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 ]
What do you mean real time? TMPG multiplexes/de-multiplexes faster that real time.
Also I don't think it can handle transport streams. You would need something like MPEG2VCR for this.
Do you mean de-multiplexing or transcoding. To de-multiplex is to seperate sound and picture streams. To transcode is to convert from one video format (type) to another format (type), or to downsample (or upsample) from the same type. If you're tanscoding, in the case of mpeg2's, it can't be done in real time, unless you own a powerful dual processor machine, and even then it's unlikely.
Griff it is possible to transcode to MPEG2 at a rate faster than realtime using my 1.7ghz p4 and 256 RAM, so nothing special there.
I regularly encode to MPEG2 at 1.5x real time using DVD2AVI>AVIsynth>CCE.
CCE is much faster than TMPG and the MPEG2 output quality is unmatched by any software encoder and beats TMPG hands down.
I thought we were discussing TMPGEnc's encoding speed, not CCE.
I've been following the interest in CCE for some time. It seemed to gain prominence around last Christmas. I was thinking of giving CCE at try, 'till I baught a piece of hardware which stopped me in my tracks. More on that later.
The thing that gets me is that many of the people who make comparisons between TMPGEnc and CCE, don't even have a DVD burner, so they never get to see what both packages can really do; and those that do have a burner then watch the DVDs on a playstation or similar.
Also,simply looking at a DVD on as computer monitor will only give some idea of what the final movie will look like on tv, but may also be misleading, since there are differences in brightness, contrast, and black level between the two.
I recently got hold of the Win Nova-t digital terrestial tv card. The mpeg2 files that this produces are breathtaking in their quality (given the right source) and I was determined to keep as much of this quality as possible, when transcoding to DVD compliant mpeg2. What I was totally unprepaired for, was the image quality I got when playing these off-air recording on my hardware player. The image quality is absolutely stunning; and I don't say this lightly. I'm a total perfectionist when it comes to video, (and audio) playback. I have a high-end DVD player (Pioneer DV717), a state-of-the-art flatscreen tv, and access to an Eizo video projector.
Some of the recordings I'm transfering to DVD are superior to many off-the-shelf titles (although I've read that some of the movie companies are releasing special editions of their movies with no extras and higher bitrates).
As for CCE, how do you better perfection?
However, you don't have to take my word for it. If you give a postal address I'm willing to send you a copy of a recent concert - if you pay for postage and the DVD, about GBP2.00 - recorded from Channel 4.
I was told that, with this program, I could convert downloaded .avi files of movies to a format that I could use to burn onto CD-Rs using my Nero software. I loaded the 2.57 program and used the Load Video CD (NTSC).mcf to set up a template. When I click Browse next to the Video Source window and then go to the location where I have a couple of movies I got from Kazaa, I click on one. I get a message that gives the files name and also says, "can not open, or unsupported". What does this mean and what do I do?
Hi,
When encode a 2-hr long AVI, the audio is gone after
about 1hr 22 mins, and came back toward the end of the
2-hr. This happens with the both versions, the old and
the newly released 2.57. They both drop audio at the exact
spot. Anybody seen this before? 2.57 supposed to fix the
audio drop for 3hr long footage. I am using the trial versions
for now, until I am satisfy with the result.
Have you tried to use the audio encoder plugins that you install in the "external tools" they probably would solve the problem cuz they are useing a totally different encodeing engine than tmpgenc, The two good ones are "toolame" and "Scmpx" just do a search for them on Google and you will find them for download they give much better quality and reliability than the encoder used in tmpgenc, You install them in tmpgenc buy going to "options" to "enviromental settings" to "external tools" then just click browse and choose the encoders .exe file and it encodes the audio first before the movie instead of as the movie is playing...
Loosing audio happens even with much shorter files: I had the same problem with a 7 minutes video where audio disappeared after 2 minutes in the *.mpg file
Using an external audio encoder (e.g. toolame) did not help either.
A remedy I was successful with:
Extract the audio as *.wav from the original *.avi with VirtualDub (Menu: File/Save WAV....), and re-encode with TMPEGEnc using the original *.avi as video source and the extracted *.wav as audio source.
Notice that video/audio synchronization may eventually require some extra efforts.
Thing is, the resulting MPEG-1 files really variated between the two...
For start with Ligos I noticed that there is an actuall shift in colors, Orange became more red and Blue became more cyan! Further more there was also an noticable extra brightness that could especially be seen in the darker areas...
The VFAPI plugin didn't have the above odd problems, but it did have few disadvantages. For start it didn't let me use SSE-2 despide haveing a Pentium-4 computer and Putting in it's directory the P4Package.dll file. Further more, unlike ligos, it didn't let me import the MPEG-2 system (program?) stream so I had to demux the original mpeg-2 file and add the extracted mp2 file as audio source.
In terms of speed the two are te same except when useing Floating point at the VFAPI plugin's setup which made it almost 8 times slower(!!!), but maybe if I could enable the SSE-2 it wouldn't be that slower.
