Movie Maker 2 and Photo
Story
PapaJohn's Newsletter
#54 - May 21, 2005
Interlaced versus
Progressive
I'll assume you're familiar enough
with "interlaced" and "progressive", the two methods of
advancing along the visual images in a video or movie.
Interlacing 'jaggies' are seen
when you look at a frame made of 2 overlapping half
frames, each half frame showing every
other horizontal line...
Progressive 'blurriness' is seen when you look
at a moving subject in a picture taken with a slow shutter speed of
1/25 or 1/30 of a second...
Sometimes the technology
dictates which of the two to use. Sometimes you get to pick.
We'll get into the jags
and blurs this week... let's first take a fun look at some of the
macro factors that helped us get to our current state of having a
choice.
-
Movies and
Hollywood, using film projectors, came a bit later.
Cameras shot on film strips using whole frames. Each
frame of a film reel flicked on the screen for a split second... with the
frames progressively changing one at a time. Most theaters use
film-based projection systems that display 24 frames a
second.
If you take a sill picture
with a camera and snap it at 1/24 of a second, you would have a slightly to
heavily smudged or blurred image if there was any kind of movement. Stopping
at any frame and looking closely would
show the blurring. We don't usually get to see a frame stopped
on screen, so it's not an issue.
-
TV sets tie
into the cycles of electricity that make their Cathode Ray Tubes
(CRT) work. Using the first 1/2 of each electrical
cycle to show half of the lines of a picture, and
interlacing the spaces with the other lines during the second half
of the cycle. The 60 cycles per second wasn't an exact number, so half of the
real number is the NTSC video standard of 29.97 frames per
second. The PAL number is a bit more precise - half of 50 cycles per
second is 25.
-
Computers,
being digital, powerful and precise, can manage make and
manage images either way. But they weren't first, so they have to
align with what was already there. If you'll be heading toward a DVD for
TV viewing, you need interlacing. For computer-based
viewing, progressive is better. But maybe you want both...
Which is
better? I've read that interlacing is considered a sign of being
associated with TV and VCR tapes, a lower class of viewers than Hollywood
and movie-goers.
The technology of
CRT computer monitors and TVs needs the interlacing, but
the goal is not to see jaggies when viewing there. One issue, which really
isn't an issue, is seeing the interlacing during the editing phase. Many
extrapolate the editing environment to the final movies, assuming what
they see in the project previewing will be what they see in the
saved movie.
Newer LCD
and plasma screens use a different technology one more
akin to computers... they can display whatever is played on them. Viewing
progressive home video is becoming more normal.
In home
movie-making activities, here are some places that you run
into interlaced or progressive:
- My digital camcorder is a Sony TRV-80, a higher end
model that offers "interlaced" recording
or "progressive". It default is interlaced but I can
change it.
-
Movie Maker 'preserves the
interlacing of DV-AVI files' as it captures from a digital camcorder to a
DV-AVI file, but 'de-interlaces' if you capture the same footage
into a WMV file.
- There are two sizes of WMV-HD (High Definition)... 720p and
1080i.... the number of lines followed by p for progressive and i for
interlaced.
- There are software options or utilities to 'de-interlace'... when
should you do it?
-
One of the options in the Profile
Editor is to 'allow interlaced processing', but my current website page says
not to check it. Why not?
-
When I use TMPGEnc to convert a file
to an MPEG-2 one for a DVD, one of the first things it does is check the
file for interlacing... bottom field first is it's usual conclusion, even if I
give it a de-interlaced WMV file. I need to go beyond it's defaults to
tell it when a source file isn't interlaced. Maybe I'd get a
better output if I told it.
When using the Profile
Editor and TMPGEnc, I'm starting to pay more attention to such
options, trying to choose more knowlegeably and wisely.
In this newsletter, I'll go
into my camcorder to see how interlacing differs from progressive.
some notes... before going
further
Notes...
I attended a Microsoft MSDN seminar in Kalamazoo on Tuesday, and
stopped by this Microsoft 'Across America'
truck.
A reader suggested the topic
of 'Video Blogging (Vlogging) using Moving Maker'
as a very timely topic. I added it to the
schedule.
There were some
hurdles to my work during the week. On Monday the server went down
for its 'longest ever' outage of 2+ days... a network card going bad. Then
my main working computer, a Toshiba laptop, had
it's screen stop lighting and is off to the CompUSA repair
center in Texas. When I brought my old Dell laptop back into service, the
power adapter chord had problems and I had to get a new one... with all
these hurdles, the content of this newsletter might take a slight hit... hope
it's not noticed.
Last night I
was checking an uncompressed NTSC AVI clip that has each frame
numbered. I was curious about which frames I could see in MM1 and MM2, which
frames were dropped in the generational issues of DV-AVI files, which frames
were used if I opted to render from an NTSC file (30 fps) to a PAL file (25
fps), and other things like that. With the Toshiba still in Texas, I
thought it would be a good topic for next week's newsletter... the subject
won't focus on the dropped frames, but more on
which frames you see. If one second of NTSC video has
30 frames and you render a PAL video from it, which 25 of the 30 frames are
used? Send me any guesses.
.... back to the main topic
I wondered, if
the world is moving from interlaced to progressive in the high definition
area, why shouldn't I flip the switch in my camcorder and do
all my recording as progressive? I asked on the SimplyDV
forum, where movie and video experts tend to go... and I read my camcorder
manual.
