Under the GOP (Group of Pictures) tab, choose I Pictures only, and lower the GOP to 0. Make sure that the bitrate is 25,000 - 30,000.
Now your 24p clips will render in a couple minutes. No compression artifacts, and faster access times while editing.
This was all based off of my own experimenting after discovering this post.
Also note that it is ideal for this bitrate to choose 10 bit Dc component instead of nine. (Not much of a diff, just thought I should say)![]()
Do you mean GOP max frames = 0? I assume that you do. I can't tell if that does a better job in terms of video quality, but it does speed up the encoding.
Make sure the bitrate is 30,000. 24 non interlaced frames is actually very easy compression, and so it requires less than 60i. If its 60i, then do not mess around with the GOP compression. Only if its 24p. W/ 60i you do have to set the compression at standard. I could post screen caps or video in 1080 if necessary. It really does work if all the settings are right.
Are you suggesting to use these settings whilst performing the pull down? ie. Set the filter to do the inverse pulldown, then on the encoder options to choose the settings you have proposed? That is what I did, but it looked horrible.
Can you perhaps just explain fully where, when, and why you used these settings as it would help me to understand. Thanks.
Yep thats it.
More compression leads to more bitrate for the image to work with, but harder access and more artifacting. And so I noticed that if the bitrate is high enough w/ no compression on 24p streams, you really do get the benefit of no artifacting w/out losing image quality. 30,000 is the sweet spot here - enough bitrate to capture every detail that was in the original file, and the filesize isn't too large. Its all good.
Its not much of a benefit to image quality, just a practical sweet spot.
I didn't know this at first, but you have been having some serious trouble w/ 24p in the first place. That makes all the diff. You see, I use an HG10, and I found out that I had to use Elecard, and not tmpg, to initially encode into vanilla HDV. Tmpg simply does not read avchd right, and as a result incorrectly writes it to HDV. And what happened next? Simply because the HDV it wrote was slightly off, the encoding took forever and the resulting video looked like crap cause it was a stuttering interlaced mess. The lesson: The File You Import Must Perfectly Match The HDV Profile. Otherwise everything doesn't work right. And so your issue must be before tmpg gets it. Whatever you use to capture is wrapping the files messily. Perhaps a setting is off. I knew when I got 24p right using the method in this thread because the video ran smooth as butter and Premiere said 23.976, two things I never had in conjunction w/ each other. Upon close examination, every frame was purely progressive.
But it took an hour to render 1 minute. And that just won't cut it when I make a full length. So I had to find a way to make it faster, while not losing any quality. (If I had a blowup distributed, this would be critical) Thats what the above is for. I cheered when the thing took 2 minutes to render and has less artifacting than before. It accesses fast too.
Everyone else has it and can benefit from this discovery. But for you pascal, I would suggest perhaps re-encoding the HDV you captured in a diff program as vanilla 60i HDV, and then importing it to tmpgenc. The prob. is happening before the file ever hits tmpgenc.
I've been using this since the OP posted (last july was it? wow), and it works flawlessly. Just in case someone is curious and peeks in on this thread, here's basically what I do:
1) Capture from the camera using Vegas 8.
2) Use this great tool called 'MPEG-VCR' to basically step through the raw .m2ts and cut out the parts I don't want, as well as since I reuse tapes I get a few frames of distortion between scenes, I cut those out. You save as another .m2t, it doesn't do any encoding or compression or anything, it just cuts out the frames you don't want.
3) Run the new .m2t through TMPGEnc 4.0 Xpress following the instructions in OP's post.
4) Throw that into Vegas. Woo!
Instead of HDV you can output to WMV and set high bitrate like 25 or 30 if you want. Windows Media Encoder is pretty smart in detecting cadence and removing pulldown. I tried converting AVCHD from HG10 into WMV using about the same bitrate as source bitrate, and result looks pretty nice.
sage
i was wondering can i use sony vegas to convert the avchd format to another format and then use tmpgenc to remove the pulldown?
basically i want to replace elecard with vegas, is this a good idea?
with the elecard demo, im not getting audio from the finished product.
anyone know why?
Last edited by thebossman; 2008 June 20th at 01:22.
So, would the technique outlined in this guide be good if you want to mess around with your footy in AfterFX?
I remember reading about how AE doesnt like formats like m2t or mpeg.
Could someone do a writeup or tell me the steps that'd be best in removing pulldown and converting the footy to a format thatd be good for AE color correcting, effects, and compositing?
Hi, I have a bit of a newbie question about HDV 24p editing:
I'm using Premiere 3.20. Rather than putting all my raw footage through reverse pulldown processing in TMPGEnc and then editing it in Premiere, would it make more sense to capture the footage in Premiere in the weird 29.97 frame rate, edit it that way, export a video and then process the final video to remove pulldown in TMPGEnc? Or would that not work?
It would just be more more convenient to process the final 5 minute video than the 90 minutes of footage that I'm editing down.
Thanks!
So can somebody confirm that this is the official way to do it in TMPgenc?
Or is there a thread that shows the official and best way?
Sincerely,
Paul Tarlevs (PT Productions)
I use this method and it works.