Do you tmpgenc users really see a difference?

Julius_ wrote on 5/5/2014, 10:44 AM
Good Day everyone,

I recently installed the latest frameserver (v2.14) and got it to framserver from Vegas V10 (64-bit).
I also have tmpgenc works 5.

My primary goal is to create a DVD from HD footage(60i) for hdtv's. My steps were

1.Frameserver in Vegas (matched media properties from footage
2.Opened tmpgenc works 5 and set +42 contrast and +2 luminance in the TMW5 filters
3.Encode use the standard dvd template. I check the "limit to 8000" bitrate and set the minimum bitrate to 4000
4. Encoded the video from vegas mpg (the average bitrate was 7800 VBR 2 pass

The encoding made use of my CPU (18% and CUDA (72%) and I must say that rendering was way faster than vegas.

I burned the DVD and played it on my 24" HDTV.

I didn't notice any difference from the same DVD burned out of vegas that uses mainconcept mpeg-2 (ntsc widescreen).

Now, I'm wondering if tmpgenc is worth it or not.....Do you tmpgenc users really see a difference?

Thanks

Comments

MikeLV wrote on 5/5/2014, 11:18 AM
I just downloaded a trial of the software and am currently experimenting with its x264 encoder, but I also have a need to encode 60i HD footage to DVD. Haven't yet messed with MPEG2 encoding. I do like the fact that VMW5 makes use of CUDA for encoding, I'm definitely seeing much faster encodes with it, even when frameserving. You can see my post on that topic here if you're curious: http://www.sonycreativesoftware.com/forums/ShowMessage.asp?MessageID=893385&Replies=41
musicvid10 wrote on 5/5/2014, 11:48 AM
TMPGENC was slightly better than Main concept mpeg-2, about ten years ago.
Today it's just another urban myth, albeit a persistent one.
malowz wrote on 5/5/2014, 11:51 AM
for mpeg2, carbon coder/procoder is my choice. "back in the day", when i tested all avaliable encoders, they where the best by far, specially with low bitrate (for like, 2h on a SL DVD)
videoITguy wrote on 5/5/2014, 12:06 PM
Not only an urban myth it is (about better quality), but the fallacy within the whole concept is what yardstick do you choose to use? At what constraints? - low bit-rate? very special source footage? NTSC or PAL? Does everyone agree when a test is performed within some constraints? At what percentage ? is it going to make a difference when DVD is played in all kinds of uncontrolled environments?

NickHope wrote on 5/5/2014, 12:18 PM
In my testing a few years ago, CinemaCraft was easily the fastest and also the best quality per bit. TMPGEnc and Procoder were joint second. MainConcept was a poor fourth. Sadly the excellent CinemaCraft Basic is not available any more. If you can somehow get a licence, get one! Until Musicvid's comment I hadn't heard any word that the MC MPEG-2 encoder has improved.
Arthur.S wrote on 5/5/2014, 12:30 PM
Definitely not a myth, that's for sure. I guess it can depend on the source footage. I'd always been happy enough with the Vegas re-size from HDV 50i to SD (DVD). It was 'OK'. No more. But then a couple of years ago I had a job whereby the Vegas resize to DVD from HDV and HD REALLY looked poor. Particularly jaggies on a bride's veil during the vows. I posted here with the problems. Unfortunately the thread is too old to find in the 'search' now. I looked around for a better option. To my surprise even some of the very cheap ($45 type) encoders made a better job than Vegas. I tried out TMPGEnc video mastering works which also made a great job. Even though that was more expensive than the cheapos I stuck with that, more for the well known name than anything.

I can appreciate that there should be controlled environments etc for REAL bonafide testing, but for me, if I can drop a clip into both Vegas and TVW5 and see a difference EVERY time, that's good enough for me. Trust me, I would NOT be bothering with it if not, especially as the TVW5 files do not smart render! That can be a real PITA.

What baffles me here is that some of the "urban myth" quotes are from folks who jump through hoops to use handbrake plus other programmes....to get a better resize/encode than Vegas can provide. Doooooh!!

malowz wrote on 5/5/2014, 12:36 PM
@videoITguy

ive tested with NTSC DV sources, 30i, static scenes, low motion scenes, texture motion, very detailed scenes. (mix from various events i recorded)

encoding with 4mbps with maximum quality possible by the encoders, 2 pass if available.

most had problems with texture ("moving" static objects) and mosquito noise (mainconcept was the worse). others where "blocky", added compression noise, etc...

cinamacraft was very good and very fast for progressive sources, but not that good for interlaced content (even tweaking internal filters options to enhance visual). even using filters in avisynth like Convolution3D to "stabilize" the video for better encoding, Procoder was better for 30i. (looks like procoder has its own internal filtering, mainly with lower bitrates, that result in good looking results, mainly low artifacts like mosquito noise

Videos where interpolated from 4:1:1 to 4:2:2 during or prior to encoding.

EDIT: i used procoder for ages, a few years im using carbon coder (same "internals", but updated). now i make h.264 blu-ray encoding with Adobe Media Encoder (DVD-A accepts 30i without recompress or errors) and with avisynth, i resize the HD to SD and encode with Carbon Coder for DVD.
musicvid10 wrote on 5/5/2014, 4:21 PM
Nick,
The thing I remember about Vegas 2 is that mpeg-2 was full of wiggle-worms. TMPG and Ulead were much better at the time. Sometime between then and when I got Vegas 8, Mainconcept got a lot better, and I'm still using it (with a high minimum bitrate).
;?)
dxdy wrote on 5/5/2014, 4:49 PM
I believe I can see a difference. When I got my first HD camera, I was, like everyone else, horrified at the downgrade from HD to SD for DVD. I experimented with various encoders and settled on TMPGEnc. The interface is quirky, settings are buried in odd places, but it puts out a very nice DVD-ready MPEG 2 product.

I tried MC again in 12, but all it did was crash. After one of those MS Tuesday downloads, MC started working again, and I still think the current TMPGEnc produces a better finished MPEG2.
riredale wrote on 5/5/2014, 4:50 PM
Back when I first got into making DVDs I used CinemaCraft exclusively. Wonderful control of all variables, and the ability to do a custom tweaking of any portion of the whole video. I remember doing lots of tests with CC and MainConcept (this was back in 2003/4) and there was no comparison. Just now I pulled out a 2003 DVD I'd made of a choir tour to Austria, and my DVD player shows that in general the average bitrate is about 5-6Mb/sec, with excellent quality. That average bitrate would correspond to a maximum of about two hours of video on a single-layer DVD, which I did on quite a few projects back then.

In later years I noted that unless I was really pushing a low bitrate then MC did fine and it certainly made my workflow faster, so I began using MC more. Nowadays I almost never do 2-hour projects on a single-layer disk and I almost never use CC any more. At high bitrates all the encoders are probably pretty much identical. Tried TMPGenc once a long time ago, nothing special compared to CC.

I do think MC has improved considerably over the years, but I have no idea how it would compare today to TMPGenc.
musicvid10 wrote on 5/5/2014, 5:27 PM
OT, but I got a newer Magnavox set-top DVD recorder to replace my old Panny.
Mpeg-2 quality at 2 hrs. is superb.