Note: I meant to say that the VFAPI plug in, when selecting the MPEG-2 program stream as video source, wouldn't allow using the audio part of the file (unlike ligos). So I had to demux the audio to mp2 file and add it to the audio source.
Go back to version 2.56 or earlier and try this codec, http://www.plmovies.com/mpeg2codec_installer.zip as it doesn't seem to work with the new 2.57 version. It's the fastest one I have used and gives good results or you could use DVD2AVI and create a d2v file.
As for importing the audio from an MPEG2 file, why bother? The audio is already in the correct format, so why encode it again. All you need do is multiplex it with your re-encoded video file, saving time and quality.
Will give the codec a try.
If not, I guess I'll try DVD2AVI... Used it only twice until now. For the avisynth/decomb and for a very complex VHS restoring process that also involved avisynth. But haven't chacked for any brightness/color problems...
And about avisynth, I tried using decomb and the video did come out progressive. However when playing the video was jumpy as if it dropped the wrong frames... Tried changing parameters but didn't help.
I can't understand why the codec didn't work unless you are usind TMPG 2.57 as I know it works.
Anyway as far as using the decomb filter in AVIsynth. I had more or less the same problem, but managed to reduce the jerking by using the decimate function to decimate every 12th frame.
Tried 12 and still was quite jumpy... Depends on the video itself too I suppouse...
And to note, that MPEG-2 plugin for which you sent me a link, not only that it didn't work but it also messed up the directshow/ligos so I can't play mp2 files... Actually they do play but change the playing speed of the files randomly while playing... Very odd sounding... Maybe Direct-x has something to do with it, or it doesnt work with win2k? who knows...
Looks like you are having more or less the same probs as me with that movie, I don't know how many hours I tried to get a smooth movie and then just settled for the best I could get it. It's funny how there are plenty of utilities and filters out there for removing 3:2 pulldown, but I have never come across anything specifically made for removing 24:1 pulldown nor have I came across any software which can actually apply 24:1 pulldown.
Maybe the problem you are having with the MPEG2 codec is due to the other MPEG2 codecs you have installed which I would guess are causing a conflict as they both use direct show. I know they work on any system as there are many other on this BBS who use them and never have a problem with them.
Another thing I forgot to mention. The MPEG2 filters I gave you the link to are in fact LIGOS filters. So it looks as if you have two sets of more or less the same filters installed which is casuing the problem.
Apparently TMPGEnc ver 2.57 had a bug fix included that would let audio for movies longer than 3 hours to encode. However, I am not having much success with this. If I convert the audio to an mp3 file (filesize ~400 mb), the audio shows up in TMPGEnc. Same audio in a wav file (filesize ~2.7 GB), no audio displays.
There is a fairly simple work around for this, you can just encode the video then encode the audio then mux them together, but if the audio still cut"s out in tmpgenc after a certain time you can use a audio encoder like "db power amp" with the mp2 codec to encode the audio then mux the audio and video together........
I read about that workaround in one of your previous posts but when I try to encode audio, TMPGenc takes about a second and generates a mp2 file of 2 KB size. Is there something I am missing?
There might be something wrong with the audio file, so try the "db power amp" work around the program is free and it is one of the better audio encoders, you can download it here:http://www.dbpoweramp.com/
but don"t forget to download the mp2 audio codec...
I do appreciate the help you have provided and just wanted to make sure that I do not sound argumentative. This is the second wav that I have had the problem with. Exact same symptoms. Leads me to believe that there has to be an issue with TMPGEnc. But, hey, I have not paid for TMPGEnc, yet :) so I will try what you suggested.
Thanks, again. Still any other pointers are appreciated.
When you say you have tried encoding the audio separately with TMPG and it only results in a 2kb file, well I think this is the problem.
It seems as though the same think is happening when you try to encode the movie with the sound. It seems TMPG is encoding the movie ok, but is only encoding 2kb of audio with it.
I guess if you examine the movie with the audio you will probably find a 2kb audio file.
I would also guess that if you can sort out the problem of why TMPG only encodes a 2kb seperate audio file then I think your problem will be solved.
Have you tried any earlier versions, try going back a few versions because you don't know when this anomalie started.
Have a look for version 2.52 on google, seemed quite stable. I can't give you the direct link 'cos it will get deleted, but if you search for
tmpgenc-2.52.34.129.zip on google you may find something.
When setting the audio gap in 2.57, it seems that if I adjust the gap just a little (100ms for example) it might skip the audio one whole GOP. The best adjustment I seem to be able to make is 500ms (one GOP at 15 per second)...
I have made a VCD with an 16/98 avi film . I encode it in MPEG VCD format ( using 16/9 option ). When I play it on my TV ( 4/3 ) the image is not complete ( left and right parts are missing ) but when I play the original MPEG ( not burn ) with a window player the image is fine ....
What do you think about it ? Do you think it come from my TV format ( not 16/9 format but 4/3 ) ? Thanks for your help :p
If you are watching the vcd on a regular 4:3 TV then you have to set the output to "4:3" with the video arange setting to "fullscreen keep aspect ratio" this will give you the proper aspect for your TV...
