Optimal MPEG2 capture settings for DVD 4.7GB
victor
find a clear answer to it and I know a lot of newbies will be asking
the same question.
Those of you that have experimented with the different bitrates and
encoding settings, what have you found to be the optimal MPEG2
encoding settings for a 2 hour analog movie to fit into a 4.7GB DVD?
I am capturing from a Sony Hi8 camcorder and into a ATI AIW8500DV on a
P-VI, 512mb memory. I am concerned with quality, but time is also an
issue and would like to avoid having to capture in one setting(like
uncompressed) and then having to enconde. I would also like to author
the DVD with minimal titles, menus and chapters.
Any information on this would be greatly appreciated. TIA.
Hughy
encoding settings have been discussed at length in the newsgroup. You could
start by searching for "TMPGEnc - Here is the mode to use for best quality of
all".
Sincere regards,
Hughy.
"victor" <asv...@yahoo.com> wrote in message
news:f44b92c1.02072...@posting.google.com...
Erratic
>I am capturing from a Sony Hi8 camcorder and into a ATI AIW8500DV on a
>P-VI, 512mb memory. I am concerned with quality, but time is also an
>issue and would like to avoid having to capture in one setting(like
>uncompressed) and then having to enconde. I would also like to author
>the DVD with minimal titles, menus and chapters.
Since you're using an AIW 8500DV make sure you install MMC7.7 because ATI
finally realized that DVD requires 48kHz audio. You don't want to encode
after capturing, so you want to capture straight to mpeg-2. I won't argue
about quality issues here, because if you're happy with the result, that's
all that matters. You want to fit 2 hours on a 4.7GB DVD-R, so you'll have
to edit the standard DVD template and set an appropriate bitrate.
According to vcdhelp a single layer DVD fits 4 700 000 000 bytes and that is
4.37 "computer" gigabytes where 1 kilobyte is 1024 bytes. 4 700 000 000 /
1024 / 1024 / 1024 = 4.377216.
You can use vcdhelp's bitrate calculator <http://www.vcdhelp.com/calc.htm>
to determine the correct average bitrate. For a 2 hour movie the correct
video bitrate is 4861 kbit/s if the audio bitrate is 224 kbit/sec. You can't
change the audio bitrate with ATI MMC, but I think MMC uses a 224 kbit/s
audio bitrate. However I'm not sure about this, nor am I sure how well MMC
maintains the selected average video bitrate, so you'll have to experiment a
little with the video bitrate settings to avoid an mpeg-2 file that is too
big to fit one 1 DVD-R (max 4.37GB).
For simple authoring I recommend Ulead DVD MovieFactory. It accepts the
mpeg-2 files captured with ATI MMC7.7. I've read that some other authoring
programs complain about the GOP structure.
If you want more information, I suggest you post your questions on the
Rage3D board <http://www.rage3d.com/board/forumdisplay.php?&forumid=24>
where you'll find lots of ATI users who can probably give you more precise
answers.
Roger H.
> about quality issues here, because if you're happy with the result, that's
> all that matters
I also recently put together this combination and I am quite
dissapointed in the quality of the capture of the supposed state of
the art ATI AIW 8500DV.
I have tried all the various settings and can't get any output that
might fit onto a DVD-R to look even close to the original qualitywise.
I was curious if anyone one has had any sucess transporting VHS/Hi-8
to DVD in a reasonably priced PC solution without a dramatic loss in
quality. If so maybe they can post some tips/advice.
It looks like stand alone units like the Panasonic DMR-e20 might be
the only way to go if you don't want to spend thousands. They claim
not only NO loss in quality but slight noise reduction that actually
gives image improvement in the DVD transfer
>Erratic <erra...@pandora.be> wrote in message news:<dpf09.30259$8o4....@afrodite.telenet-ops.be>...
Erratic
>I also recently put together this combination and I am quite
>dissapointed in the quality of the capture of the supposed state of
>the art ATI AIW 8500DV.
>I have tried all the various settings and can't get any output that
>might fit onto a DVD-R to look even close to the original qualitywise.
I agree. Real-time mpeg-2 encoding needs a hardware encoder or a much faster
PC than what we've got now.
>I was curious if anyone one has had any sucess transporting VHS/Hi-8
>to DVD in a reasonably priced PC solution without a dramatic loss in
>quality. If so maybe they can post some tips/advice.
What you might try is Half D1 resolution: 352x576 PAL or 352x480 NTSC.
