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 ]
Hi, hopefully some1 can help me out. I have an avi film which is 29.970fps, which i want to convert into 25fps(pal). Heres what ive tryed and what happens..
Firstly i extract a .wav file from the avi film while its still at 29.970fps(Using virtual dub in full audio process mode). Then using Besweet i set all the standard setting but for FRC Presets i select NTSC2PAL 29.970 to 25.000. I then convert the wav file to an mp2 file. Now this is where the problem starts,
I tested the wav file against the full original avi file, and both sound perfect the speech is at the right pitch. BUT when i listen to the mp2 file the female voices sound like mens voices and the mens voices are very deep pitched. Ive encoded a small part of the film with tmpgenc just to test sync and its fine, its just the voices are very deep.
The link below is the method i use for converting 23.976ftp avi films to (PALsvcd) films and it works every time.
Should i be setting Besweet up differently ?
Please can some1 help.
I couldn"t get through to see the link but the way I do it is ,extract the audio to wav then use "avi frame rate changer" to change the frame rate to 25fps then encode the video to a "m2v" video file, then I find out the video length then use "cool edit" or "sound forge" to stretch the audio to the exact length of the video file,makeing sure that i use the "preserve pitch" option so I don"t get the same effect that you are getting, then encode the stretched wav file to mp2 the mux the files together...
Minion has hit the nail on the head.
The reason you are getting the audio problem is because a 29.97 to 25 fps conversion means the audio has to be stretched quite a lot thus having the effect of lowering the pitch. I don't think Besweet corrects the pitch so this is your problem.
The best way is to use Cooledit pro or 2000 as Minion says. This will preserve the pitch of the audio.
I am experiencing unstable video when playing back via my Pioneer DVD player.
Firstly i edit my footage using Premiere 6. I then create an AVI file using Mainconcepts DV codec. Then using a high quality setting with Tmpge i encode the the AVi into an Mpeg2 and then using Nero burn as an SVCD movie for PAL (UK) .
The quality is excellent if there is no movement but where the video `moves` ie panning the camera the video image begins to stutter and partially break up. But once the camera remain still the image is perfect. This `break up appears as lines. Sorry i cannot describe it any better.
I am using 2 pass VBR rate for quality which is my main aim. I am not bothered about file size. Playback is for a TV not computer.
I have experienced a simular thing on many occations and it comes down to the dvd player that your playing the files on,I have noticed with my player if I put the bitrate to high I get the stuttering effect when there is any movement at all but when there is no movement then it looks perfect, so now for svcd I make sure the max bitrate is below 5000kbs for my player, another thing that causes this is if you encoded your avi file to a different frame rate as the avi file, so if you encoded a ntsc file to a pal file you can get this effect...
This sounds like interlacing artifacts.
You need to de-interlace your source AVI using the de-interlace option in TMPG.
Click the de-interlace filter and choose a scene in the movie with some action like a car moving across the screen or someone walking then choose the'even-odd field(field)' option .
Click the right hand arrow button a few times and have a look at the motion. if it is smooth the field order is correct, if it appears to stutter( move forward then backward) then the field order is wrong and you need to change it.
Once you have selected the correct field order change the de-interlace option to 'even field'.
This should cure your problem.
This sounds like a wrong Field Order in the Video. DV uses Field order B (Bottom Field First).
Don't do any deinterlacing, the only effekt is lowering the Quality of the Video.
I figured out how to do a single avi>>mpeg conversion but I would like to know how I can set a batch task to convert a whole bunch of avis to mpegs at the same time.
What you do is load each file in tmpgenc seperately do your settings and go to "file" to "save progect" then repeat for each file, then when you are done saveing progect files for each avi file then you go to "file" to "batch encode" then load all the project files into the "batch encode" window then click "run" ans thats it accept wait.....
How can I eliminate a stutter condition on the last frame of a clip?
using TMPGEnc version 2.56.39.143 downloaded June 30, 2002
Have a 5 min 13sec clip that I've used to test various settings and encoders. This clip is NTSC, 29.97fps, 720x480 resolution, 4:3 aspect.
The last 1 second of this clip contains rapid motion. Each movie consists of the same 5:13 clip, authored with DVDit and played back on a Sony DVP-NS700P player. Outlined below are my test movies which all use this footage:
Movie1
Clip exported from Sony Giga Pocket as MPEG2
plays fine
Movie 2
Clip exported from Giga Pocket as AVI, encoded with DVgate Assemble
plays fine
Movie 3
Clip exported as AVI,encoded with TMPGEnc (using Rui's settings)
plays fine until last frame. Last frame stutters on screen for approx. 1 second before the movie ends.
For example, when playing back, I pause the player at the 5:13 mark. I then frame by frame forward 8 times. For all movies tests, the clip plays ok up to this point. When I frame forward the 9th time, Movie 1 and Movie 2 displays ok. When I frame forward the 10th time, the frame stutters. If I continue to frame forward, on the 21st time, the movie ends.
