MJPEG vs. AVI uncompressed- difference great?
Hooper
is a big difference in the final mpeg-1 when compressing from mjpeg
and avi uncompressed?
Hooper
Kyong Song
Well, I tested it, but I don't think the difference is that great. Although
I didn't put it on my homepage, half size MJPEG vs. half size uncomp. AVI
for VCD size MPEG, there was noticeable difference (at least, for me). But,
when you capture full size then resize it down, difference wasn't that big.
Still there WAS more artifacts compared to uncompressed AVI (and quality
score is far lower than uncomp. AVI when you encode), but, probably it is
not worth getting separate uncompressed AVI capture or faster HD for that
purpose, if you already have MJPEG capture.
However, if you are deciding what to buy, I certainly recommend you to get
AVI capture since it is far cheaper.
You can find full size capture comparions in my homepage, under "Capture
Methods" section.
K.S.
--
------------------------------------------------------
For MPEG creation, you can visit my homepage.
http://members.xoom.com/ksong1222/start.htm
------------------------------------------------------
Hooper <a...@def.ghi> wrote in message
news:a0prbs4vf7c61e5gj...@4ax.com...
Jeff Carbello
I saw someones VCD MPEGs that were created from a DV capture using a Sony
Analog to DV Converter and the quality looked very good, about equal to my
clips done using uncompressed AVI.
Also, Is the DV codec better than MPEG2?
Kyong Song
I have no experience with DV codecs, so I have no idea about them...
K.S.
--
------------------------------------------------------
For MPEG creation, you can visit my homepage.
http://members.xoom.com/ksong1222/start.htm
------------------------------------------------------
Jeff Carbello <jfca...@NOSPAMameritech.net> wrote in message
news:pfov4.1016$Pq2...@nntp0.chicago.il.ameritech.net...
yep
>I want to capture and compress to mpeg-1. Can someone tell me if there
>is a big difference in the final mpeg-1 when compressing from mjpeg
>and avi uncompressed?
I did not do experiments with half frame size but at full size at
least you can get impeccable reslts if you have a fast CPU, so I'm
sure it works well at half size, even with a smaller CPU.
BTW I cannot do any meaningful tests of full size from S-VHS or Hi8
sources because they are so bad compared to DV or a good TV station.
----------------------------------
http://www.freenet.de/codecpage
remove nospam- from e-mail address
yep
<jfca...@NOSPAMameritech.net> wrote:
>How about DV captured video vs Uncompressed AVIs?
>
>I saw someones VCD MPEGs that were created from a DV capture using a Sony
>Analog to DV Converter and the quality looked very good, about equal to my
>clips done using uncompressed AVI.
>
>Also, Is the DV codec better than MPEG2?
>
That comparison is a bit unfair. MPEG2 is 1/4 the size of DV. It's
certainly a lot worse. DV is the best but it's 200 MB/minute. MJPEG at
100 MB/minute is a bit less but still way better than S-VHS. Any
analog consumer tape format is much worse than DV. You can easily see
this when making an analog copy from an analog tape. I also could not
deinterlace any such source without compromising quality because of
the tiny noise and edge jitter it has, With DV, no problem.
BTW do not spend one buck for analog equipment any more (except a
simple TV card). Go strictly DV.
Hooper
difference when you make an mpeg-1 file from mjpeg or a avi
uncompressed? By the way when I said I have an s-vhs quality source I
meant I have a very good source like dss, I don't actually have an
s-vhs vcr or tape, but I thought saying s-vhs quality would mean the
same as a good dss picture.
On Thu, 02 Mar 2000 14:03:08 +0100, yep <nospam-...@email.com>
wrote:
>On Wed, 01 Mar 2000 22:50:23 -0500, Hooper <a...@def.ghi> wrote:
>
>>I want to capture and compress to mpeg-1. Can someone tell me if there
>>is a big difference in the final mpeg-1 when compressing from mjpeg
>>and avi uncompressed?
>
>I did not do experiments with half frame size but at full size at
>least you can get impeccable reslts if you have a fast CPU, so I'm
>sure it works well at half size, even with a smaller CPU.
>
>BTW I cannot do any meaningful tests of full size from S-VHS or Hi8
>sources because they are so bad compared to DV or a good TV station.
