Comparison of TMPGEnc 2.56 vs. CCE 2.64 vs. Main Concept

[adam]

i was told that it does a 2 pass in the same way that a hardware encoder does a two pass - that is that it goes forward several frames before encoding them

All encoders do that regardless of what mode they are encoding in. That is what motion searching is. Anytime you encode in mpg you have I, B, and P frames that make up each gop. Your bitrate settings determine how many bits get distributed to each gop and from there the encoder distributes those bits to each frame depending on its priority. I frames have a set bitrate. With P frames the encoder looks forward to future frames to determine how many bits that P frame needs. With B frames the encoder looks both to future as well as past frames.

An encoder never just looks at a frame and says, ahh it needs this many bits. It always has to look either forward, backwards, or both to make its decisions.

It seems MainConcept is only doing one pass which explains its speed. If you were to encode in 1 pass in CCE my guess is that it would be faster. Whether or not it would be as high a quality is another question.

[BJ_M]

that is 1 gop motion sensing which yes i know all encoders do or attempt .. i meant it goes forward at times a lot -- 10 gops or more ..

my hardware optibase encoder does 2 pass vbr in real time by using its buffer .. but of course it cant buffer more than a few sec at a time ..

the main concept encoder acts like the hardware 2 pass enocder - which i know is doing 2 pass for sure ..

[ajp]

BJ_M,

maybe that's how they do it. The quality, speed and ability to set max/avg/min would certainly suggest it uses some other method compared to cce/tmpg etc. It certainly sounds like it's possible to do in software too using clever programming...

What I don't like in normal CQ modes like Tmpgenc or CCE uses is that you can't set avg bitrate, I like that in the Mainconcept encoder. On another note, CCE 2.66 is also very very fast in CQ mode... probably even faster. Not sure how the quality compares though...

[ajp]

Updated review for the new 1.2 version of Mainconcept encoder which should be released shortly, hopefully this week. Most problems seem to be fixed and there are even a few new features added/improved...

http://www.softdeko.8m.com/mainconcept_mpeg_encoder.htm

[DVWannaB]

Updated review for the new 1.2 version of Mainconcept encoder which should be released shortly, hopefully this week. Most problems seem to be fixed and there are even a few new features added/improved...

http://www.softdeko.8m.com/mainconcept_mpeg_encoder.htm

Thats interesting. No mention of it on their official website. Go figure.

[ajp]

Well, I think this is sort of a preview of the new version. They wrote on the support forum that it will be out at the end of the week, so I'm hoping tomorrow ...

[Lamont Cranston]

There remains the problem of pulsating artifacts in the high contrast areas problem that appears in MSPV6.5a and VS6 V6.02 and if the problem is also in the MainConcept encoder. Maybe the problem is limited to the SDK. There are also several other known bugs that were not mentioned. IMHO It is a very impressive encoder and the bugs will be corrected... That's why I sprung for $149.

ajp Have you tried to batch encode several avi's into 1 meg file?

[ajp]

I haven't noticed any pulsating artifacts, not yet anyway. The only real serious bug I've come across is the large file bug. Let's see the official bugfix list first. I'm just glad they got the new version out so fast! I'm sure new versions will continue to come out now that they've fixed the worst bugs, glad I bought it...

I've tried batch encoding several DV files into one mpeg but only once and it worked fine?

[Lamont Cranston]

I've tried batch encoding several DV files into one mpeg but only once and it worked fine?

I have tried batch encoding 2 NTSC 720 x 480 DV type 1 avi files to one mpeg file (DVD default settings) on 1 win 2k system and 2 win 98 systems. On all threee systems windows aborted MainConcept V1.1 for perfroming illegal operations. Based on your experience it must work on WinXP.

[ajp]

I think the ones I tried was DV type-2, I also use Win2k though and not XP...
The files weren't very bigh though. Maybe it has something to do with the directshow priority problem in 1.1? You'll have to try again with 1.2...

[sherlock]

I don"t know what the exact criteria you used for your test or the exact specifics of your system but the results that you got Totally conflict with the results that i and probably al lot of poeple get..On my system the speed of CCE is about 25%- 300% faster than Main Concept standalone..Averageing about 3-5 times real time or up to 150FPS, when encodeing Uncompressed YUV2 avi files..and I allways have a lot of Macro blocks on low bitrates in High motion scenes with Main Concept,low bitrates meaning under 2000KBS with Full Resolution AVI, were I can get away with useing CCE at 2000KBS at 704+480 with minimal artifacts..I guess every system acts differently with different sources so these "Test" don"t really show the overall quality and Compatibility of a Product...

[Lamont Cranston]

I allways have a lot of Macro blocks on low bitrates in High motion scenes with Main Concept,low bitrates meaning under 2000KBS with Full Resolution AVI, were I can get away with useing CCE at 2000KBS at 704+480 with minimal artifacts.

sherlock, please post the mpeg encoded clip, 704 x 480 video data rate = 2000kbps, of your test clip with high motion scenes with minimal artifacts and no macroblocks. That has to be a first.

[sherlock]

No Problem....How do I post it??Well i guess that is the problem..Should I post it as a Single frame mpeg2 file or a Jpeg of the Mpeg2 file, and were to post and I would be more than happy to...