Movie 4
Using Premiere 6, I took the same footage and added a 20 frame, black fill title clip at the end. Encoded with TMPGEnc. Performing the same frame by frame replay, when I frame forward the 9th time I see the final scene appear ok. I then have to frame forward to the 41st time until the movie ends.
Using BitRate Viewer, I have verified that each clip contains a total of 9399 frames, except for Movie 4 which contains 9419 frames.
I have tested with 8M CBR, 2-pass VBR (hi 8000, avg 6000, low 2000). Same results.
Tested with 8M CBR, Max GOP length 15. Same results
Tested with 8M CBR, unchecked "Detect scene change". Same results.
Any suggestions on how I can eliminate this "stutter" condition?
I have also used other clips created by Premiere 6. All appear to have this same condition.
I thought I had cracked the video audio sync issues with TMPGEnc (even v 2.56 does this). Then I found that even the raw captured AVI on the hard drive was suffering badly from the same problem.
I had a 3 hour video which I captured thru a win tv card using MGI Videowave 4 capture software.
At the beginning of the capture all is well but by half way thru (about 1.5 hours) there is about a 10 second difference.
If this is happening at the AVI capture stage what hope do I have for encoding the whole thing? Why is this happening?
This is probably the most common problem with captured files, the de-sync is usually caused by dropped frames and in turn the dropped frames are causeing the video file to be gradually shorter than the audio file so your file gets slowly but gradually out of sync, to fix this you can use a audio editing program to speed up the audio file and shrink it so it is the same exact length as your video file you can use "cool edit" or "sound forge" for this ,There is a detailed "how to" at "www.vcdhelp.com" on how to fix sync issues like this....
Your capture software may have a setting somewhere regarding interleaving with the Video with the audio. If it has make sure it is set for full interleaving this should prevent the audio going out of sync when frames are dropped.
The other and more likely reason is that your sound card's clock frequency is running at a different rate than your capture card. This will cause precisely the problem you are having.
Your best bet is to use another capture software which corrects this. Virtualdub is able to syncronize your capture card with your soundcard.
Start Virtualdub for capture and then click Capture>Timing... and make sure the option to sycronize is checked.
This should sort you problem if it's down to the frequency mismatch.
Your capture software may have a setting somewhere regarding interleaving with the Video with the audio. If it has make sure it is set for full interleaving this should prevent the audio going out of sync when frames are dropped.
The other and more likely reason is that your sound card's clock frequency is running at a different rate than your capture card. This will cause precisely the problem you are having.
Your best bet is to use another capture software which corrects this. Virtualdub is able to syncronize your capture card with your soundcard.
Start Virtualdub for capture and then click Capture>Timing... and make sure the option to sycronize is checked.
This should sort your problem if it's down to the frequency mismatch.
Hello!Although you cannot belive I've another problem with audio when I encode a film!As you teach me I'm using virtual-dub to encode the audio and then I encode all film (audio + picture) And it worked well in my first film!but in the second one, when I encoded the second frame of film audio isn't syncronized with picture. What can I do to solve this problem?????????????
Another personal question: tmpgenc plus have so many problems? or if I buy it I will encode all my films correctly at once?
It seems that most problems encodeing stem from the quality and condition of the avi file you are encodeing and not bugs in tmpgenc,if you encode movies downloaded off the net you will have a lot more problems than with avi files produced by your computer or captureing, this is because while downloading movies off the net there is a certain amount of corruption in the file,but for you out of sync problems, it depends on what type of sync problem you have as to what method you use to fix it, if the audio and video are out of sync the same amount the whole way through the movie then the problem can usually be fixed by de-multiplexing and re-multiplexing with a program like "bbmpeg" or "mpeg2vcr" and use the feature in the multiplexor that lets you off-set the audio or video to sync up the file......
I wish there could be an option to Pan and Scan a movie in a way that I'd tell TMPGEnc the periods and the amount of Pan and Scan to do for every section.
Its very helpfull when converting from one format to another NTSC to PAL, 16:9 to 4:3 and such.
I've been trying to make an SVCD compliant MPEG2 file using FlaskMPG 0.594, Avisynth 0.3 (w/Premiere Plugin), and TMPGEnc 2.55. However, TMPGEnc keeps fooling me:
*If I select CBR, everything works, but the mpg file gets huge. For instance, I encoded a 50min video at a specified average bitrate of 1920kbps (and a max of 2300kbps), which should produce an mpg file of roughly 700..750MB. The file I got eventually was about 1GB!
*If I select 2-Pass VBR, the behavior is even more curious. The resulting mpg file was about the size I'd expect, but it displayed only the last frame throughout the entire playing time, while audio was okay.