>
>
yep
>I'm not sure what you're saying. Are you saying there's not too much
>difference when you make an mpeg-1 file from mjpeg or a avi
>uncompressed? By the way when I said I have an s-vhs quality source I
>meant I have a very good source like dss, I don't actually have an
>s-vhs vcr or tape, but I thought saying s-vhs quality would mean the
>same as a good dss picture.
>
Indeed, no big difference between mjpeg and uncompressed if you run
mjpeg at a moderately high quality setting. Look at the stills I put
on that website (link below).
S-VHS is really not a synonym for top quality, it's good for consumer
use and may be much better than many US (not European) TV stations but
Betacam (analog pro format), DV and DVD are all much better. It
becomes obvious when you make copies. The heavy signal processing
necessary to get analog video out of jitter and tape noise has much
worse effects than a well tuned digital compression.
Hooper
wrote:
>Indeed, no big difference between mjpeg and uncompressed if you run
>mjpeg at a moderately high quality setting.
What do you consider to be a high quality mjpeg setting? Would
capturing mjpeg at 3-4mb's/sec be good?
yep
For a good software codec, 1.5 MB/sec will be a very good setting with
the 2 fields option off (e..g. fields are compressed together, but all
lines, nothing omitted, may still be deinterlaced). This equals a
quality setting of luminance 30, chrominance 25, 4:1:1 with the PIC
codec or 50%, 4:1:1 with the Main Concept codec. Source format YUY2.
Both codecs perform equally fast and good. You can get the PIC for
free while the other is $25 but compatible with most hardware
encoders. Demo version available
CPU should be Intel >=550 MHZ for PAL or about >=500 for NTSC.
If you encode 2 fields separately, you will have to tune up quality
quite a bit, but probably no more than 2..2,5 MB/sec. Only that you
also need a faster CPU for real time capture, that would be at least
700 MHz for NTSC I guess.
Have a look at my web page, it's all explained there.
bal...@my-deja.com
> How about DV captured video vs Uncompressed AVIs?
I am interested in this one as well. If one starts with a DV type AVI
file and want to go to MPEG-1, will there be a significant difference
in quality than if one started with uncompressed AVI?
Olushola
Sent via Deja.com http://www.deja.com/
Before you buy.
James Yan
low compression ratio (even lower than my Marvel G200). But I didn't
compare the quality systematically.
On Fri, 03 Mar 2000 11:03:16 +0100, yep <nospam-...@email.com>
wrote:
James Yan
very low compression ratio (even lower than Marvel G200), but I
havent' compared it with other encoder systematically.
yep
wrote:
>How about the quality of Morgan mjpeg encoder? It can compress with
>very low compression ratio (even lower than Marvel G200), but I
>havent' compared it with other encoder systematically.
I guess you mean high compression ratio (small files).
I've checked out Morgan some weeks ago and it was 2-3 times slower
than PIC or the new Main Concept (beta). Not able to capture full
frame size in real time.
MJPEG is not for distribution but for capturing and editing, so if a
codec, hard or soft, can't do real time it's not very useful to me.
Morgan is typically used to edit hardware captured MJPEGs or capture
half size.
BTW frame by frame, PIC and MC are already near the ratios a still
picture JPEG can achieve. I guess that can't be topped very much.
Rich A
a...@def.ghi says...
> I want to capture and compress to mpeg-1. Can someone tell me if there
> is a big difference in the final mpeg-1 when compressing from mjpeg
> and avi uncompressed?
>
comments based on experience. The short answer is that you need
to define "big". A difference? Yes. The long answer follows:
MJPEG is a compressed video. Simply put it shrinks the size of
the video by treating each frame as an image and applying
compression to it similar to the JPEG compression we all know and
love in the graphics world. So every frame is a "key" frame and
every frame is compressed.
A straight non-compressed AVI file (of the same frame size) has
NO compression done to it. If you took the same number frame
from each, and blew them up enough and compared them the MJPEG
would have blocks and artifacts and the straight AVI would not.
Now when your Mpeg encoder does it's thing, it's going to do some
fancy frame comparisons and (overly simplified) examine each
frame of the source in MUCH greater detail than your eye/brain
can possibly do. Then it will divide the frames into different
types of frames .. some having video content divided into macro
blocks and compressed and some having no video content but just
mathematics that will allow the "player" to rebuild those frames
during playback. There will be adjustments for "motion" from
frame to frame and some redundant information will be either
thrown out or converted to mathematical equivalents.