The manual has a
single page with an interesting title: "Recording with all the
Pixels - Progressive Recording Mode". You would think that using all
the pixels is better than fewer. But a footnote says the
camcorder won't do a couple things as it records progressive... one is
widescreen, and another is 'steadyshot' compensation for normal
camcorder movement. It also says the progressive recording will be a
slightly larger field of view.
The comments I
received at SimplyDV included a caution about losing the 'steadyshot'
feature, advice to stay with interlaced if the video is to be played on TV, use
progressive if you want clearly defined stills, interlacing looks smoother
on TV than progressive, and the image quality is far better in progressive
mode...
To check the
manual and advice at SimplyDV, I shot some footage
using one mode and then the other.
3 Sets of Clips - First
Impression
Here are snapshots from 3 sets of clips, using
the MM2 snapshot feature with the clips being previewed in the
collection...
The first and most obvious thing is the field of view. I
routinely shoot with a wide-angle lens on the camcorder, but the only change I
made between the clips at the left and right was to change the option for
interlaced versus progressive. The Sony manual said something about it using
more pixels when it shoots in progressive mode, but I hadn't expected the
difference to be as significant as this. The progressive option adds
considerably more wide-angle to the already wide-angle view.
If you don't have a wide angle lens but you have a progressive
option, then maybe you also have a built-in wide angle lens.
It's pretty easy to determine how much the pictures at the
right differ from those at the left... the left image is comparable to
cropping 480x360 pixels and resizing it to 640x480, a 33%
magnification in each direction.
Interlaced Clips at
Left
Progressive Clips at
Right
Camcorder steadied on arm-rest of big cushy chair...
pointing straight ahead... partially zoomed
Maximum 10x optical zoom... steadied on arm-rest but panning the
shelves for minor movement
No zoom... wide-angle lens on camcorder, hand-held with the
Toshiba on my lap
Selected Closer Views -
Magnified
I used TMPGEnc to extract each frame of the
clips to sets of BMP images, and browsed the pictures with
IrfanView to see the differences.
Items of interest were cropped and
considerably magnified to better see the jaggies of the interlaced clips
versus the blurriness of progressive.
interlaced
Progressive
Low Light Conditions
With the
camcorder on a tripod, in a low light situation... night-time, shooting
toward a closet that had darker crooks and crannies... Because the
progressive mode adds to the wide-angle view, I magnified the
cropped area of the progressive frame an
additional 33%.
I'm not
seeing any differences to indicate one is better than the
other.
Interlaced Progressive
- magnified additional 33%
Custom Profiles...
After checking the camcorder, I tried making a
custom WMV profile that passed the interlacing of an interlaced DV-AVI
file to a WMV. Could I? The Profile Editor only let me pick that option
with the Media Video 9 Advanced Profile, but then it did let
me.
I made a profile with settings comparable
to a high quality MPEG-2 file for a DVD... the picture at the left shows a
cropped section of the saved WMV movie. It doesn't look like the
jaggies were passed along.
As a check, I opted to not 'allow
interlacing' and rendered it again... it's the picture at the
right.
The results are similar and the profile/codec
doesn't result in interlacing being passed along to the WMV file.
Checking or unchecking the option doesn't matter.
Doing this exercise left me off at a
good starting point toward 2 new custom profiles as WMV
alternatives to DV-AVI files.
Conclusions and Closing
I'm going to continue shooting in interlaced mode. I like
widescreen and prefer steadier video.
The manual doesn't come right out and say it, but gives hints
about progressive frames being taken at 1/30 of a second and interlaced at 1/60
in order to get the two half-frames. If so, I'd expect the interlaced frames to
be sharper than progressive... but if the shutter is open twice as long for
a frame, why isn't it better in low light? I'd guess it's because of
the step back to a wider angle view, which uses the extra light rather than
giving it to the selection. That means zooming in tighter would give the pixels
back to the area of interest and the lower light situation might be better... I
need to do some more testing. The next 'last dance' at a wedding might be better
shot in progressive mode.
While playing with the WMV profile to see if the interlacing of
a DV-AVI source file would be passed through to a WMV file, I noticed that the
profile I made emulated a high quality MPEG-2 file. Pulling on that thread a
bit, I used TMPGEnc to convert the 10 second WMV file into files needed for a
DVD and compared the file sizes (the WMV file was 10.7 MB.... the .m2v was 8.9
MB with an associated 1.9 MB .wav file = 10.8 MB total). If bit-rate correlates
to quality and file size, then the WMV made with the profile
correlates to high quality DVD MPEG-2 files.
Could this be a better work-around of the generational
loss issues with DV-AVI files, and the audio issues in DV-AVI files... avoid
rendering to DV-AVI by rendering to a single high-quality WMV file
that aligns with the bit-rate and other parameters of an MPEG-2 file made for a
DVD? The bit-rate is 4 times as high as the highest WMV profile included in
Movie Maker 2.
I made custom profiles to use when saving a WMV file headed for
a DVD, for both NTSC and PAL. The settings include quality-based VBR
audio and video, a quality setting of 95 for the video, and lossless
audio. If you want to try them, there are new links on the Saving
Movies > Custom WMV Profiles page of the website. I checked
them by rendering WMV files and running them thru TMPGEnc to
get MPEG-2 files.
Have a great week...
PapaJohn
