I have found no noticeable difference between the Vegas MainConcept encoder and TMPGEnc. But a BIG difference between Vegas and Procoder 3. Vegas and TMPGEnc both wash out the image slightly, to my eyes anyway.

I used Debug Mode's frameserver.

TMPGEnc has worked flawlessly for me. I started using it when I was working with Liquid because Liquid doesn't do a double-pass and that makes a substantial and noticeable difference. I've never seen any sort of washed out effect. If anything, the image looks sharper and crisper than prior to encoding. Perhaps it's the settings. Regardless, I'm happy with the results.

**Vegas and TMPGEnc both wash out the image slightly, to my eyes anyway.**
I know what you mean, but it seems to be a product of lower bitrates (4mbs and lower). At higher bitrates (7-8 mbs) I am pretty happy with the results in Mainconcept. I haven't tried Procoder.

I was unhappy with the Vegas HDV to SD Mpeg results I was getting so I started using TMPEG Plus 4.5. It uses the Lancos-3 method for resizing and I have found the results to be better on the material I have been outputting to DVD.

**it seems to be a product of lower bitrates (4mbs and lower)**

Hmm, can't recall what settings I was using, but at the time was experimenting with SD to SD frameserving. Now I am entirely HDV, so I'll do some more experimenting.

I'm not complaining about the results from Vegas/TMPG, I'm content with them, but I *know* I could achieve better results if I bothered to investigate how.

Do your own testing. There are definitely significant performance differences between encoders: some are much faster than the MainConcept encoder built into Vegas. However, at the risk of contradicting what I just wrote, it is actually quite difficult to do performance testing because all of the encoders "expose" different controls to the end user, so you cannot set them all to exactly the same specifications.

Also, I frameserve almost every day, and I can tell you that even after you sort of know what you are doing, it is VERY easy to screw up the settings and end up with something which has different contrast from the original, either looking washed out, or else too contrasty. I always do a test encode and then bring the result back into the Vegas timeline and do a quick A/B comparison to make sure I don't see any difference, either in the preview screen or on the scopes. The levels should look 99.999% the same (i.e., no discernible difference whatsoever).

The things you must get correct when using any encoder outside of Vegas are the settings having to do with color space and levels. When you frameserver from Vegas, you first need to decide whether to use RGB24, RGB32, or YUV. Then, in your encoder, you need to look to see if there is some setting labled "0 ... 255" or "16 ... 235" or something similar. In TMPGEnc, there is a setting "output yuv data as basic ycbcr not ccIR601" which you must play around with until you get the correct results.

So, any differences having to do with contrast or levels is almost certainly cockpit error in dealing with these settings and NOT a problem inherhent to the encoder. Instead, the differences between the encoders will primarily be seen around fast-moving, high contrast edges; in transitions and dissolves; on any scenes involving smoke and fog; and possibly in fine details, especially when motion is involved.

Unless you have a specific problem with the Vegas MPEG-2 encoder, or have heard that another encoder can cut your encoding time by a significant amount, I wouldn't worry about it. The MainConcept encoder is pretty darn good, and the default settings (other than changing average bitrate) work for most situations.

**Vegas and TMPGEnc both wash out the image slightly, to my eyes anyway.**

I think this is a colorspace issue. With the default MainConcept settings I get slightly washed out videos too, but when I render with the16..235 option enabled, the black levels look right to me.

EDIT: Sorry, I should have read further down before replying. I just realized the previous post says the same thing more thoroughly. :)

DRG

--
D.R. Greenlaw
Little Green Dog
www.littlegreendog.com

D.R. Greenlaw
Senior Digital Artist
Rhythm & Hues Studios - The Box
www.rhythm.com

I think the MainConcept encoder is good, but I find TMPGEnc to be better. I notice the difference easily in things like one second fades up or down from black. In general, I think TMPGEnc's motion detection is better. And when I get to lower data rates in long projects, there are more controls that can improve things in TMPGEnc that aren't in the MainConcept encoder, such as block noise reduction, for instance.

I typically frameserve all my projects to TMPGEnc.

If I'm doing something short where I'd only be using CBR, I sometimes use the MainConcept encoder, but once a project is longer and needs 2-pass VBR, I only use TMPGEnc, based on my own test results.

I notice the difference easily in things like one second fades up or down from black. In general, I think TMPGEnc's motion detection is better. Yes, that's exactly what a better encoder will do.