays it on an interlaced display. Shoot a resolution chart and you'll really see it.

January 20th, 2005, 01:53 PM	#1
Brian Broz Regular Crew Join Date: Oct 2002 Posts: 61	Is CF30 really 1/2 resolution?? Hi, I've been testing/shooting sample footage with an FX-1 viewing on a Sony LUMA HD monitor (via HD analog component). I have used the DVX100A and PD170 a lot so I have a pretty good idea what the different camera's advantages/weaknesses are. Form the tests I have shot so far, I really can't see a 50% loss in resolution that people are claiming (in CF30 mode)...viewing in HD, not SD/DV! Yes CF24 seems pretty useless IMHO. For example, when shooting with a PD150/170 you see a noticeable softness/loss of resolution in 1/30 shutter mode...and the introduction of stairstepping and "rough" edges. To me, CF30 is basically indistinquishable from the standard 60i mode. With the FX-1 I really can't see the an appreciable loss of res in CF30 (viewing on a $3000.00 Sony pro monitor). On relatively wide shots I don't get any stair-stepping and can still read incredibly fine details/print that of course I can see in 60i/std interlace mode. Again, this is with a locked-off shot. Does anyone else have any experience or care to comment? It's possible I need to have my eyes checked :-) but to me the "loss" in resolution is negligable. Hoping for some input everyone, Thanks, Brian Broz www.llsr.com

January 20th, 2005, 02:26 PM	#2
Chris Hurd Obstreperous Rex Join Date: Jan 2001 Location: San Marcos, TX Posts: 27,366 Images: 513	Howdy from Texas, Well, I have a comment, Brian. << I really can't see a 50% loss in resolution that people are claiming (in CF30 mode) viewing in HD, not SD/DV! ...to me the "loss" in resolution is negligable. >> Regarding resolution loss in Sony's CineFrame mode (or Panasonic's standard definition Frame mode found in Canon and Panasonic SD camcorders), my comment is "so what." I firmly believe that the proof is in what you see. If the the loss in resolution appears to be negligable, then it's no big deal. Too many people tend to get hung up on the number of pixels they're dealing with, and it seems a large portion of these folks have never actually seen the results; they're just going by the numbers. Are you happy with what you see -- and more importantly, is your intended audience (your business clients) happy with what they see. The answer is most often yes, it looks great. So what if there is a loss of res. Can you see it? Does it make a huge difference in image quality so much as to make it unusable? Can you create viable, productive material with sufficient quality? Almost all the time, the answer is yes. I personally know scores of people who shoot in nothing but Frame mode all the time with Panasonic and Canon camcorders -- the real number is probably a significant percentage of all owners of these camcorders -- who are more than pleased with the results they're getting. Resolution loss in CF mode? So what -- the important thing is, are you happy with the image. That's all that matters. Some folks tend to be too consumed with numbers to realize how good they've got it with these pseudo-progressive modes. __________________ CH Search DV Info Net \| 20 years of DVi \| ...Tuesday is Soylent Green Day!

January 20th, 2005, 02:52 PM	#3
Brian Broz Regular Crew Join Date: Oct 2002 Posts: 61	Well said Chris, I agree completely...However, I would like to know simply for peace of mind:-) For example if a project is going to be projected onto a large screen, or heavily tweaked/color-corrected in post...it may affect my decision of what mode to shoot in. Basically I think I need to get a hold of a resoltion chart and see for my self. Personally I don't care if I can see a loss in res zoomed at 200 or 400% of the frame, simply at 1440x1080 res. I am really surprised to read all the trashing and nit-picking of the FX1/Z-1. Of course it's not a Varicam or HDcam...but it's groundbreaking for the money (again viewed in HD). All the BetaSp and Digibeta shooters (that have also shot mini-dv) have seen our test footage on the Luma monitor were both surprised and impressed with the image. Well, does anyone know where I could find a download-able resolution chart? Many thanks, Brian Broz

January 20th, 2005, 04:22 PM	#4
Barry Green Barry Wan Kenobi Join Date: Jul 2003 Location: North Carolina Posts: 3,863	First, what Chris says is exactly true -- if you like what you see, so what? But since you ask: there is a loss, but it's nowhere near 50% as compared to the 60i mode. As tested by a resolution chart, 60i delivers about 775 lines, CF30 delivers about 575.

January 21st, 2005, 01:46 AM	#5
Brian Broz Regular Crew Join Date: Oct 2002 Posts: 61	Hi Barry, I have read many of your posts regarding the DVX and appreciate all the time you put on these boards. I have a question (which feels like a dumb question) regarding the resolution. How does the resolution of HDV become 775 lines? That doesn't sound like much...especially considering the images I've seen. I'm not questioning your findings, I'm curious because most DVCAM or DVCPRO (2/3") camera heads shoot 800-850 lines before the signal is recorded. So higher resolution than hidef...doesn't seem right, or am I missing something obvious? Many thanks for any response, Brian HDV Fan

January 21st, 2005, 03:13 AM	#6
Toke Lahti Major Player Join Date: Oct 2004 Location: Helsinki, Finland Posts: 158	I'm also little a bit confused about these resolutions. How can interlaced resolution be so high as 775 lines? If there are details less than 2 lines high, aren't they having interlace flickering (25/30Hz)? If details shorter than 2 lines are low pass filtered, wouldn't the resolution me more like 540?

