[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [AV Media Matters] Compression
On Saturday afternoon (Nov 18), I will be chairing a session at the AMIA
Conference in LA and session title is: Compression in the Archives
Heather Weaver will be one of the panelists and she was the person
responsible for the video format tests reported in DV magazine. A few
months
ago, I posted a piece about these tests and I'm enclosing that piece here.
Jim Wheeler
Comparison of Quality of Several Videotape Formats
The May 1999 issue of "Digital Video" magazine had a report of tests of
several videotape formats. If you have Acrobat, you can download the report
from the Digital Video magazine website:
www.dv.com/magazine/1999/0599/formats0599.pdf
The Tektronix PQA-200 was used in all of these tests. Figures 1, 2, 3 show
the high number of "bad" pixels caused by the H--8 format as compared to the
non-compressed 4:2:2 DDR format.
3/4" U-Matic and Hi8 did not do very well in these tests as shown in several
charts that summarize the test results. Digital Betacam came out the best.
If D1, D2, D3, D5, and D6 had been included, they would have been in the
"best" category because they are uncompressed digital formats. That was
known by the people at Digital Video magazine so why waste time and money
testing these formats.
Betacam-SP, DVCAM-25, DVCPRO-25, and Digital-S varied in quality depending
on
the type of information being reproduced.
S-VHS was not tested because they believed that it would be similar to Hi8.
DVCPRO-50 was not tested because it was assumed that it would be similar to
Digital-S.
The Tektronix PQA-200 makes picture quality measurements using the Just
Noticeable Difference (JND) method developed at the Sarnoff lab to emulate
the perceptions of the human visual system. Several different test patterns
with motion are used and the Tektronix test unit produces a number that
represents the number of "bad" pixels. A Picture Quality Rating (PQR) of
zero means that there is no difference between the test video frame and the
same frame recorded on a uncompressed 4:2:2 disc or tape machine.
1.0 = Degradation with a small perceptual impact
3.0 = Mild (but observable) image degradation
10 and higher = Clearly observable image degradation
Digital Betacam was .92-1.14 on two of the test images. That's the
BEST of the compressed formats and also a very expensive
machine--about $40,000 US.
These tests confirmed what I already knew and that is that
compression will degrade the video quality. I am surprised that the
mild 1.6 to 1 compression ratio of Digital Betacam produced a PQR
over 1.0.
There is another type of compression (wavelet) that is "loseless" up
to about 15 to 1 compression ratio. Wavelet compression is used in
the TI E-cinema system.