*Then I tried to encode just a tiny part of the video. I set Flask to convert the first 5000 frames. Result: I got the audio information out of frame 0..4999, but the video of frame 5000..9999. It looks as if TMPGEnc rewinds the audio stream after Pass 1, but not the video stream.
Is this a bug, or what am I doing wrong?
In most cases if the file plays fine on your computer then it should burn to a cd-r and play in your dvd player if your player supports svcd, so All I can think of is if the mpeg plays fine on your computer the problem must be with the way you burned it to cd-r or a problem with your dvd/svcd player, There are a few good svcd authoring program out there that if used correctly should author your file to play , "Nero Burning Rom" is probably the easiest and most reliable for burning vcd/svcd, and "VCD Easy" is good if you want to do chapters, if you burn it correctly then the only place left that the problem could be is possibly your dvd/svcd player.....
Ohh, and if you didn"t check the mpeg file after encodeing and before burning to disk,I don"t know why you wouldn"t but in cas you didn"t the problem could be that you need to raise the priority of the "Direct Show File Reader" in your "Vfapi Plugins".Usually if the movie is displayed in the tmpgenc screen while encodeing then it will be there in the file but you should allways watch the mpeg on yer computer before burning cuz sometime the file can have problems like your problem or no audio or out of sync or any other one of a dozzen problems...
Go to "www.vcdhelp.com" and there there are a lot of step by step manuals that will help give you and idea on how to go about it and there are links to the software that you will need to do it correctly, makeing vcd/svcd/dvd"s can be a lot harder than it looks so it is good to read up on all you can and get a good idea of the principles involved before you start encodeing......
Hi,
I am new to the world of video editing, I am currently trying out
the TMPGE, I follow some of the template I found to get the best
DVD quality from AVI (VBR, 10bit, high bit-rate, Best quality, 2 pass...)
My 2-hr video need 50 hr of encoding time,
is this normal and can I do anything to speed it up. My system is:
- AMD Athlon 1.2 with 512 Mb RAM
- Windows XP on Seagate 9 Gb, 7200 ATA 66
- Captured AVI is on a separate disk Seagate 40GB, 7200 ATA 100
First thing you can do to increase speed and maybe even quality is to not use the 2 pass encode mode, most poeple that are farmilliar with tmpgenc know that the CQ method will give you as good results with less than half the encode time, and useing the "high quality" mode instead of "highest quality" will speed things up a lot, and try not to use very many filters, and make sure that the cpu optimizations are enabled (3D-now,sse,sse-2), if you are doing everything correctly then it shouldn"t take more than 10-12 hours to encode a 2 hour avi file on your system...
Yes, me too (as read in yesterday's thread). I've tried everything and still can't get less than 69 hours on a 2:24 video with a P3-800 & 256mb (480x480 SVCD). It's particularly frustrating because so many others are able to easily achieve what I consider to be stunning performance. If I could get it to 24 hrs even, I'd be elated. But unfortunately, 69 hours makes me less than eager about the entire process.
Currently, I'm in the process of defragmenting my boot HD with Norton Speed Disk. I think the disk is fragmented just enough that the windows swap file is having issues. It's only a theory but I'm running out of ideas. It can't hurt at least. If I find out anything, I'll be sure to post it here.
My test encode using VBR (hi 8000 avg 6000 low 2000) took approx. 25 minutes for each minute of footage on a WinXP P4 2G, 512M system. Windows Task Manger showed TMPGEnc consuming 95-100% CPU. Disk light blinked on and off slowly. It appears that the limiting resource is CPU
So based on your results, a movie the size of mine (2:25) would have taken you approximately 19 hours and 20 mins. With a processor about 3.7 times as fast as mine, that makes my 69 hrs seem accurate. So why is everyone else with relatively low speed processors reporting such fast encode times? Hmmmm...
Now wait a minute here. I mis-calculated. At 25 minutes encode time per 1 minute of video, that comes to 60 hours and 25 minues for a movie 2:25. If that's with a 2G P4, and mine takes 72 hous with a P3-800, I would say that CPU isn't quite as big a factor as I once thought. Clearly, the P4-2G performs better, but I would have expected a much larger improvement timewise. But my original question still applies, "How do some of you get such low times on your encodes?". Lower res? Maybe an older version of TMPG? No wonder there's nothing I can do to speed things up!
I agree entirely. Like I earlier said, the difference between encoding times on my home machine: PIII 450MHz 314MB RAM 2 IDE hard disks, and my work machine: Athon 1.4 GHz, 1.5GB DDR RAM, 3 SCSI hard disks is only around 70-90%. i.e. ~a factor of 2, but my work machine should encode at least 3 time faster.
dpweber,s test movie took 25 minutes to encode 1 minute of video, i.e. thats a factor of 25:1 so a 2 hour movie on his machine will take 50 hours.