January 21st, 2005, 03:29 AM	#7
Barry Green Barry Wan Kenobi Join Date: Jul 2003 Location: North Carolina Posts: 3,863	The specs you'll see about a broadcast camera's head resolving 800 or 850 lines or whatever refers first of all to "TV Lines", not scan lines, and secondly, that's talking about the camera head, not the recording format. The camera head may be capable of resolving 850 lines, but by the time it gets recorded on DV/DVCAM/DVCPRO/DigiBeta tape, the absolute maximum possible is 540 lines. That's a limitation of the recording medium, regardless of what the camera head is putting out. The FX1 can resolve and record somewhere around 850 lines horizontally, and about 775 lines vertically, as measured on a resolution chart. Now, 775 sounds low, when talking about a format that's supposed to have 1080 pixels, right? Well, that's where Toke's observation comes in, about interlacing and field-combining and low-pass filtering. Steve Mullen postulated that the FX1's maximum vertical resolution should theoretically be 820 lines, as he calculated that a field shouldn't be capable of resolving more than 410 lines. When I observed 775, I figured that's close enough to his 820 to call it "close enough". Consider that a normal DV camera has ostensibly up to 480 scan lines of resolution, but in observed practice about the most you can get resolved is around 360-380. In progressive-scan mode you can actually get all 480 on a DVX or XL2 in thin-line detail. But in interlaced, those same cameras will max out at about 360 to 380. So, 775 may not sound as promising as 1080 would make you think, but it's still around double what an SD camera can do. Concerning the 540/575 lines of CineFrame 30, I measured 575 on a resolution chart. I don't know exactly how the camera accomplishes this; you'd think that if it was a straight de-interlace of the interlaced signal, you'd get around 400 lines (depending on whether my observed 775 or Steve's theoretical 820 were the more accurate number). However, it's clearly much better than 400, it's around 575. That leads me to wonder if it isn't just a straight raw field right off the 1080 CCD, which would give it 540 resolved scan lines. 540's close enough to what I saw (575) to make me think that that's probably what it's doing.

January 21st, 2005, 05:35 PM	#8
Betsy Moore Major Player Join Date: Mar 2003 Location: Los Angeles Posts: 444	I agree the proof is in the pudding--but in the case of any frame mode I've ever seen the allegorical pudding is terrible to my eyes. And just in case you think my opinion might arise from a numbers bias, the first time I ever saw a Canon frame mode I assumed it had the same resolution (or higher) before I looked at it, then when I looked at the image I was appalled, I thought maybe I finally needed glasses. Later when I found out it had significantly less resolution and saw how appallingly blurry the chart tests were, it all made sense. My only point is, sometimes the numbers do bear out--and sometimes just because we aren't bugged by a particular flaw doesn't mean others will be as forgiving (whether or not they know the resolution beforehand). Unfortunately, as TVs get bigger and bigger and the casual viewer develops more and more of a demanding eye, that lost resolution could really hurt. That said, if you love it, go for it!

January 22nd, 2005, 05:29 AM	#9
Toke Lahti Major Player Join Date: Oct 2004 Location: Helsinki, Finland Posts: 158	<<<-- Originally posted by Barry Green: Consider that a normal DV camera has ostensibly up to 480 scan lines of resolution, but in observed practice about the most you can get resolved is around 360-380. In progressive-scan mode you can actually get all 480 on a DVX or XL2 in thin-line detail. But in interlaced, those same cameras will max out at about 360 to 380. -->>> If xl2 has 480 lines with pNTSC and 576 with pPAL, how come I haven't seen any interlace flicker (25/30Hz) in the interlaced crt? Shouldn't there be flicker with details that are one line tall? Or does all dv-equipment have low pass filter in the da-converter that removes the flicker and lowers the resolution for analog domain. That way one could get full resolution only out of firewire.

January 22nd, 2005, 03:54 PM	#10
Barry Green Barry Wan Kenobi Join Date: Jul 2003 Location: North Carolina Posts: 3,863	Depends on what setting you put "vertical detail" on. "Normal" bypasses the interlace flicker reduction filter, "Low" employs the filter. If you shoot 24p or 30p with vertical detail "normal", and there's fine thin horizontal lines in your image, yes you'll see flicker, ranging from mild to overwhelming. That'll happen with any camera that shoots progressive footage but displays it on an interlaced display. Shoot a resolution chart and you'll really see it.