Shooting progressive with HC1? - Page 3 at DVinfo.net
DV Info Net

Go Back   DV Info Net > Sony XAVC / XDCAM / NXCAM / AVCHD / HDV / DV Camera Systems > Sony HDV and DV Camera Systems > Sony HVR-A1 and HDR-HC Series
Register FAQ Today's Posts Buyer's Guides

Sony HVR-A1 and HDR-HC Series
Sony's latest single-CMOS additions to their HDV camcorder line.

Reply
 
Thread Tools Search this Thread
Old December 1st, 2006, 03:05 PM   #31
 
Join Date: Jan 2004
Location: Stockton, UT
Posts: 5,648
Quote:
Originally Posted by Graham Hickling
Yes, fine to all of that.

But the question that kicked off this thread still stands.....I fail to understand how interlace lines could be VISIBLE in footage shot at 1/30.
The NLE may not be interpreting the frame/two fields correctly. There are a few other reasons, but that would be my first suspect.

[edit] as Greg succinctly worded, the difference is entirely in the acquisition timing. The HC1 is always offset, which is why you can't move the camera at all, or have motion if you don't want to see interlacing offsets. Canon have a proprietary (and brilliant) implementation similar to Psf that works wonderfully well to give the look of progressive in an interlaced image.
__________________
Douglas Spotted Eagle/Spot
Author, producer, composer
Certified Sony Vegas Trainer
http://www.vasst.com
Douglas Spotted Eagle is offline   Reply With Quote
Old December 1st, 2006, 07:37 PM   #32
Regular Crew
 
Join Date: Mar 2006
Location: Baltimore, MD
Posts: 161
Just to take a wild guess without having a clue about how Sony reads the CMOS chip in the HC1 -- Could it be that, even at 1/30sec shutter, Sony is reading the odd and even lines from the CMOS chip at different time points (differing by 1/60th of a second), so the two fields really are temporally distinct? That is:

frame 1, odd: exposed for time points 1 & 2 (each time point representing 1/60sec)
frame 1, even: exposed for time points 2 & 3
frame 2, odd: time points 3 & 4
frame 2, even: time points 4 & 5

I think you could see this with a strobe light (or even just a camera flash), which half the time would only illuminate one field and not the other.

Or did I just reveal my stupidity? (don't answer that)

-Terence
Terence Murphy is offline   Reply With Quote
Old December 1st, 2006, 09:09 PM   #33
Trustee
 
Join Date: May 2004
Location: Knoxville, Tennessee
Posts: 1,669
I think I may have the answer.

HC1 footage shot at 1/30 is indeed composed of pairs of identical fields ... except for the first one!

So the fields are: abbccddeeffgghhiijjkk and so on.

Displayed in pairs on an NLE, this shows up as ...ab bc cd de ef... which of course looks interlaced.

But if you clip off that first field (which I did just now in avisynth) this converts to ... bb cc dd ee ff gg hh ii jj kk ... which LOOKS perfectly progressive, with NO lines, even though the underlying structure is of course still interlaced.

You are stuck with a 1/30th shutter, but the result really looks rather nice! Someone wiser than I should write a little app to implement this!
Graham Hickling is offline   Reply With Quote
Old December 1st, 2006, 11:49 PM   #34
Trustee
 
Join Date: Mar 2004
Location: Milwaukee, WI
Posts: 1,719
Quote:
Originally Posted by Graham Hickling
I think I may have the answer.

HC1 footage shot at 1/30 is indeed composed of pairs of identical fields ... except for the first one!

So the fields are: abbccddeeffgghhiijjkk and so on.

Displayed in pairs on an NLE, this shows up as ...ab bc cd de ef... which of course looks interlaced.

But if you clip off that first field (which I did just now in avisynth) this converts to ... bb cc dd ee ff gg hh ii jj kk ... which LOOKS perfectly progressive, with NO lines, even though the underlying structure is of course still interlaced.