So knowing all that, which will make the better final VCD (if you
really nit pick and examine them "closely") The MJpeg that is
already full of compression artifacts? Or the pure uncompressed
AVI source? It's rather logical to me.
It's the same reason so many are disappointed with their efforts
at creating VCDs from their home recorded VHS tapes. VHS tape is
certainly the lowest end of the storage media quality wise.
While you watch it on your TV it may look pretty good. But try
examining it frame by frame like your mpeg encoder is going to
do. The poor quality of VHS along with noise, shakiness,
over/under exposure and all the other things not normally noticed
while being viewed real time on the TV just drive that Mpeg
encoder crazy.
You can't make VHS "better" than it "is" by simply encoding it to
a VCD bit rate Mpeg video. To get the end result of a "near VHS"
quality you have to START with something BETTER than VHS.
Sorry to run on .. but I've always thought all this was rather
logical and from lurking these news groups for the past year or
so I see a lot of newbies who really need a plain language
explanation else they quickly become frustrated and paranoid
because they can't get desired results.
Okay, I'm off the soap box. Sorry for the long winded diatribe.
--
Rich A. http://www.members.home.net/richa/
Hooper
source though, I'm not messing around with that bad vhs crap. My plan
isn't to capture to a low vcd rate, but to capture at a high mpeg bit
rate and burn it to dvd-ram as an mpeg file. Then whenever I want to
watch I will just play it from my computer to my tv.
Rich A
a...@def.ghi says...
> Thank you for the informative post. I did say I had an s-vhs quality
> source though, I'm not messing around with that bad vhs crap. My plan
> isn't to capture to a low vcd rate, but to capture at a high mpeg bit
> rate and burn it to dvd-ram as an mpeg file. Then whenever I want to
> watch I will just play it from my computer to my tv.
>
for sure. And yes, if you are going the DVD-RAM way, (without
the 650 mb CDR limit) you can make some very nice stuff. I'm not
sure what the transfer speed of those DVD-RAMs are but I guess it
would be okay for maybe 3 or 4 Mb/s Mpeg-1.
I went another direction for my "higher quality" stuff. I have
an Indigita digital tape drive that takes inexpensive DAT tapes.
But it also can record and play back Mpeg-1 or 2 up to around 6
Mp/s. I can get about 2 hours of DVD quality Mpeg-2 on a single
tape that costs around 4 bucks. Of course it's still "tape"
<grin> And not as "tough" as CD. But it's a lot cheaper than
those DVD-RAM discs. I'm still waiting for the DVD-R media and
drives and authoring software to come down in price to where it's
nearer to the "pro-sumer" market so we can burn our own "real"
DVDs.
Hooper
>I went another direction for my "higher quality" stuff. I have
>an Indigita digital tape drive that takes inexpensive DAT tapes.
>But it also can record and play back Mpeg-1 or 2 up to around 6
>Mp/s. I can get about 2 hours of DVD quality Mpeg-2 on a single
>tape that costs around 4 bucks. Of course it's still "tape"
><grin> And not as "tough" as CD. But it's a lot cheaper than
>those DVD-RAM discs. I'm still waiting for the DVD-R media and
>drives and authoring software to come down in price to where it's
>nearer to the "pro-sumer" market so we can burn our own "real"
>DVDs.
That's a great idea, but yea it is just tape and prone to wear out. As
for the DVD-RAM I was thinking of getting the Creative Labs one (just
$235 or so), but check out these specs from their site. Does this mean
it can't do 3 to 4mb/s of mpeg-1 (or mpeg-2 for that matter)? If it
can't, could a Hollywood Plus decoder card overcome this problem?
Maximum data transfer rates
- DVD-RAM: 1,350KB/second (1X)
- DVD-ROM: 2,700KB/second (2X)
- CD-ROM: 2,400KB/second (16X)
Average access times
- DVD-RAM: 180ms
- DVD-ROM: 260ms
- CD-ROM: 150ms
2MB data buffer
Jeff Carbello
I am going to compare:
DV -> MPEG-1
Uncompressed AVI -> MPEG-1
MPEG-2 I-Frame -> MPEG-1
DV -> MPEG-2 DVD compliant stream
I am also going to try multiple sources DV tape, cable feed, LaserDisc, DVD,
S-VHS Taping from DirectTV
bal...@my-deja.com wrote in message <89obrt$bdt$1...@nnrp1.deja.com>...