MainConcept MPEG Encoder 1.4

[bb]

The brand new version 1.4 of MainConcept's MPEG Encoder is really worth a try. It's easy to use and provides excellent quality.

What do you think?

bb

[lilhobo]

roll out the rest results thanks

[waldok]

yes,

Toll out the rest tesults rhanks

Waldok:D

[bb]

Be patient, pals...

bb

[bb]

Ok, here are some testing results:

Source was a progressive 16:9 DV type-1 low light shot with lots of noise. It was 5342 frames long. I compared four encoders at bitrates of 4000 kbps and 6000 kbps video, CBR / 1-pass VBR, and 224 kbps MP2 audio each: Ulead MediaStudio Pro's integrated MPEG.Now Encoder, CCE 2.66 SP, TMPGEnc 2.52, and MainConcept MPEG Encoder 1.4.

Encoding times:
1. MPEG.Now____________________7min 40sec
2. CCE (no filtering)__________3min 44sec
3. TMPGEnc (Motion Estimate)__10min 17sec
4. MainConcept_________________4min 22sec

Filesizes 6000 kbps:
1. MPEG.Now_____195.966 KB
2. CCE__________168.685 KB
3. TMPGEnc______164.491 KB
4. MainConcept__165.764 KB

Filesizes 4000 kbps:
1. MPEG.Now_____100.350 KB
2. CCE__________114.547 KB
3. TMPGEnc______111.653 KB
4. MainConcept__112.928 KB


To compare the quality of each encoding I used the following script and saved to a Huffyuv file:

cropx=110
cropy=140
vcce=MPEG2Source("C:\Test\Test_CCE.d2v").Crop(cropx,cropy,-cropx,-cropy)
vmsp=MPEG2Source("C:\Test\Test_MSP.d2v").Crop(cropx,cropy,-cropx,-cropy)
vtmp=MPEG2Source("C:\Test\Test_TMPGEnc.d2v").Crop(cropx,cropy,-cropx,-cropy)
vmco=MPEG2Source("C:\Test\Test_MainConcept.d2v").Crop(cropx,cropy,-cropx,-cropy)

vcce=vcce.Subtitle("CCE")
vmsp=vmsp.Subtitle("MediaStudio Pro")
vtmp=vtmp.Subtitle("TMPGEnc")
vmco=vmco.Subtitle("MainConcept")

StackHorizontal(StackVertical(vcce,vtmp), StackVertical(vmsp, vmco))


In a second encoding at 4000 kbps I tried CCE with "Low pass filter" set to 8 and "Effect restricted vertical filter" set to 4. The filesize turned out to be exactly the same, but the quality was better because of the noise reduction. Processing time increased, too: 4min 17sec (almost identical to MainConcept).

All settings were kept according to the source, and they were equalized among the different encoders wherever possible. Of course I know that my quick evaluation does not meet the criteria of a professional encoder test - much more different settings and sources would have to be evaluated, especially some interlaced sources, too. But I think it's a first step, and it shows some tendencies.

The quality of all encoders were comparable - no big differences were visible. Ulead's encoder seems to use some internal smoothing, which is not configurable, but increases quality - and processing time. CCE and MainConcept are clearly the fastest encoders, and especially the latter has many settings and therefore some tweaking potential. I wasn't surprised about the medium performance of Ulead's encoder, but I thought TMPGEnc would be faster. Although its slowness is well known, I had expected some more speed at the fast "Motion Estimate" setting. All encoders except Ulead's MPEG.Now can be tweaked a lot, e.g. there are different encoding modes like multi-pass VBR.

Conclusion:
The source quality and format have a much higher impact on the final quality than the encoders tested, if you create DVDs at average bitrates above 4000 kbps. Price and speed should trigger your purchase decision more than quality aspects, because the quality differences are marginal.

Opinions?

bb

[ronnylov]

Why don't you test with SSIM in avisynth?
http://forum.doom9.org/showthread.php?s=&threadid=61128

I did a comparison but using 704x576 huffyuv capture of a VHS-C source. I posted some results here:
http://forum.doom9.org/showthread.php?s=&threadid=61025