You are not logged in.
im a video cruncher enthusiast. i have been working with virtualdub primarily for years.
lately, i have been trying to capture from satellite, or vhs, and author in only two swoops. im close, but still a few kinks to work out. i have the capturing setup down, i just need some suggestions on the latest method that im trying to do for authoring the dvd files.
what i am currently doing is;
step1-
capture directly from source using virtualdub and a pinnacle dv-500 capture card with breakout box via s-video/composite and on-the-fly compressing into mpeg-2@2500kbps 720x480 resolution interlaced video, ac3@128kbps audio. my system is handling this 65:1 compression ratio without losing any sync. captures are superb. i used to capture into huffyuv, make edits, then compress, but for now im trying to keep the quality without any kind of filtering.
step-2-
im trying to find a program, or command line that will help me convert these files directly to .vob's with menus and chapters. does anybody know a good program that will work for this without having to recompress the files at all?
devede problems - wont handle the mpeg-2 files very well, locks up when i try to get into the properties of the files, doesnt have success using source without forcing a recompress
dvdstyler problems - recompresses the files, i cant figure out how to use the files as is, without recompress.
k3b disk burner - i dont see any options to make a video dvd from .avi's or .mpg's
handbrake - im not sure if i can make vobs with this program, i only see mkv and mp4
as far as command line programs
dvdauthor - i think this might work, still trying to figure out how to get this to work.
dvd-shrink, again, this might work, still trying to figure out the potential on this one as well.
edit..
i do want to add, i do like atleast a semi-automated menu and chapter process, so prefer a gui for this where possible, rather than mess with xml for the menus where possible. i wouldnt mind incorporating a middle man where using a gui program just for the menu, then back to the command line for the final author.
is there any other video enthusiasts out there that has some direction?
thanks.
NOTE--
for those of you who are atleast semi-interested in this topic, once i figure this out, i will post a tutorial if possible to help simplify this process for others. :-)
Last edited by wolfdogg (2011-07-19 06:25:08)
Offline
I've done something similiar a few years ago not sure how much of the below mentioned stuff is dated. But, I think you're on the right track w/dvdauthor. Anyway some ideas....
transcode for DVD compliant video
tcrequant to shrink if need be
mplex to put the video and audio back together when needed
dvdauthor which created chapters at set intervals
mkisofs to create an ISO to burn
I had a bash script that checked the video, transcoded it, shrank it if need be, authored it, etc. Then cleaned up the temporary files once complete. I didn't care about a menu, but, I think dvdauthor can do the menu too (basic probably) but you'd have to fiddle w/an xml file each time.
Offline
well, today i found how easy the xml file is using dvd author, but i cant seem to get that far, because im running into problems further processing the captured file in any way. although its an mpeg at the correct resolution, and what i thought was the correct audio, its kicking back an error using dvdauthor, i was just trying briefly to see if the file would work there as is, then i tried mplex to see if it would demux it without recompressing the mpeg stream, but im getting errors there too.
dvdauthor error said no vobu's found on the stream, thats understandable, i guess i need to properly transcode it first with a known dvd video technique which will introduce the vobu's into the mpeg stream.
but when i ran it in mplex, it was saying that the "file is unrecognisable", so thats where im hitched up at the moment.
i would prefer to stay on this track if possible, capture as mpeg-1 or mpeg-2, transcode or remux with something like mplex, then run through dvdauthor to make the vob files.
where i need the specific help is to figure out is if 1) if im using the proper codec, and if there is any fine tuning in the settings of the codec (ffdshows mpeg-1 and mpeg-2 in virtualdub) where i might be going wrong, 2) if maybe the audio is wrong, (if 128kbps is ok, using ac3 codec, and if 48k is ok, vs. 44k) and 3) what to run that file through next, whether mplex, or what.
for those that arent sure about what im really trying to do here, dont get me wrong, there is many ways to do this, such as automated programs that will gladly recompress the file and get the job done no problem with loss of quality on the recompress, but not only does this cause quality loss, it renders useless the fact that i am already compressing during capture, and causes me to have to go back to the standard way of doing things (capture lossless at source quality with huffyuv codec) then it doesnt matter from there if a program wants to recompress the file. BUT, the reason im trying to avoid this should be obvious, by compressing to the mpeg codec on-the-fly during capture the file is about 32 times smaller allowing for many more hours to fit on the hard drive, AND is already a DVD valid mpeg stream. kills two birds with one stone which saves not only time, but eliminates a step (having to recompress the video into the mpeg stream after capture).
