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 ]
With "Real Media" Files you have to convert them to avi first before Tmpgenc will encode them, there is a program called "Tinra" that will encode RM to avi,I think it is a freeware program and you should be able to find it on google search engine...
I have been reading a bit about this format and from what i have read you have to use a program called "ATI Multimedia Center" and it can encode the "ATI VCR" Format into mpeg2, but what would probably give you better quality is to use a different codec to capture too, then encode with tmpgenc, maybe a codec like "huffyuv" if you have a big hard drive...
audio sync problems.i have tryed to back up my dvds useing flask then tmpg to mpeg the file but the audio sync to the video is out by say four seconds at the start to say 15 seconds at the end why is this ?can you help??thanks
I have recently encoded the lord of the rings from avi to mpeg1, and the file played fine, with no audio sync problems. However, the file was too big to fit on one cd, so i cut it into 3 seperate files. But on each new file (1of3, 2of3, 3of3)after about ten minutes the video will go all messed up for about ten seconds, and after that, the audio jumps out of sync by about one or two seconds for the rest of that file. This is the first time i've ever run into a problem like this while making vcd's. The frame rate is at 29.97fps. Does anyone know what may be causing this problem, and if so how i could fix it. Your help would be much appreciated. (By the way, the audio is in sync for the first ten minutes of each file, it's not until the video goes all messed up that it suddenly losses sync)
I'm afraid there is usually only one way to circumvent this problem and that is to use some other software.
MPEG2VCR will allow you to cut the file, but only the full version. A free alternative is to use BBMPEG. First you need to demultiplex the MPEG with TMPG and then use BBMPEG to multiplex the file.
When doing so BBMPEG has an option to split the file at certain points either by file size or time.
This should solve your problem.
I just started my first SVCD encode of a 2.5 hour source and TMPEGEnc tells me it's going to take 72 hours. Is this normal or does this indicate something's wrong? It's been 19 hours so far and only 28 minutes has been completed.
My system is a Micron Clientpro PIII-800, 256k, Win2K.
Settings I used:
DVD2AVI .d2v file as source video (Took 35 mins to capture the 6Gb)
SuperVideoCD (NTSC).mcf template
Motion Search precision = Highest Quality (very slow)
Encode Mode = Interlace
Rate Control Mode = 2 Pass VBR
Avg Bitrate = 1842
Max Bitrate = 2516
Min Bitrate = 1145
Max Pass = 2 Pass (old type)
Video Source type = progressive
Field Order = Bottom field first
Source aspect ratio = Full Screen (keep aspect ratio)
If there's something I can do to make this go faster, I'd certainly appreciate any info. Thanks for any help offered.
If this is your first SVCD encode why are you using 2 pass and the highest quality setting. It's no wonder it's taking so long, this is just overkill.
There are two things you can do which will dramatically reduce the encoding time and give you the same quality or maybe even better.
Firstly the highest quality setting will not give you any noticible improvement if any than the high quality setting, this in it's self will halve the encoding time.
Next forget about the 2 pass. It is now widely accepted by most people on this board that the 'Constant quality (CQ)' VBR method will in most cases give better results vs file size than the 2 pass method.
following these steps should reduce your encoding time considerably.
Thanks for the reply Ashy. I was pretty excited to try your suggestions and they both make perfect sense to me. After the first disk finished,I thought the quality was terrific, but it took 24 hours to make. So I stoped the process and restarted it using high quality and CQ. Incredibly, this only took the job down from 72 hours to 69. 3 hours was all I could gain so I'm just as concerned now as I was before. Doesn't it seem like 72/69 hours is overly excessive? I'm using the process outlined on afterdawn.com at:
Thanks Griff. I'm pretty sure that's exactly what I'm doing. I created a separate wave file using DVD2AVI. Inside TMPEGEnc, I use SSRC as the sampling converter and tooLame as the audio encoder. I can't imagine what could be causing it to process so slowly. It's very frustrating because it seems so straight forward. 30 hours is closer to what I had expected. I am puzzled by your mention of the CQ settings ramped right up to 100%, the max bitrate set to 8MB. I'm not aware of those options. I just used TMPEGEnc pretty much out of the box, so to speak.
It seems to me that some other program or programs, other than tmpg, is using processor power. What programs are running in the background? How much disk space is available? What size is your page file? How much free Physical memory is available? Just a thought!!!
Nothing else is running and the hard disk is a brand new 80gb. Page file is auto (windows managed) and the memory is 256mb (152mb free). Another thing I found interesting is that I can launch multiple IE sessions and surf the web pretty much at will. When I do, the estimated time remaining increases slightly but not significantly. It added about 45 minutes after 4 hours of web use. I started it again last night at 9pm with nothing open and went to bed shortly after. When I awoke this morning, 8.5 hours had elapsed and the 1st disk (of 3) was only 29% complete. The timeframe was still on 00:00:00 so the video encoding hadn't even begun yet. If there is a setting that will enable 100% CPU utilization, then maybe that's the problem.
