View Full Version : MPEG-2 encoding question
neil wilkes
4th October 2012, 14:11
I've seen this done on commercial releases so I know it is possible - but how to do it?
MPEG-2 file (m2v) of a 14-track album running at 68 minutes, yet is consisting of just a single I-Frame image for each song (chapter) thus greatly reducing the size of the MPEG-2 file.
It was done on Bjork's "surrounded" box set, and used to be often done on Audio-only DVD-Video discs.
Encoders available to me are CCE-SP3, Sony Vegas Pro 11 and Adobe's CS6 Media Encoder (through Premiere & standalone)
Would greatly welcome any suggestions as to how I do this. Seems to me the trick is in telling the app how long the chapter is - and it must be possible to access individual tracks, so chapter marker placement is a must.
smok3
4th October 2012, 19:32
well, I dont remember technical details, but its about importing bunch of images into encore..., a task for authoring app really.
rik1138
4th October 2012, 22:04
You can encode a still image with a duration... That's probably how it's done. I believe Scenarist (for DVD) will do this for you, you just drop the image in the timeline and tell it how long it's to be displayed.
For Blu-ray, Scenarist comes with a still-image encoder that will create the files for you.
For tools other than Scenarist, there should be ways... Look for settings like 'still with duration' or 'still image timeout' maybe... But it is something that's likely to be done in the authoring software using a .TIF image (or whatever format) as the source.
SeeMoreDigital
5th October 2012, 09:26
You can add duration to (.bmp) images using MPUCoder's MuxMan (http://www.mpucoder.com/) DVD authoring application...
neil wilkes
5th October 2012, 10:11
Thanks for the suggestions, but Scenarist is NFG for this one. It just does not work that way.
All I would be able to do is add a bunch of single image timelines with Audio - not good enough.
2 reasons.
1 - Spec violation for importing Video_TS into a DVDA app. (All titles in a VTS shall be contiguous)
2 - Segued tracks might glitch at track point, as Scenarist cannot handle seamless slideshows with multiple audio streams.
Best bet is DVD Lab Pro, which creates the right type of stream - even though cells are not set seamlessly PGCEdit can do this, but not from Scenarist output - even Sonic tell me what I want to do is not possible in Scenarist, only DVD Creator. Trouble with DLP is that it has some nasty bugs.
All I want to do is create an MPEG-2 file with single images that will not be a monstrous size.
Obviously it is not possible. Strange.
SeeMoreDigital
5th October 2012, 10:45
Hmmm....
Here's a link to a 60 second sample I've just created using MuxMan: http://www.sendspace.com/file/zxbmsf
Guest
5th October 2012, 12:44
All I want to do is create an MPEG-2 file with single images that will not be a monstrous size. Obviously it is not possible. Strange. It's easy to create an MPEG2 stream with a single I frame, but there is nothing in MPEG-2 syntax that allows signaling to a player that a frame should be repeated for N frames where N is arbitrary.
bigotti5
5th October 2012, 13:53
Do you have access to Spruce Maestro?
Place bitmaps to timeline, drag to preferred length, set chapters and use "Create Sync Audio Track" to create audio.
Emulgator
6th October 2012, 12:07
I did such projects, 2 for test reasons, 1 for a customer in DVDLabPro2 some years ago.
DLP 2.19 for a 84-tracks Audio-Only with stills DVD.
DLP 2.24 for a 12-Album Audio-Only with stills DVD.
DLP 2.51 for a Live Concert Audio-Only with stills DVD.
It worked on various players back then and still does.
The project files are still there.
Depending on bug status of the DLP muxer some PGCEdit work (Seamless flag ? IIRC) was to be done.
But you will have to let the still image encoder in DLP do the encoding, which is a bit suboptimal.
The Muxman still encoder is said to be improved for the recent versions from MuxMan 1.4 on.
My version 1.4.3 seems borked, touching Navigation -> VTS01 aborts MuxMan, so I did not try yet...
P.S. Expanding VTS01 works though, and the encoding result is not too bad.
Link coming soon, testing better images...
Late Edit:
Yes, if stills are brought in perfectly resized at final resolution
the MuxMan outcome is quite good, comparing to DLP.
If Muxman's internal resizer shall be bypassed,
I suggest to use FritzPhoto, so AviSynths choice of comprehensive resizers for that..
SeeMoreDigital
6th October 2012, 12:49
But you will have to let the still image encoder in DLP do the encoding, which is a bit suboptimal.
The Muxman still encoder is said to be improved for the recent versions, I did not try yet...I've done the same thing with Muxman...
You can drop several 5.1 Ch AC3 audio files with different track run-times into MuxMan. Then add your different (720x576) .bmp image files. Then allocate the required run-time data to each image file to match the run-time of the audio file. Then add chapter times to match the run-time of the audio and image files. Then mux...
As long as your images have been created correctly (ie: to allow for DAR changes), it's a simple muxing procedure.
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.