Last edited by wolfdogg (2011-07-16 04:29:16)
Offline
I've been wrestling for some time with video processing, with limited success (probably due to limited knowledge!!!!). In particular, tryin to get .m2t files into some kind of usable format. IMHO, something that grabs video into a valid mpeg stream is desperatley needed in Linux. The Windows drivers that came with my Hauppage card can do this successfully on my ancient Dell, so it should be possible.
Anyway, during my attempts (particularly in my pre-Arch, OpenSuSE days), I had a great deal of help from "Microchip" (http://microchip.bplaced.net/), and used ProjectX successfully for overcoming a/v sync problems. I hope this information is some use.
Offline
Use mediainfo on an example file and post it here. The output of ffprobe from FFmpeg can also be useful. This may be helpful in determining what's wrong with the stream.
Last edited by skottish (2011-07-16 15:27:43)
Offline
An ac3 file at 128kbps and 48k frequency is fine (shouldn't be your problem) at least for a dvd-ntsc compliant audio, IIRC. 44k is generally for music...
Offline
well, its my understanding that virtualdub cant process or output to mpeg file type. so maybe this is where im going wrong. im starting to get a hint that i have confused an .avi with a mpeg-2 codec for a mpeg-2 file. is there a difference? it makes sense to me after thinking about this that the .avi's interleaved file type even though the video stream is the proper format for dvd using teh mpeg-2 codec, is the improper format for authoring programs. thats leads me to wonder what exactly is a mpg file.
previous to this, i HAVE renamed my .avi files to .mpg, they still work, and it doesnt change the errors at all. this doesnt mean much i know.
well, let me get the media info, i did that yesterday using mplayer info, and the avi with mpeg codec looks quite a bit different from from an actual mpg that i processed years ago using the pinnacle studio suite.
heres an .avi stream just captured using an ffdshow mpeg-2 codec with fourcc of MPG2 which is the closest i can figure so far to capture correctly in the proper format
# mediainfo test_mpeg-2.avi
General
Complete name : test_mpeg-2.avi
Format : AVI
Format/Info : Audio Video Interleave
File size : 685 KiB
Duration : 5s 239ms
Overall bit rate : 1 072 Kbps
segm : c:\Vdub\Caps\
Video
ID : 0
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : No
Format settings, Matrix : Default
Codec ID : MPG2
Codec ID/Info : (MPEG-1/2) FFmpeg
Codec ID/Hint : Ffmpeg
Duration : 5s 239ms
Bit rate mode : Variable
Bit rate : 915 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 3:2
Frame rate : 29.970 fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Bottom Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.088
Stream size : 585 KiB (85%)
Audio
ID : 1
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : 2000
Duration : 5s 184ms
Bit rate mode : Constant
Bit rate : 128 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Compression mode : Lossy
Stream size : 81.0 KiB (12%)
Alignment : Aligned on interleaves
Interleave, duration : 32 ms (0.97 video frame)#ffprobe
Input #0, avi, from 'test_mpeg-2.avi':
Duration: 00:00:05.23, start: 0.000000, bitrate: 1071 kb/s
Stream #0.0: Video: mpeg2video (Main), yuv420p, 720x480 [PAR 1:1 DAR 3:2], 104857 kb/s, 29.97 fps, 29.97 tbr, 29.97 tbn, 59.94 tbc
Stream #0.1: Audio: ac3, 48000 Hz, stereo, s16, 128 kb/sand here is the old .mpg file that i think is a true mpeg file, ignore the audio portion, since its a mp3 i think
# mediainfo test.mpg
General
Complete name : test.mpg
Format : MPEG-PS
File size : 5.81 MiB
Duration : 8s 541ms
Overall bit rate : 5 709 Kbps
Video
ID : 224 (0xE0)
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : Yes
Format settings, Matrix : Custom
Format settings, GOP : M=3, N=15
Duration : 8s 541ms
Bit rate mode : Variable
Bit rate : 5 403 Kbps
Nominal bit rate : 6 000 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 4:3
Frame rate : 29.970 fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Top Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.522
Stream size : 5.50 MiB (95%)
Audio
ID : 192 (0xC0)
Format : MPEG Audio
Format version : Version 1
Format profile : Layer 2
Duration : 8s 232ms
Bit rate mode : Constant
Bit rate : 192 Kbps
Channel(s) : 2 channels
Sampling rate : 48.0 KHz
Compression mode : Lossy
Delay relative to video : -66ms
Stream size : 193 KiB (3%)i can see that the video stream is almost identical however, so it seems that this is possible. it seems that if i demux the file that i will be left with a near proper mpeg stream. i do see now that the BVOP setting is different, i need to look into this more, although i have brushed over this in the past, i dont know what it is offhand. and the field order is opposite also. im going to look for these settings in teh codec, and do some research on BVOP for starters.