You are stuck with a 1/30th shutter, but the result really looks rather nice! Someone wiser than I should write a little app to implement this!
Very interesting Graham! I have tried to figure out a way to us this type of video but I have never tried to set the order of the fields in a different way. I also use Avisynth a lot and I will have to play around with the fields now that you figured out the pattern. Very interesting indeed. I guess now I will not get any sleep tonight. I need to try this out.
Thomas Smet is offline   Reply With Quote
Old December 2nd, 2006, 12:12 AM   #35
Trustee
 
Join Date: May 2004
Location: Knoxville, Tennessee
Posts: 1,669
Thomas,

Just FYI, here's the script I used to access the raw fields:

avisource("original.avi")
assumetff
separatefields

I then manually set the in-point 1 frame in, rendered to a new file named trimmed.avi, and then converted that back to 1080i with this script:

avisource("trimmed.avi")
assumefieldbased
weave

There is undoubtably an elegent way of doing all 3 steps with a single script - but I was rushing a bit this evening ....
Graham Hickling is offline   Reply With Quote
Old December 2nd, 2006, 12:53 AM   #36
Trustee
 
Join Date: May 2005
Location: Wyomissing, PA
Posts: 1,141
Images: 57
So it's higher quality way of deinterlace, but for only 1/30 footage?
Peter Ferling is offline   Reply With Quote
Old December 2nd, 2006, 01:32 AM   #37
Regular Crew
 
Join Date: Feb 2005
Location: Rainham, Kent, UK
Posts: 69
Quote:
Originally Posted by Graham Hickling
I think I may have the answer.

HC1 footage shot at 1/30 is indeed composed of pairs of identical fields ... except for the first one!

So the fields are: abbccddeeffgghhiijjkk and so on.

Displayed in pairs on an NLE, this shows up as ...ab bc cd de ef... which of course looks interlaced.
That's exactly what I interpreted Mikko's comment that "the fields are flipped" to mean. Shame we couldn't have arrived at the answer without the unnecessary sarcasm...
Steve Brady is offline   Reply With Quote
Old December 2nd, 2006, 10:55 AM   #38
MPS Digital Studios
 
Join Date: Apr 2003
Location: Palm Beach County, Florida
Posts: 8,531
I have no issues shooting with CF30 or 25, or even not shooting with it and de-interlacing. The video looks great!

heath
__________________
My Final Cut Pro X blog
Heath McKnight is offline   Reply With Quote
Old December 2nd, 2006, 06:28 PM   #39
Regular Crew
 
Join Date: Aug 2006
Location: London United Kingdom
Posts: 77
so in conclusion what i derive from this thread is:

with a shutter speed of 1/30 sec and appropriate post processing of the captured HDV signal the HC1 can produce video that is indistinguishable from one captured using a progressive (i.e. 30p) camera? is that right?

surely it cant be that easy! how cool would that be?
Mauritius Seeger is offline   Reply With Quote
Old December 2nd, 2006, 07:31 PM   #40
Trustee
 
Join Date: May 2004
Location: Knoxville, Tennessee
Posts: 1,669
I'm not sure if this approach produces something that's identical to CF30 (which the HC1 doesn't have although the A1 does).

I can say, however, that it's an improvement on the "Cinema Efect" digital effect on the HC1, because Cinema Efect disables manual exposure control.

(By the way - I'm typing 'Efect' intentionally as that's what the HC1 menu says - ha!)
Graham Hickling is offline   Reply With Quote
Old December 3rd, 2006, 01:51 AM   #41
Regular Crew
 
Join Date: Feb 2005
Location: Rainham, Kent, UK
Posts: 69
The main limitation, of course, is that you're stuck with a 1/30th shutter speed, which means that you'll get twice the amount of motion blur that you'd see with "real" progressive video under "normal" conditions. As Graham already said, most interlaced cameras can do this, so try it out and see what you think of the look before getting too excited...
Steve Brady is offline   Reply With Quote
Old December 3rd, 2006, 02:27 AM   #42
Major Player
 
Join Date: Mar 2004
Location: Portland, OR
Posts: 314
I really love the footage from the HC1 but it only shoots interlaced footage, plain and simple.