I apologize for wasting so much BBS space on this, but I was hoping it was something more obvious. I just don't know what else I can do.....
I'm amazed that you say the encoding time only went down to 69 hrs from 72 hrs when using CQ. If anything this should have halved the encoding time straight away as you are only doing 1 pass, are you sure you set the motion precision to high and not highest. You could try motion estimate as this sometimes gives similar results but faster.
As a note I used to have a P3 450 which would take 12-15 hours for a 2hr MPEG2 movie when using the CQ method and precision set at normal which would probably equal to about 25 hrs using the high quality setting.
I have pretty much the same system as you "800mhz p3 512mb-ram" and a 2 hour film only takes about 12 hours , maybe you haven"t got the system optimizations turned on" go to "options" to "enviromental settings" to "cpu" and make sure the "mmx,mmx-2,sse" boxes are checked.....
Wow! You have double the memory which would help some, but that's a huge difference. Yes, all those things are checked off in the CPU tab. In fact, it checks them automatically so I'm sure it does some hardware checking. I'm going to have to really go after things tonight and try to figure it out. I'm using SmartRipper to capture the VOB's, and then DVD2AVI to grab the audio and create a video project. But those things shouldn't contribute to the prolonged encoding. I'll report what I find, assuming I make some progress. :)
This is an interesting problem. All things being equal, more processing power should equal faster encoding time. Having said that, my work machine (1.4 GHz Athlon, 1.5 GB RAM, all SCSI IO devices) designed to do some serious number crunching, doesn't turn in the blistering performance with TMPGenc that I thought it would - it's about 70% faster than my home machine (yet on paper it should be nearer 3 x faster).
"hkcmicro" might be getting close. Since you're running Windows 2k, open the task manager (CTRL+ALT+DELETE) click on processes, and then start TMPGEnc encoding the source vid file.
In the task manager look for TMPGEenc under "image name"
The processor utilisation should be around 90-100%. If not, look for other processes using significant amounts of CPU time.
If TMPGEnc is using 90-100% CPU time, then the problem most is probably your source vid. If, on the other hand, TMPGEnc is using only around 30% CPU utilisation, then find the other application(s) that are using the CPU time and close them down.
This may be a silly question, but do you have DMA checked for your hardrives.
Also there was a post a while back from somebody with more or less the same problem. The reason turned out to be that he was using different drives for the temp file and output file.
As soon as he changed everthing back to his main drive the encoding time was reduced and went back to normal.
Ashy as highlighted a very good point,also if both of your hard drives share the same IDE channel this will considerably slow things down due to the fact that drives cannot read and write at the same time. Try what Ashy suggested, and put everything on the same drive. GOOD LUCK.
>Ashy as highlighted a very good point,also if both of your hard drives share the same IDE channel this will considerably slow things down due to the fact that drives cannot read and write at the same time.
Yes, but no-one's going to do that. If you're going to have two IDE hard drives, you'll give your system and application disc (master) on IDE1 (no slaves) and put the second drive as master on IDE2, with the CD/DVD/CD-R/DVD-R as slave.
This is how my system is at home.
Even if you did put two hard drives on the same IDE, the performance hit would be nothing compared to having to wait for the drive heads to move from read sector, to write sector and back again, in a single drive copy. I've copied some very large files on a SCSI system, both to the same disk and to another drive and there is no comparison between the waiting times. There is nothing more frustrationg than having to listen to a hard disk churn away as it copies from then writes to, itself.
Well, I'm almost out of ideas on this. I just can't imagine why better performance can't be achieved. And as Ashy said, I should have seen a much more significant improvement going from 2 pass to CQ. I verified that my CPU does read 99% while working. I've also verified that there are no other process stealing any clock. Memory is also fine at 256 with less than half actually utilized. I'm defragmenting my C: (video work is done on a 2nd disk) right now with the thought that the Windows swap file is causing the slow down. The disk was fairly fragmented and nearly full (2G avail). At this point, I'm beginning to grasp for straws. :(
Thanks to all of you who tried to help and for offering the various suggestions!
Well Griff I would agree with what you say for the most part, but it all depends on the actual system configuration. The suggestion was merely made from another post which had a similar problem and was solved this way.
It is not necessarily true that two drives are faster or the same speed as one, and I know this from experience, even if they are on different channels.
It depends whether the drives are matched in thier specs. If one drive is slow then no matter how fast the other drive is the data can only be transferred at the rate the slow drive is able send or recieve it thus creating a system slow down. Whereas a fast drive with enough available memory and a higher bandwith may be able to read and write the data at a higher rate thus increasing the speed.
I too don't want to get into an argument here. However, even if you mismatch drives, true you can only read and write as fast as the slowest drive. But, you can have the fastest drive in the world (and by fast were talking data transfer here) and the the process of copying large files (and lets be honest mpeg creaton moves around large files) will slow it to a snails pace. This is because of the physical limitations of the drive heads.
It's a bit like the difference between virtual and real RAM. The virtual RAM relies on the mechanical hard disk, real RAM is pureley electronic. The more real RAM you have the faster your system. The less real RAM you have the slower your system, as you HD churns away IOing.