oh yes, also, i remember reading something in the mplex or dvdauthor man, that the ID is used also when the stream is read, and there is an ID of 0 on the avi file, this might be a problem.
Last edited by wolfdogg (2011-07-16 21:11:10)
Offline
The second one has MPEG 2 audio, which is part of the DVD specification. It sounds like crap, but it works.
Your top file is an avi file and dvdauthor doesn't deal with those. You have a 3:2 ratio in your example. DVD only supports 4:3 and 16:9 ratios. GOP is important for successful playback and I don't see any evidence of it being set (GOP 12 is a reasonable standard). The frame rate is good and you have interlaced content, but I'm a little concerned of the "Bottom Field First" of the "Scan Order". I can't remember if that's part of the NTSC DVD standard or not. Whatever encoded your video spelled FFmpeg wrong and that never inspires confidence
In short, the settings on your encoder are not generating NTSC DVD compatible videos.
Last edited by skottish (2011-07-16 21:20:47)
Offline
ok, and i concurr with this, and i think i have something to go with now.
i noticed the 3:2 ratio, and i dont know how it derives this since they are both 720x480. do you know why?
i found that i can convert the file using ffmpeg -i -target , but i have yet to look if it recompressed the video, this is one of my requirements, that the video stream is NOT recompressed lossy in any shape or form.
i had a bit of success with this converting the 2nd or bottom file using this;
ffmpeg -i test.mpg -target ntsc-dvd test.vob.mpgthen when i ran this through the dvdauthor it successfully create the dvd. this is farther than i got last time. but, the issue is not with this file so this is just for reference.
so regarding the top file, which is my virtualdub captured file, i tried to run it through this ffmpeg -i test_mpeg-2.avi -target ntsc-dvd /tmp/test_mpeg-2.mpg and it said
test_mpeg-2.avi: Invalid data found when processing input so i will see what i can do with the
a) ratio
b) GOP
c) field order (ill leave the field order for last, since some processes do account for this and adjust accordingly)
as far as the spelling, you mean the [Codec ID/Hint ]? maybe thats just a data field that needs only first case upper only? it is spelled correct in the [Codec ID/Info ] atleast i think.