Quote:
Originally Posted by Graham Hickling
I think I may have the answer.

But if you clip off that first field (which I did just now in avisynth) this converts to ... bb cc dd ee ff gg hh ii jj kk ... which LOOKS perfectly progressive, with NO lines, even though the underlying structure is of course still interlaced.
Huh? Any movement or slanted lines shown using pairs of identical fields will definitely NOT be the same as progressive.

Quote:
Originally Posted by Peter Ferling
So it's higher quality way of deinterlace, but for only 1/30 footage?
No way. You are better off shooting 60i at 1/60 (or faster) and using a smart deinterlacer that adaptively interpolates to remove interlacing from areas with motion while maintaining the full resolution of the stationary parts. The 1/30 method just gives you field-doubling at best.

However, if you do like the way the 1/30 footage looks then that is really all that matters.

Last edited by John McManimie; December 3rd, 2006 at 03:42 PM.
John McManimie is offline   Reply With Quote
Old December 3rd, 2006, 05:04 PM   #43
Trustee
 
Join Date: Mar 2004
Location: Milwaukee, WI
Posts: 1,719
Ok I tried tried a lot of this over the weekend.

1. shutter of 30. This will give you progressive images although with a lot of motion blur. This will work for some people who happen to like this look.

2. Flash digital effect at the lowest setting. Does not work. What it does is give you 1 frame then 1 blended frame and so on. By getting rid of the blended frames you end up with 30p but only at half resolution. So basically you get the same as you would by shooting 60i and deinterlacing.

3. Photo mode. The camera in photo mode seems to work in progressive but it uses the same odd field setup as the 30 shutter mode. By cutting the first field and putting them back together I get 30p. I'm not sure what the shutter is like in this mode since it is really hard to compare with this mode. This mode also only works by capturing into an uncompressed system live. There is no way to record to tape when in this mode. The motion did look a little blurry but not as bad as a shutter of 1/30. Of course this mode would alsmost be useless unless you can capture live. You also have no control of the camera at all in this mode and it basically becomes a point and shoot camera. You can manual focus and zoom but thats about it. This also explains why some people have said it looks like 30p to them when they hook it up to their HDTV. On a digital HDTV the 30i shifted field video would get bobbed to 60p so it would actually look like 30p. Very interesting. Actually there is a lot more control in this mode then I thought there was. You can still set sharpness, exposure, white balance, focus, zoom, program AE and AE shift. The only problem is what to capture it to.
Thomas Smet is offline   Reply With Quote
Old December 3rd, 2006, 07:53 PM   #44
Regular Crew
 
Join Date: Aug 2006
Location: London United Kingdom
Posts: 77
Quote:
Originally Posted by John McManimie
I really love the footage from the HC1 but it only shoots interlaced footage, plain and simple.
well, i dont think it's either plain or simple - just look at the length of the thread.

and perhaps not even true, given the fact we just established that it is possible to derive progressive images when using a 1/30 sec shutter speed (with the right post processing)
Mauritius Seeger is offline   Reply With Quote
Old December 3rd, 2006, 10:03 PM   #45
Trustee
 
Join Date: Mar 2004
Location: Milwaukee, WI
Posts: 1,719
2. When I said frames I meant fields. The odd field is a true field while the even field is a blend of the odd field and the next odd field. Very odd. SONY really came up with some interesting ways of dealing with interlaced frames.
Thomas Smet is offline   Reply
Reply

DV Info Net refers all where-to-buy and where-to-rent questions exclusively to these trusted full line dealers and rental houses...

B&H Photo Video
(866) 521-7381
New York, NY USA

Scan Computers Int. Ltd.
+44 0871-472-4747
Bolton, Lancashire UK


DV Info Net also encourages you to support local businesses and buy from an authorized dealer in your neighborhood.
  You are here: DV Info Net > Sony XAVC / XDCAM / NXCAM / AVCHD / HDV / DV Camera Systems > Sony HDV and DV Camera Systems > Sony HVR-A1 and HDR-HC Series


 



All times are GMT -6. The time now is 07:49 PM.


DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network