That may be so Griff, but I know from my own experience that two drives isn't necessarily faster than one. I have tried it with a fast ATA 100 IBM deskstar ans a crappy old quantum. Using the the quantum together with the IBM slowed things right down. It just couldn't write the information fast enough whereas the IBM deskstar's are known for fast access times and was much quicker even though it was reading and writing to itself.
Look at this way.
Once the faster drive has written all the data to RAM there is nowhere else for it go, so it has then to use the swap file which will be on the faster main drive, so all it is doing in the end is still writing the data back to itself until the slower drive is ready to receive it which means it has to write the data to RAM twice, once initially and then from the swapfile. This buffering is what will slow down the process.
On the other hand the faster drive will write the data to RAM and will then write it back to itself much quicker than the slower drive can take it from RAM via the swapfile.
ASHY
I hope you understand this as it's not meant to be an attack.
>That may be so Griff, but I know from my own experience that two drives isn't necessarily faster than one. I have tried it with a fast ATA 100 IBM deskstar ans a crappy old quantum. Using the the quantum together with the IBM slowed things right down.
Well of couse it would. If you introduce a bottleneck (such as a much slower hard drive) in the system it's not going to speed things up.
How do you think you're system would perform if you had two Deskstars?
Here's an idea, get another Deskstar and put the two HD's in RAID 0.
My area of work generates absolutely huge files (many GB in size) and I/O speed is critical. I have 3 SCSI hard drives (not in RAID) on my work machine, which transfer thes massive files betweeb them. If I had one big SCSI hard drive - even a SCSI 320 - the software which uses these files, and currently utilises 99% CPU time, would have to wait, and the the run times would increase - wer'e talking run times of up to a week, as it is!
Griff you are missing my point entirely. I absolutely agree with you that 2 Deskstars would definitely out perform one and especially on a RAID system.
>Well of couse it would. If you introduce a bottleneck (such as a much slower hard drive) in the system it's not going to speed things up.
This is the point Iam trying to make. If the user is experiencing a slow down using two drives then Iam simply suggesting that the bottleneck, as you so aptly refer, may be causing this.
I simply wanted to make clear that 2 drives isn't ALWAYS better than one.
WHEN I USE TMPGENC TO CONVERT AVI FILE TO MPEG FILE, AFTER 6 OR 7 MINUTES ENCODING I HAVE THIS MESSAGE.
"READ ERROR OCCURED AT ADRESS 0047310c OF MODULE TMPGENC.EXE WITH 12618A8F
PLEASE HELP ME.
Start the program, go to bed, wake up in morning with a beautiful MPEG all ready and encoded.
Ok, if you don't want to do that, here is a round about way to acheive what you ask.
When you decide to stop the program just click the stop button, it will then ask you to confirm. Don't click any thing else. Now put your computer into hibernation.
When you decide to restart the encoding just boot up your machine and cancel the abort and TMPG will continue encoding.
I know this questions comes back several times on the BBS but i'm having trouble opening my video source. When I select the avi-file i get the famous "can not open or unsupported" error. The VFPAPI plug-ins in the enviromental settings are all set to priority 0 exept for the Directshow Multimedia File Reader who is set to priority 2. I have the Divx 5.0.2 codec and working with Win2k Pro. Can someone help me ?
I am encoding AVI @ 23.976 fps to MPEG-1 for VCD, Error '-537403781 164192' I Press OK and encoding continues, but the error keeps coming up every few seconds.
I am trying to use the TNPGenc inverse telecine filter on a 29.97 Quicktime movie (8-Bit 4:2:2 Component Y'CbCr (720x486) Quicktime captured via Adobe Premiere 6.0 (Windows 2000 Pro) with a Targa 2000 DDR card from DigiBeta). The frame count of the movie is 945 frames, but the TMPGenc IVTC filter sees it as 94599 frames!! If I perform the IVTC process, it takes much longer than it should (since it is processing phantom frames) and the process never really works. If I convert the Quicktime movie to an AVI first, the IVTC works fine. How do I get the TMPenc IVTC filter to recognize the correct frame count of a Quicktime movie and perform properly? Or is this a bug?
AVIsynth will not open a component 4:2:2 Quicktime movie. AVIsynth will only open files that Windows Media Player (DirectShow) supports. Unfortunately, component (SDI) video is not on the list.
I have made an avi movie. The time is 42.59 min. If I start the wizard in version 2.53 and 2.54a the time is also 42.59 and decoding is ok. But if i start the wizard in version 2.55 and 2.56 (plus version) the movie is 66.32 min long (page 4 of 5 in the wizard) and the file is about 2/3 to long.
Yes, I've been having this problem too. It's been driving me bonkers.
I've spent the last week trying to use TMPGenc to transcode an MPEG-2
to a lower bitrate and it always ends up about 10 minutes longer.
I have the excact opposite problem. All the files are far too small.
Im going back to an older TMPG since it works in that.
The ones i have had problems with so far are the 2 latest versions.