Last edited by wolfdogg (2011-07-16 22:15:56)
Offline
ooh, look at this
$ mediainfo test-mpeg-2-bframes.avi
General
Complete name : test-mpeg-2-bframes.avi
Format : AVI
Format/Info : Audio Video Interleave
File size : 568 KiB
Duration : 3s 804ms
Overall bit rate : 1 223 Kbps
segm : c:\Vdub\Caps\
Video
ID : 0
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : Yes
Format settings, Matrix : Default
Codec ID : MPG2
Codec ID/Info : (MPEG-1/2) FFmpeg
Codec ID/Hint : Ffmpeg
Duration : 3s 804ms
Bit rate mode : Variable
Bit rate : 1 068 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 3:2
Frame rate : 29.970 fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Bottom Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.103
Stream size : 496 KiB (87%)
Audio
ID : 1
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Codec ID : 2000
Duration : 3s 584ms
Bit rate mode : Constant
Bit rate : 128 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Compression mode : Lossy
Stream size : 56.0 KiB (10%)
Alignment : Aligned on interleaves
Interleave, duration : 34 ms (1.02 video frame)note, the BVOP is now ON, how i did this was in the compression settings;
i turned on b-frames, max consecutive bframes to 1, dynamic checked, refine unchecked, closed GOP unchecked , threads to 1
$ ffmpeg -i test-mpeg-2-bframes.avi -target ntsc-dvd test-bframes-vob.mpg
ffmpeg version N-30954-g1986380, Copyright (c) 2000-2011 the FFmpeg developers
built on Jun 22 2011 13:28:34 with gcc 4.6.0 20110603 (prerelease)
configuration: --prefix=/usr --enable-libmp3lame --enable-libvorbis --enable-libxvid --enable-libx264 --enable-libvpx --enable-libtheora --enable-postproc --enable-shared --enable-x11grab --enable-libopencore_amrnb --enable-libopencore_amrwb --enable-libschroedinger --enable-libopenjpeg --enable-librtmp --enable-gpl --enable-version3 --enable-runtime-cpudetect --disable-debug
libavutil 50. 43. 0 / 50. 43. 0
libavcodec 52.122. 0 / 52.122. 0
libavformat 52.110. 0 / 52.110. 0
libavdevice 52. 5. 0 / 52. 5. 0
libavfilter 1. 80. 0 / 1. 80. 0
libswscale 0. 14. 1 / 0. 14. 1
libpostproc 51. 2. 0 / 51. 2. 0
Input #0, avi, from 'test-mpeg-2-bframes.avi':
Duration: 00:00:03.80, start: 0.000000, bitrate: 1223 kb/s
Stream #0.0: Video: mpeg2video (Main), yuv420p, 720x480 [PAR 1:1 DAR 3:2], 104857 kb/s, 29.97 fps, 29.97 tbr, 29.97 tbn, 59.94 tbc
Stream #0.1: Audio: ac3, 48000 Hz, stereo, s16, 128 kb/s
[buffer @ 0x1bbb1c0] w:720 h:480 pixfmt:yuv420p tb:1/1000000 sar:1/1 sws_param:
Output #0, dvd, to 'test-bframes-vob.mpg':
Metadata:
encoder : Lavf52.110.0
Stream #0.0: Video: mpeg2video, yuv420p, 720x480 [PAR 1:1 DAR 3:2], q=2-31, 6000 kb/s, 90k tbn, 29.97 tbc
Stream #0.1: Audio: ac3, 48000 Hz, stereo, s16, 448 kb/s
Stream mapping:
Stream #0.0 -> #0.0
Stream #0.1 -> #0.1
Press [q] to stop, [?] for help
frame= 114 fps= 0 q=2.0 Lsize= 1016kB time=00:00:03.58 bitrate=2322.3kbits/s dup=1 drop=0
video:785kB audio:196kB global headers:0kB muxing overhead 3.584905%Last edited by wolfdogg (2011-07-16 22:28:49)
Offline
The spelling thing was a bit of a joke and has nothing to do with successful encodes.
It's been a long time since I've done DVD authoring, but I think that I manually padded 720x480 ratio videos to 720x576 using FFmpeg. 720x480 is a standard resolution as long as it's letter-boxed, but I never figured out a way to do it without adding the borders and still having the video not stretched. The FFmpeg options of GOP and ratio are straight forward:
-g 12
-aspect 4:3 or -aspect 16:9
Offline
i think we posted at the same time, check out the last post! partial success with b-frames turned on, i noticed this because it has the GOP setting in it mentioned above. thanks for that!
ahh, ok, ill check into the -g 12, incase i have problems on playback, do you think i should just shove that setting in there? ok, im excited, i want to check whether the file was recompressed or not, it looks like the file is just about double of what it started with, so im assuming it just doubled the frame rate, be back.