That's a valid DVD format so your DVD player should play it. It's called DVD
Medium in MMC's custom settings. You'll probably have to edit it and use a
higher variable bitrate to get reasonable results. By default all ATI's
MPEG-2 settings have deintelacing switched on. You may try switching that
off, because interlaced material is supposed to look better on TV. But I've
also read that deinterlaced material produces less mpeg-2 compression
artifacts at low bitrates.
I convert my avi files to mpeg-2 with VV3. That's a lot slower, but as I
wrote in another message, with software encoding you have to choose between
quality and speed. I hate those blocky mpeg-2 artifacts, so I choose quality
over speed. As far as I'm concerned, on a regular TV screen even lo-res
352x288 PAL looks better than blocky hi-res.
>It looks like stand alone units like the Panasonic DMR-e20 might be
>the only way to go if you don't want to spend thousands. They claim
>not only NO loss in quality but slight noise reduction that actually
>gives image improvement in the DVD transfer
That's the easy way, I guess. I have no experience with standalone DVD
recorders, so I can't say anything about the quality.
victor
Thank you both for replies. Every bit of information really helps.
Hughy
basically spent all of this year in experimentation in how to get a good
result, only recently finding a method that I am satisfied with.
The secret to a good result isn't necessarily a hardware encoder. If you're
prepared to wait for a slower software encoder (such as TMPGEnc) and
experiment to find the correct settings, you should get almost the same
quality as the source material. TMPGEnc is a low cost solution (with a free
trial period).
Regards,
Hughy
"Erratic" <erra...@pandora.be> wrote in message
news:Y8E09.31912$8o4....@afrodite.telenet-ops.be...
René
"Hughy" <airways_electronics@take_this_out_bigpond.com> wrote in message
news:kiI09.45650$Hj3.1...@newsfeeds.bigpond.com...
> Amongst other things, I produce small runs of concert videos on DVD-R.
I've
> basically spent all of this year in experimentation in how to get a good
> result, only recently finding a method that I am satisfied with.
Ok, I'm waiting...
> The secret to a good result isn't necessarily a hardware encoder. If
you're
> prepared to wait for a slower software encoder (such as TMPGEnc) and
> experiment to find the correct settings, you should get almost the same
> quality as the source material. TMPGEnc is a low cost solution (with a
free
> trial period).
Very nice, but WHAT are the settings, that's what we are waiting for!?
> Regards,
> Hughy
Grtz,
René
Hughy
impossible to mathematically predict what the final file size will be.
However if you shoot similar material often, you'll get to be able to
"guesstimate" roughly what your final file size will be after a bit of trial
encoding and practice. Yes, initially you'll have to re-encode a few
projects - possibly several times. But if you *must* have quality, there
isn't really any other way (other than to spend a heap more money).
Set max bitrate to 8 Mbits/sec and min bitrate to 2 MBits/second (can get
problems with older DVD players above around 8 to 8.5 Mbits/sec). Oh ... I
forgot. I also convert my audio to .ac3 file. This is actually quite
important, as it frees up quite a lot of space that can then be dedicated to
video. You will still get good results without this, but won't be able to fit
as much on a DVD.
"CQ percentage" will depend on how many minutes of material you have to encode
and also, how much "action" is happening in the image, whether a stage
production (can mean that much of the background is black) etc. etc. However
for a start, you might try around 65 percent. You'll get satisfactory results
at 65 percent, good results at 70 percent and very good results at 75 percent.
These "ratings" are my own - satisfactory is the rating I apply when I think
that my clients won't notice any blocky artifacts. I might see an occasional
brief one. My "very good" rating is around about where I personally cannot
detect any artifact.
Motion search precision - highest quality (slowest).
Your other parameters will depend on whether you are encoding in PAL or NTSC.
So far as quantize matrix settings, I accept the default, as I don't know at
this stage the effect of these.
Next settings are all for PAL: I set number of frames in a GOP to 15. Frame
size is 720 x 576, interlaced, bottom field first. Aspect ratio 4:3, 625
lines.
I have had spectacular results lately - but they've been a long time coming.
Regards,
Hughy.
"René" <blad...@hotmail.com> wrote in message
news:ai05fe$69d$1...@test-news2.dbsch1.nb.home.nl...
René
Many thanks for the settings, I will try these!
Grtz,
René
"Hughy" <airways_electronics@take_this_out_bigpond.com> wrote in message
news:OxW09.46228$Hj3.1...@newsfeeds.bigpond.com...