This forum is for users to exchange information and discuss with other users about a TMPGEnc product.
In case you need official support, please contact TMPG Inc.
Pegasys Products BBS [ Sorted by thread creation date ]
I am still searching for a way to create quality mpeg2 from (interlaced) dv sources.
If I leave my dv sources (high quality) interlaced the quality of the resulting (interlaced) mpeg2 is very poor (on movements which is clear). Deinterlacing helps but I would like to keep the videos interlaced since I intend to view the mpeg2 on a regular tv only.
Any idea about a way to achieve this?
To my knowledge mpeg2 is designed for interlaced video. Don't know how this works technically. If it's based on frames I am not surprised about the low quality since interlacing artefacts will be difficult to compress. So it should be on a field base instead. Not sure if that's the way it works. I looked for answers on the web but everyone just looks for the best de-interlacer which is in my eyes not necessarily the right solution.
If your having problems with artifacts and you can't perform IVTC then the only way to solve them is going to be some sort of de-interlacing.
Evev if you did de-interlace the movie the final output MPEG2 would still be interlaced.
All you need to do is de-interlace the movie prior to encoding to remove the artifacts.
Also it seems you may be using the wrong kind of filters.
If you wish to maintain the best quality of the AVI then your best bet is to use the AVIsynth 'BOB' filter. This is probably the best de-interlacing filter there is and gives very good results, it's far better than TMPGs filters.
My ntention is to avoid deinterlacing.
I know that if you look at both fields from an interlaced frame in a still capture the result looks like the samples on your site. But when watching the interlaced video on a regular tv you won't see any artifacts in the (interlaced) dv source but in the interlaced mpeg2.
I know that most people try to ged rid of interlacing because of the pour mpeg2 results. But I would like to keep the smoothness in motion of the original interlaced dv source while avoiding compression atrifdacts resulting from interlaced video source.
My ntention is to avoid deinterlacing.
I know that if you look at both fields from an interlaced frame in a still capture the result looks like the samples on your site. But when watching the interlaced video on a regular tv you won't see any artifacts in the (interlaced) dv source but in the interlaced mpeg2.
I know that most people try to ged rid of interlacing because of the pour mpeg2 results. But I would like to keep the smoothness in motion of the original interlaced dv source while avoiding compression atrifdacts resulting from interlaced video source.
I agree with you that the interlacing artifacts will not be visible on a normal TV, but seeing as it is these artifacts which are causing the problem then the only alternative seems to be to remove them the best you can.
I would try the BOB filter. I think you will be happy with the results.
As for IVTC (Inverse Telecine) you would use this on a source which has had pulldown applied to it.
For example most DVD movies are originally 24 fps then have telecining applied to turn the movie into 29.97 fps by usage of flags for frame duplication.
This effect creates an interlaced output, but these duplicated frames can be removed by either DVD2AVI's ForcedFilM option or IVTC.
This returns the source back to a progressive frame movie and therfore no interlacing artifacts.
Depending how your AVI has been created you may or may not be able to use IVTC.
Hi Ashy,
IVTC is not an issue for me since we've got PAL in good ol' Europe!
The minutes we get from playing movies 1/25 faster are typically used for playing another Marlboro commercial in movie theatres.
Currently I am using Tmpgenc's built-in deinterlacing. Both double and upper-lower with decent results.
Nevertheless I wouldlike to stay as close as possible to the dv source and I thought this should be possible in MPEG2. Maybe I have to dig into more detail to see whether MPEG2 allows to compress upper and lower fields separately thus avoiding the loss during compressing interlaced frames.
I'm surprised you are having any problems at all with PAL sources. I too am from Europe and deal with mostly PAL sources.
I have never had any compression problems where interlacing is concerned. Maybe DV is different, but I cannot see why. I don't deal with DV much so maybe this is out of my league, but I don't see how this can be any different to other high quality sources.
Are you sure you have the field order set correctly when encoding?
What are you using as your visual reference, your T.V. or your Monitor.
If you are using your monior as your reference then this is a big mistake.
Also what software are you using to view your MPEG2?
You do realise most MPEG2 software players incorporate a de-interlacing filter which will cause some blending and motion artifacts.
First of all my reference always is TV (100 Hz Sony) since I want to watch my edited dv's on tv not pc - I am aware of CRT's being non-interlaced.