---------
edit
well, the bit rate went from 1067kbps to 1690kbps after running it through FFmpeg, do you think that was a lossy recompress, or is that just from the frame changes? the cool thing is it is showing as progressive now in gspot, not interlaced.
oh yes, and heres what its looking like after running it through ffmpeg as you see in that last code snippet above. it looks like the GOP is now there too, so probably wont need to add that to the command after all
$ mediainfo test-bframes-vob.mpg
General
Complete name : test-bframes-vob.mpg
Format : MPEG-PS
File size : 1 016 KiB
Duration : 3s 804ms
Overall bit rate : 2 188 Kbps
Video
ID : 224 (0xE0)
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : No
Format settings, Matrix : Default
Format settings, GOP : M=1, N=18
Duration : 3s 804ms
Bit rate mode : Variable
Bit rate : 9 000 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 3:2
Frame rate : 29.970 fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.869
Stream size : 4.08 MiB
Audio
ID : 189 (0xBD)-128 (0x80)
Format : AC-3
Format/Info : Audio Coding 3
Mode extension : CM (complete main)
Muxing mode : DVD-Video
Duration : 3s 584ms
Bit rate mode : Constant
Bit rate : 448 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Compression mode : Lossy
Stream size : 196 KiB (19%)
Menu----
edit
and we HAVE .vob's!
it successfully ran through dvdauthor and created the fiel structure.
so what i need to know if that was a losless transistion when i ran it through ffmpeg or might have all been a useless way to make a mpeg-2 file, then recompress it as such, lol.
Last edited by wolfdogg (2011-07-16 22:53:06)
Offline
FFmpeg definitely re-encoded it. Although, if the bitrate is much higher, you didn't lose much (or any) data. Of course, this is a waste if you can do it the first time.
I'm curious if the audio and video are synced on your copy. I've had terrible luck with dealing with FFmpeg and MPEG2 video and AC3 audio as one file. I always had split them until I was ready to author.
Offline
ill check the audio and video sync when i do an actual video, but there is a way to set the delay when authoring i think, once you have this offset number, it will probably be the same everytime.
ok, so you think ffmpeg re-encoded it? i wasnt sure if it just split the streams, and added some information, since it was an avi, the streams were already seperate, so im not completely sure it had to resample the video. i was hoping that it just split it using the interleave table, and added some information to it, rather than resampling it. i need to know if this is lossles. or not. this will determine the best way to use this process, and if i need to capture differently.
for instance, if this is lossy;
1)if i am to be able to fit more than 12 hours of capture time on my capture hard drive, i will need to download and capture using the mjpeg codec, because i hear that this resamples well to the mpeg-2 format without much loss. i think this will be higher quality than capturing at mpeg-2, them resampling back to mpeg-2
2)otherwise i will have to go back to capturing to huffyuv (losless) then processing it.
how do we find out if the ffmpeg conversion re-sampled the pixels or not, as opposed to just adding more frame information?
Offline
The way to use lossless in FFmpeg would be to pass -vcodec copy and -acodec copy to the command line. Your results may not work out with this method.
I highly doubt that you're losing a significant amount of data if you are at all. My guess right now is that FFmpeg is up-sampling it to make room for the extra information. The simplest way to get excellent answers to questions like this is to join either ffmpeg or libav at freenode (IRC). There's almost always FFmpeg developers watching those channels.
Offline
ok, so this is what need to be found out then. ill see if i can figure that one out.
im looking at the ffmpeg process as noted above in the code snippet, and this portion of the line looks like its suggesting that the video was resampled using the quantizer method, does it look like that to you?
Stream #0.0: Video: mpeg2video, yuv420p, 720x480 [PAR 1:1 DAR 3:2], q=2-31, 6000 kb/s, 90k tbn, 29.97 tbcand
frame= 114 fps= 0 q=2.0 Lsize= 1016kB time=00:00:03.58 bitrate=2322.3kbits/s dup=1 drop=0if im not mistaken i used the quantizer of 2 to capture that one instead of the bit rate method, so im wondering if thats just refelcting the source. do you know more about that process output info? are those settings that reflect the stream, or are those the actual settings that its resampling with?
Last edited by wolfdogg (2011-07-17 00:48:12)
Offline
watching closer, i notice that the quantizer is an active process, after processing a larger file, it was variable, changing during the process between q 2.0 to q 2.5 or something like that, it finally settled out at q 2.0, but this capture was q 1 for a higher quality, so its nto reflecting the capture, its reflecting the resampling. hmm not i guess i just need general opinion from the ffmpeg folks if this is causing much loss, or the improper way to do this.