What is the VBV setting I should use to accomodate the "112 x 2048" specification that Alcorn McBride gave me for this parameter? Originally I tried a setting of 230 KB, but I got an error message "The VBV buffer size is too large for the selected level".
Thanks in advance for your help.
When encoding a film, playback from a DVD R results in studdering motion. The video has been encoded as both a and b field. The source is film, telecined to digibeta, digitized into an Avid Meriden (b field) Output to quicktime using the Avid codec.
The best results seem to be using inverse 3:2 pulldown on playback, b field, but then when audio is added, the video breaks up. Different frame rate?
How can I get smooth motion from film originated footage? removing telecine won't work because the footage is edited and the pull in frame will change from shot to shot.
Are there any tricks to improving encoding speed when using TMPGE? Any preferred environmental settings or Windows OS tricks to improve encoding speed?
Thanks
Haveing a really fast processor helps,I encode vcd at real time on a 1.7ghz, but if you set the "motion search precition" to "motion estimate search" this can double the speed of encodeing from high quality but the quality won"t be as good......
Thanks for you prompt reply. I have a 2.2 GHz P4 and 1 Gbyte of RDRAM using Windows XP. I brought up Windows XP "Task Manager" while encoding on TMPGE and noted that only 200 Mbytes of Cache was being used with ~ 800 free for use. Either TMPGE or the Windows XP OS is limiting the amount of RAM being used during the encoding process. Any was to increase usage of Cache memory rather than accessing hard drive so much?
I use Windows XP also with the default cache settings accept I use "Norton System Works" to optimize and manage my cache and I think it seems to help cuz my encodeing speeds are fairly good and you have a faster cpu than me so you should be able to encode vcd at close to real time also(high resolutions take a little longer)With CinemaCraft I can encode at twice real time..So maybe it is Norton that is helping with my cache settings so go and download a copy it might help...
Thanks for the information. I decided to run a test using TMPGEnc version 2.57 on two different platforms (Athlon 1.2 GHz with 768 Mb of SDRAM and a P4 2.26 GHz with 1 GB of RDRAM). Using the same TMPGEnc settings and encoding the same file ( 3.3 GB AVI File), the results weren't all that dissimilar. The P4 system system came in first with a total encoding time of ~ 1 hour 10 minutes while the Athlon system completed in 1 hour 39 minutes. What do you think?
I tried an experiment that had interesting results. I installed "RAM Disk" software that sets up a portion of System RAM to act just like a hard drive. I installed the TMPGenc application and the AVI file to be encoded on to "RAM Drive". I then started TMPGenc, selected the AVI file to be encoded from "RAM Drive" and stored the outputed MPEG2 rendered file on "RAM Drive". There was virtually no encoding time difference between running the application and it's data completely from RAM versus running application and it's data from hard drive. The TMPGenc encoding process is clearly CPU dependent (atleast when encoding high quality MPEG 2 files.
If you use TMPGEnc Plus, and enable "Saves analyzing result of multi pass VBR to cache", and specify the RAM disk at "Folder setting" in Environmental setting, it may be faster comparing to HDD. But you need a lot of space for cache if you encode large file.
I agree that encoding is mostly CPU dependent. I'm not suprised you got the results you did.
It would seem pretty obvious that there is no way that even the fastest CPU could encode fast enough to cause a hard drive lag even when reading and writing to the same drive. Look at it this way my hard drive can sustain a transfer rate of 40 MB/sec.
If a processor could match this rate with processing speed it would mean a 2hr movie would be completed in roughly 40-60 secs. Highly unlikely wouldn't you say. I know there are other factors to take into account such as temp files, cache, virtual memory and reading and writing to the same drive which all slow down the hard drive and the process, but still no processor is going to match it for speed thus causing the hard drive to struggle supplying the data fast enough.
The bottle neck IS the processor, but a reasonable amount of RAM would almost certainly help.
Hello,
I am trying to encode AVI capture file into NTSC Film VCD. The file info: size ~35GB, AVI 2.0, codec Huffyv. After loading it into TMPGenc (not so fast but OK) and setting filters (Noise reduction, Inverce telecine, Sharp, Crop, Resize) it goes to encode. First it process Inverse telecine (slow, but working), after TMPGenc goes to actualy encode file and strage things starts - it working for ~30-40 seconds using 100% CPU and after it stops (0-2% CPU) and becomes "Not Responding", after some time it wakeups and load CPU with work again until next "sleep" session. The remaining time become very long and very bad thing it constantly increasing and does not go to end. So anybody knows where is problem and what the solution? The hard drive is fast, memory is enough - so what is wrong?
I can't tell you what the problem is, but may have a work around.
Use Virtualdub to frameserve the file to TMPG. You could also use Virtualdubs filters on the file, (Noise reduction, Inverce telecine, Sharp, Crop, Resize) as I think they are faster than TMPG's filters.
When i convert a DivX codeced movie into vcd , there seems to be this flickering of white lines in the output video. It looks like white lightning or something, and sometimes its even in the margin (if on widescreen. this problem started when i downlaoded the new free version of tmpgenc.. and i re-installed the old version and it now its doing the same thing..do you guys know anything about this.. thanks