I guess you always encoded (interlaced) DVD sources. Whether a dvd is interlaced or non-interlaced doesn't change anything because both fields match perfectly in a frame since the source typically is a 24fps movie. This is at least true for PAL dvd's. (For NTSC you do have interlacing probs due to 3:2 pulldown.) Thus ripped DVD's converted to svcd compliant mpeg2 do not have those artefacts that I am talking of.
DV is different because consumer cams take only one field of a frame at a time. That means you have a difference between both fields due to motion. For compression this is bad. For a TV it's great because this results in smooth motions and high res stills (at least in terms of current tv standards). That's what I want to maintain.
I regularly encode material captured from a HDTV receiver which is similar in resolution to DV (even after down conversion) and has lots of field-to-field motion.
This compresses with virtually no loss. Your troubles with DV must be caused by some bad link in the chain which needs to be isolated.
Average camera scenes are a piece of cake for Mpeg compression. Even severe camera shake is easily detected as a simple motion vector. And yes, the Mpeg 2 standard added native support for field based encoding.
Make sure you completely understand the field order issue. If the footage looks fine on WinDVD type players (which deinterlace), but bad on a TV, the field order is set wrong. (I think DV is bottom first.) The settop DVD player must know which field comes first in time since either is possible.
Beyond that could be bitrate issues. Try 2-pass VBR with 6Mbs ave, 8Mbs max.
Other possibilities: The DV decoder in use is poor. Coversion from 4:1:1 to 4:2:0 color sampling introduces loss.
I agree wcpaul. You are thinking along the same lines as myself.
I wouldn't have thought there would be issues with compression with regards to interlaced DV footage.
This would sort of defeat the purpose of capturing to a high quality source such as DV wouldn't it.
There must be other factors at work here which could be causing these problems. The field order issue being the first one to look at and I agree bitrate shortage.
As regards the DV decoder. I'm not sure what format you DV AVI is, but if possible maybe it would help to convert the DV AVI to Canopus DV file format with the Canopus DV File Converter. http://www.dvideohome.com/encoding.htm
Thanks for your feedback.
I capture and edit with Premiere and export lower field first. All settings in Premiere look the same, no mismatch in the project settings. Via avisynth (just for reading from IPC) I frameserv to Tmpgenc. Reading and compressing again lower field first (Does the dvd player read lower first?).
Bitrate is always limited to svcd standard (or 3Mbit max. for svcd with my Pio player). But that's no issue since deinterlaced in Tmpgenc the same clip looks fine on svcd.
I think I'll have to dig deeper into.
Wcpaul are you sure that mpeg2 compresses on a field level?
>Wcpaul are you sure that mpeg2 compresses on a field level?
It has to. Video fields occur at different points in time. It would be wasteful (impossible?) to apply motion detection and DCT algorithms to 8x8 blocks with interleaved fields. Remember, DCT works because most 8x8 blocks contain very simple patterns.
Since my main goal is to encode with zero loss, I have spent much time comparing pre and post encode video in every possible way. Seperating the fields with AVIsynth or Virtualdub (or bob deinterlacing) shows that the encoded video can reproduce each field with no loss even at low bitrates.
Here is a new experiment I have never tried: Encode interlaced frames as progressive and see if the individual fields are faithfully reproduced after seperation. Do this at a bitrate that just satisfies the interlaced case.
I actually experienced a similar problem trying to encode a 50 fps progressive video to 25 fps interlaced, and just find the solution.
My source video is a computer animation, the 1993 Amiga's "Desert Dream" demo, if you ever heard about it :-) I recorded it using WinUAE.
This video is made of very fast animations scene, where each of the 50 fps is different.
My concern was to keep the animations as smooth as the original, because this is precisely the smoothness of the animations that made those demo so striking.
My source video was 720 * 576 at 50 fps.
The original resolution is actually 720 * 288, at 50 fps, but on my video, each field have been doubled (even field is the same as odd field), giving the progessive 720*576 source.
What I wanted is a 25 fps interlaced video using odd field of frame 1, and even field of frame 2, to keep the animation as smooth of the original when watching on TV (I don't care of watching it on PC).
My first try was the open the 50 fps AVI file in TMPGEnc, with source type "Non-Interlaced (progressive)" and target format 25 fps "Interlaced".
I though TMPEGEnc would render the interlaced fields using odd field of source frame 1 (at 1/50th second) and even field of source frame 2 (at 2/50th second).
This would recreate the original smoothness of the video, but this is not what TMPEGEnc made.
It simply skipped even frames from the source video, and encoded odd frames as interlaced. :-(
I finally resolved my problem using AVISynth and the following AVS file:
==============
AVISource("DesertDream.avi")
AssumeFieldBased
VerticalReduceBy2
Weave
==============
It produce a 25 fps interlaced stream from my 50 fps progressive video.