Last edited by wolfdogg (2011-07-17 01:07:25)
Offline
The target switch will automatically use the quantizer method to attempt to produce the best possible stream and that's what's happening here. So, what you're seeing in the output is what FFmpeg is encoding and not what it was decoding.
--EDIT--
We always seem to be writing at the same time.
Last edited by skottish (2011-07-17 01:11:32)
Offline
ok, i have been messing with this today again, and i have been able to consolidate my thoughts.
if i capture lossy, and then have to re-encode, it causes generation loss. i dont want this.
if i capture lossy, then transcode to losless, its a destructive process, because the file will inevitably have to be re-encoded, i dont want this.
if i capture lossy, but at the proper format which also includes all the necessarry flags that dvd author is looking for, i wont need to re-encode, and will have my problem completely solved. I would then be able to convert that .avi capture file into .vob's using dvdauthor and the dvd will be complete for burning.
So, the only solution is to identify EXACTLY what flags are missing from the capture and figure out how to either enable them in the mpeg-2 encoder settings during capture, or figure out a process that can successfully introduce them into the .avi just after capture, and before dvdauthor.
Skottish, you mentioned the GOP(group of pictures), and i believe this is related to B frames, and I frames, etc... i did notice that when i enabled B frames in the encoder it almost got me home. But there is atleast 1 more thing missing before dvdauthor will take the file. its complaing about the elusive vobu's. what is that? i think they are blank spaces for vob information set into the stream at intervals. How to i set up the encoder to supplement this, or what process can be added to add them lossless?
heres the data so far, and its just copied from code the snippets above, just shortened for easier viewing
before; (this is the capture that wont run in dvdauthor, needs to go through the ffmpeg ringer first)
Video
ID : 0
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : No
Format settings, Matrix : Default
Codec ID : MPG2
Codec ID/Info : (MPEG-1/2) FFmpeg
Codec ID/Hint : Ffmpeg
Duration : 5s 239ms
Bit rate mode : Variable
Bit rate : 915 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 3:2
Frame rate : 29.970 fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Bottom Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.088
Stream size : 585 KiB (85%)after;(this is the re-encoded version that successfully will convert to vob's using dvdauthor, its actually a different file altogether, so ignore the sizes)
Video
ID : 224 (0xE0)
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : No
Format settings, Matrix : Default
Format settings, GOP : M=1, N=18
Duration : 3s 804ms
Bit rate mode : Variable
Bit rate : 9 000 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 3:2
Frame rate : 29.970 fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.869
Stream size : 4.08 MiBi need to figure out what exactly the diffeence is between the two.
ps, i found the setting for top field first, so thats not a problem, that will be enabled on my next test capture
Last edited by wolfdogg (2011-07-18 10:29:43)
Offline
You used FFmpeg's -target ntsc-dvd on the second one, right? That will create the VOBU's in the correct places. In fact, it looks as if FFmpeg handled everything including the GOP, except for the aspect ratio. That file is progressive, but I don't think that any modern DVD player cares.
Offline
well, not sure, but dvdauthor might fix that. yes, the 2nd one represented the file after it was processed with ffmpeg.
ok, i have refined this a bit, and found a few crucial settings in the encoder, and also asked a few questions in teh ffmpeg irc. they dont exactly answer questions there, but if you wait an hour you will a one liner now and then, lol.
edit.. - ALSO, i found a way to get ffmpeg to direct copy the video codec, without recompress [-vcodec copy] see below..
so, this is what i got.