On my machine a 3 hour movie took 2 days - in fact it took 2x16 hours = 30 hours, at absolute top quality.
And the version I'm using - beta12a (which in certain quarters is still said to be the best version).
I have performing a number of tests over the past week to obtain performance and quality info. I've included below some tests that I recently completed with TPMGEnc on my Sony P4 2G, 512M, 2-100G drives. All test files are 5:13 in length.
(the "AVI file created in Step 3" refers to my use of Sonys' GigaPocket capture hardware/software. If you're not familar with this, it's functionality somewhat like a Tivo unit. Once the video is capture, it supports export of AVI and MPEG2 files)
Step 6
Processed AVI file created in Step 3 using free version of TMPGEnc. Used default settings with 8m CBR specified.
length 14:50 (just a little longer than Movie Shaker)
cpu 70-95%
disk light on then off slowly
file size 319M
Step 7
Processed AVI file created in Step 3 using free version of TMPGEnc. Used default settings with 8m CBR specified.
Created elemental output files
length 14:30
cpu 85-100%
available memory 320k
disk light on then off
file sizes: video 299M, audio 14.3M
Step 8
Processed AVI file created in Step 3 using DVgate Assemble to produce 8m CBR
length 5:30
cpu 85-95%
available memory 345k
disk light blinking
file size 299M
Step 9
Processed AVI file created in Step 3 using free version of TMPGEnc. Used settings with 2 pass VBR specified. Max bitrate 8000, min 2000, avg 5000, motion precision set to "Motion estimate search (fast)"
length 27:41
cpu 90-100%
disk light on then off slowly
file size 204M
This final test results in a relatively quick encode; however, the image was only fair and pixelization occurred during pans. The image quality improved significantly when I switch to motion precision Highest. This also resulted in the best quality of all others listed above. However, this difference is quality was not overly significant and I was happy with the results of all tests except for the Step 9 test.
I am currently attempted to resolve another problem where the last frame stutters (you can find the details in another post on this bbs). Once I get that resolved I will conduct the same tests with the CQ setting.
I think there must be something wrong with your guys systems ,Wow 25 minutes to encode 1 minute of mpeg on a p4 with 1.5gb-ram,definately something wrong you should be encodeing in real time or faster, on my 800mhz it is about 4 or 5 to 1 with tmpgenc,I get as fast performence on my p1-133,it takes about 1 hour to encode 3 minutes on my 133.....
I agree with Minion their is surely some thing wrong with your guys systems. 25min encode time for 1min of video is insane. I average 3-5 Hours dependeing on bitrates I use for 1 Hour of Video in Mpeg format for conversion and alot less for Divx 4.12 encoding.
My system comprises of AMD Athlon 1.33Ghz,256MB 266MhzDDR ,15&40GB 7200Rpm Maxtors, 4xAGP32MB ATI Rage Furry Pro On a MSI KT266Pro Mother Board.
Could the reason for the fast encode times be due to your video cards? I've been out of the video card hardware advances for quite a while but if you have hardware MPEG2 encoding/decoding, could it be that your system is utilizing it? I realize how stupid this must sound, but there has to be a logical explaination for this huge difference in encode times. I know my system runs slow for TMPEG, but I can compile a complex application just fine. I don't think it's as much of what's "wrong" with our systems as it is what's "right" about yours. Now....if we can only identify what that is, we'll all be in business. Well, the slow ones will be anyway. :)
None of us is using hardware encoding and the Video card has absolutely nothing at all to do with it.
I'll give you some specs that my system is using with TMPG.
Hardrive- IBM deskstar 40gb ATA 100 (using 80 pin cable a must) DMA enabled
Graphics- Nvidia TNT2
Processor- P4 1.7
Ram- 256 DDR 266
**************************************************
TMPG settings. Under Option>enviroment setting.
GENERAL:
All unchecked except,
Canapus DVD codec- 'enable reading reffered AVI through network'
Panasonic IO Data DV GigaAVI CODEC- 'enable reading reffered AVI through network'
Temporary file- same drive as operating system.
CPU:
All unchecked except MMX, MMX2, SSE, SSE2
SOUND:
Non checked.
EXTERNAL TOOL:
Non checked
VFAPI PLUGIN:
ALL checked and at priority 0
***********************************************
Under Option>preview option.
Display with thinning is checked.
Fix to 320x240 is checked.
Automatically resize window for preview is checked.
*************************************************
Under Option>Task priority.
When active - Normal priority is checked.
When not active - Idle time only is checked
**************************************************
When encoding all filters are unchecked and motion search precision is set to Normal.
Floating point DCT is checked.
**********************************************
Thanks Ashy. We have some different options with the new software that I couldn't set, but overall, your list was helpful. I did some testing with a :30 promo and it looks like I'll be getting 9 minutes per 1 min of vid. Not what I had hoped for but an improvement just the same.