If the source video was 720*288, "VerticalReduceBy2" would be unnecessary.
I opened the AVS file on TMPEGEnc, with the source type as "Interlaced (Bottom Field first)", and target type as "Interlaced"
I checked the result on my TV, and it was perfect. :-)
Ashy was true concerning the preview of the result using TMPEGEnc.
If the result is interlaced, TMPEGEnc preview use a "Smoth Deinterlace" methode, preventing to check whether result is correclty interlaced.
To be sure your result is interlaced, you can use Mpeg-VCR, from Womble Multimedia, the picture quality is bad, but the scanlines are showed without any filter :-).
I hope this will help you, because I searched myself 2 days to find the solution, mainly because i checked the result inside TMPEGEnc, I though my (correct) result was wrong since 2 days...
I wanna know how i can let an mpeg file with more of 80 minutes in 79 minutes. That's so i can burn with nero in one cd of 700 MegaBytes. Cuz i have some vcd's with a length of 90 minutes or 100 minutes in one cd and it seems like the length was it of 79 minutes or 78 minutes aprox.
So the point is burn in 700 MB cd's mpeg files with more of 80 minutes and not cut the movie and burn it in two cd's and please don't tell me there is no way cuz i have a lot of vcd's in one cd
of 700 MB
First of all it would help if you talk English.
Gathering what I can from the broken gibberish you have posted I'm guessing you want to fit a movie which is longer than 80 mins on to an 80 min CDR.
You cannot fit most standard VCD movies on to one disk. If your movie is say 100 mins long and larger than 800mb then you won't be able to fit it on one cd.
You say you have movies which are 100 mins long yet are on one cd? Yes this is possible, but they are movies which have been encoding with a lower bitrate to accomplish this.
This makes the movie smaller so you can fit it on 2 cd, but this also makes the movie look absolutely awful and I for one wouldn't even consider it, but if size is more important than quality to you then the only way you can fit you movies on to one cd is to re-encode them at a lower bitrate or use the so called 'KVCD templates' http://www.kvcd.net
You can use a bitrate calculator to find out what bitrate you need to set to fit the movie on one cd.
>This makes the movie smaller so you can fit it on 1 cd, but this also makes the movie look absolutely awful and I for one wouldn't even consider it, but if size is more important than quality to you then the only way you can fit you movies on to one cd is to re-encode them at a lower bitrate or use the so called 'KVCD templates' http://www.kvcd.net
You can use a bitrate calculator to find out what bitrate you need to set to fit the movie on one cd.
Correction: The first sentence should have read 1 cd not 2.
I want to keep this movie original and intact. I want to play this on my Sony DVD home player for a Sony NTSC TV with intentions to update to a wide screen later. this will be burned in SVCD format.
1) Since this is in PAL format, should I select SVCD PAL or DVD PAL for the file to create?
2) For the source file, I assume the video type is interlaced. Is this correct?
3) What aspect ratio should I select?
4) In the video arrange section, should I select "full screen keep aspect ratio" or".......aspect ratio 2" or ???
I am confused by so many settings that pop up. Are these repetative?
Your source is most likely progressive. TMPG should detect it automatically anyway.
Your input aspect ratio is 4:3 PAL
Select SVCD PAL (Recommend MPEG1 VCD)
Full screen (keep aspect ratio)
Note: There is no point in burning this movie as a SVCD as the resolution is already correct for VCD. Resizing to a higher resolution will not improve the image quality, but will waste more bitrate. You would be better off encoding to MPEG1 VCD and raising the bitrate.
Ei, ive encoded a few anime episodes from avi uncompressed format to vcd and using the saved presets that ive loaded (Video-CD Film 352x240 CBR 24.97 fps) and when i encoded a file that is 180 mb long, it expanded to about 700 mb+ and has no sound? ive tried some avi files and they work just as well, but some came up with a large file size mpeg. how do i correct this problem?