Virtual Dub
capture mode
resolution - 720x480
video compression settings;
ffdshoww video codec
encoder mpeg-2 FOURCC MPG2
bit rate (one pass - average bit rate)
your choice, im using 6000kbps for capture currently
b. frames on - max consecutive b frames 1
interlaced encoding on
top field first
Output settings;
pixel aspect ratio 40:33 (ntsc 16x9)
audio
48khz raw capture format
compression settings
ac3-128kbps(probably will switch to 190kbps on following runs)
mediainfo in.avi
Format : AVI
Format/Info : Audio Video Interleave
File size : 12.9 MiB
Duration : 24s 320ms
Overall bit rate : 4 453 Kbps
segm : c:\Vdub\Caps\
Video
ID : 0
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : Yes
Format settings, Matrix : Default
Codec ID : MPG2
Codec ID/Info : (MPEG-1/2) FFmpeg
Codec ID/Hint : Ffmpeg
Duration : 19s 486ms
Bit rate mode : Variable
Bit rate : 5 378 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Top Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.519
Stream size : 12.5 MiB (97%)ffmpeg
ffmpeg -i in.avi -vcodec copy -acodec copy -target ntsc-dvd out.vobmediainfo out.vob
Format : MPEG-PS
File size : 13.0 MiB
Duration : 24s 320ms
Overall bit rate : 4 492 Kbps
Video
ID : 224 (0xE0)
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, BVOP : Yes
Format settings, Matrix : Default
Duration : 2s 769ms
Bit rate mode : Variable
Bit rate : 4 275 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 fps
Standard : NTSC
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Interlaced
Scan order : Top Field First
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.413
Stream size : 1.41 MiB (11%)dvdauthor
dvd.xml file contents;
<dvdauthor>
<vmgm />
<titleset>
<titles>
<pgc>
<vob file="/path/out.vob" />
</pgc>
</titles>
</titleset>
</dvdauthor>export VIDEO_FORMAT=NTSC
dvdauthor -o /output/dir -x /dvd.xmlresults a seeming working dvd dir structure, ready to burn.
so we have a 3 step capture to dvd structure, and we have a compressed capture allowing for about 6 hrs capture time on 200GB free space. this allows alot of flexibility for capture time, thats a complete 6 hr VHS tape if archiving, a trip across town for the day and back, or whatever your poison is for capturing for long durations.
sync issues
i must note some crucial settings that i struggled with for a long time that caused sync problems for years in vdub. if you have a fast system, atleast in all my setups, you MUST turn off the automatic resync in video timings since this is responsible for the bulk of sync problems. and if your capturing vhs and its bad quality, you probably want to do some experimenting with the 'drop frames' and 'insert null frames' because this can potentially mess up the timings as well, just not sure if on or off is better for either.
im going to test some real captures now, and then test the output for sync, seems good so far, and then burn test
Last edited by wolfdogg (2011-07-19 06:00:53)
Offline
I know that it's only been a few days, but I'm curious how this has turned out for you.
Offline
thanks for the interest. Well it turned out well. I am able to run the avi, that was captured properly with mpeg-2 codec, and acm codec's , on-the-fly, then edit out commercials etc... as just do a stream copy, instead of a recompress, therefore that edit is done in seconds and is losless, then after that, i send it through the convert to dvd process that we created, which only takes a minute assuming the video is long, otherwise it takes a few seconds.
I think im on to something here. The most important things to this whole process is to capture properly(dvd format codecs), and then that (-vcodec copy) command option for ffmpeg.
The only thing that i havent done yet is menus. This is primarily because i recorded about 8-10 hrs of the last space shuttle mision sts-135 and have a ton of editing to do. Im mostly finished with that, and ready for converting to dvd soon.
I will need to do some burn tests to figure out the menu's still.
Last edited by wolfdogg (2015-09-25 04:26:26)
Offline
i did stumble across this today when running a dvd through dvdshrink. I checked out the way it handled the chapters on the dvd.xml. this looks like the proper way to do handle the chapters. ill just have to merge this format into my bash script.
<dvdauthor dest="/tmp/PAUL/BUILD">
<vmgm />
<titleset>
<titles>
<audio lang="en" />
<pgc>
<vob file = "/tmp/PAUL/PAUL.mpg" chapters="00:00:00.000,00:05:25.200,00:11:20.033,00:16:51.866,00:23:52.699,00:28:56.666,00:31:22.666,00:36:48.367,00:40:52.167,00:43:55.034,00:51:10.201,00:56:27.201,01:00:38.667,01:06:57.366,01:11:37.699,01:18:23.666,01:21:01.866,01:26:47.866,01:30:40.698,01:35:23.865,01:43:42.032" />
</pgc>
</titles>
</titleset>
</dvdauthor>Last edited by wolfdogg (2015-09-25 04:38:51)
Offline