If you Files have No audio then use something like Virtual-Dub to extract the audio to WAV Format and encode the Wav file as the audio source...The Size of a AVI has no effect on how big the Mpeg file will Be..It doesn"t matter if your avi file is 10mb or 10,000GB what matters is How LONG the Source File is..If you used the Standard VCD Template then that is the Standard Compliant Bitrate for VCD and to lower the Bitrate to Lower the File Size will Greatly effect the Quality, you Can Only get a Max of 75-80 Minutes of Video on a 700mn 80Min CD-R, and You can Fit 800MB on a 700mb CD-R...
tnx minion...but how do i make the target file size smaller to fit in a standard cd-r 700 mb? ive encoded some anime and the standard size should be at least 240+ mb in dat file format (vcd). but ive noticed some avi files that exceed 190 mb when i encode it turns out to be 700+ mb in size. how can i correct the source file to fit in a standard cd-r 700 mb? tnx for the reply...=D
Didn"t you read my post ,Most of my whole Point in my last post was explaining that the SIZE of the AVI file in Mega Bytes has to Bearing on How Big the Mpeg file will Be . So saying that you have an AVI file that is 190MB and it turns into a Mpeg file that is Over 700MB and Then asking me how to stop that doesn"t make sence.A 190MB AVI file could be 3 hours long and you would NEVER Fit 3 hours on a Disk, You might be able to get 100 minutes on a Disk but the Quality would be Crap, To do this Like I said you have to Lower the Bitrate, To lower the Bitrate you will have to first load the "Unlock.mfc" Template from the "Extra" Folder, and you will have to set the "System" setting to "Video-CD Non-Standard"...and it does not Matter if your Mpeg file is 700MB or Bigger because you can put 800MB on a 700MB 80 Min CD-R....
errr...i did read it but i forgot to place in my last message that the avi file i was trying to encode has about 23 minutes of anime, meaning one episode,i just cant see y the program when it converts to mpeg, it will expand an one 25 minute episode to about 700 mb for only just one episode? isnt it a bit too much for a vcd file? anyway...tnx minion...
What type of Mpeg??? Meaning a Just a Mpeg file of the Vob Files in the Video-TS Folder?? Or do you want a Mpeg file that You can Burn as a VCD or SVCD??? If you want to make a VCD/SVCD you can use a program called DVD2AVI to Frame serve the Vob files to Tmpgenc and it will extract the audio track of your Choice that you encode in Tmpgenc and in the end you will have a VCD or SVCD..Go to "http://www.dvdrhelp.com" and there will be a bunch of Step By step Tutorials to show you how to do it....
There are many programs which can do this. If the DVD is encrypted then most rippers can rip the VOBS to one VOB then just demux with TMPG.
If the VOBs are already decrypted to your hard drive then you can use VOBedit or VOBsnoopy to convert the VOBs to an M2v file. http://www.digital-digest.com/dvd/downloads/vobedit.html
I tried to use virtualDub to frameserve Tmpgenc Plus. I installed the readAVS to tmpgenc. And follow the instruction of it. But tmpgenc will not read it. Just say "not supported". The readme in readavs said it was tested in win98. I use win2000. Is that the problem?
I tried to use virtualDub to frameserve Tmpgenc Plus. I installed the readAVS to tmpgenc. And follow the instruction of it. But tmpgenc will not read it. Just say "not supported". The readme in readavs said it was tested in win98. I use win2000. Is that the problem?
hello hope someone can help me im trying to convert avi to mpeg using tmpg.but im just getting a black screen with sound,ive encoded a few films and all been ok.i recently restarted my pc and now nothing ive tried raising direct show to 2 but still a black screen can someone please help. THANKS FOR ANY REPLYS
hello hope someone can help me im trying to convert avi to mpeg using tmpg.but im just getting a black screen with sound,ive encoded a few films and all been ok.i recently restarted my pc and now nothing ive tried raising direct show to 2 but still a black screen can someone please help. THANKS FOR ANY REPLYS
If Raiseing the "Direct show" Didn"t help then make sure you have the Proper codec installed and if you do have the proper codec then Maybe re-install the codec and if that doesn"t work see if you can open the avi file in Virtual Dub and if you can then try frame serveing the file with V-Dub and if V-Dub will not open the file then there is something wrong with the File....
DONT INSTALL ANGEL POTION!!!!!! It is a Hacked Codec that can Cause Major Problems when Running Tmpgenc, Your files can end up with a Purple and Pink Picture, and If you have Madia Player 9 then that is Your Problem ,,,,
I agree with Minion don't install that crap on your PC unless you like pink blotches on your movies. Your problem is Mediaplayer 9. Roll back your system with system restore.
I tried to use virtualDub to frameserve Tmpgenc Plus. I installed the readAVS to tmpgenc. And follow the instruction of it. But tmpgenc will not read it. Just say "not supported". The readme in readavs said it was tested in win98. I use win2000. Is that the problem?