tmpgenc VS cce VS mainconcept VS procoder: THE ENCODERS TEST

[crespo80]

I searched a lot for a good comparison among these popular encoders, but it's hard to find something useful.
So I decided to make my own test.
I tested the full retail versions of the best encoders of the moment:
CCE-SP 2.67.00.27
TMPGenc Plus 2.521.58.169
MainConcept MPEG encoder 1.04.02.00
Canopus Procoder 2.01.30.0

Here is the page of the test
http://www.crespo80.altervista.org/site/the_encoders_test_1.htm


Vote the poll after you have read the test!
The test is open to all possible critics, help me to improve it!

[Piper]

@crespo80: Thanks for posting a current MPEG-1/2 encoder comparison, I've been looking for something like this recently as well.
Your .rar files appear to be corrupted though. Can you post high quality jpg's?

[crespo80]

I tried to download the files from another PC and was all OK!
Did you right clicked the files and selectde "save object"?
Because if you left click, It doesn't dowload anything!

The images are Tif compressed, this is the best solution because is a lossless format with some kind of compression (like HuffyUV codec).
The raw images were around 1100K, in Tif format the size vary from 350K to 800K.

The jpg format isn't usable because, with 5x zoom, become visible the artifacts of jpg alghortitm (even with the lowest possible compression of 1) and the test is falsed!

[Piper]

The files download fine, the issue is with uncompressing them. Winrar reports "The archive is either in unknown format or damaged".

Your first file is also only about 26kb in size which suggests something's wrong. It's also unlikely any space savings will be made compressing to RAR in the first place. Is your original TIFF approx 26kb? ;)

EDIT: There are download issues using Firefox. Link works properly with IE.

[buzzqw]

archives are correct, make sure you have latest winrar installed. (btw link isn't fuctional with FireFox)

BHH

[rui]

I saw your comparison. My main interest was between CCE and Procoder, but with a source consisting of DVD material.

I am a DVD-RB user (awsome software), and it has the option of choosing between CCE or Procoder to do the encoding part (to reduce a DVD-9 to DVD-5).

But, by watching your test, CCE and Procoder are very similar in results, being Procoder much more slower than CCE.

[dplaton]

The best encoding is a fairly simple chalange.
The real chalange is what is the best trade-off.
Encoding is just a simple step.
Let't say the task is preserve VHS memories, you have to:
1. capture VHS;
2. process the raw avi;
3. encode;
4. authoring;
5. play.
Every step is not easy.

Supose that NuEnc is the best encoder, you have to feed it by Avisynth,
but AviSynth give on my computer some random jumping or green frames.
So NuEnc is useless.
With VirtualDub I can not capture at an apropriate resolusion.
With WM9Cap I can not use codecs, but it captures at the desired resolution.
It is also important the final player or players.
For me what is working <best> is:
1. Capturing 720x576 with WM9Cap;
2. Process with VirtualDubMod MPEG2
crop the bottom 10 pixels
deinterlace MAP is very slow but is the <best>
resize bicubic 720x576
color enhance MPEG2 4:2:0
3. CEE encode 1 pass VBR Q=30, max=9000, matrix MPEG2,DCT=10, no sound
4. BeSweet avi -> mp2
5. guessing the audio delay
6. muxing with a MPEG2 editor
7. TMPEG authoring.

That is best?

[crespo80]

In my case was not so complicate to find the best settings, there are no many possibilities:


1. Capture in fly200TV, the best program for Philips SAA713X chip
YUY2 color space (best for this program)
720x576 res (best for this chip)
right crop 16 pix -> 704x576
real-time compress to HuffyUV (best lossless codec)
master stream audio (because some times the tape jumps and the cap drops frames, so audio gets a little desynch during capture)

2. open the file in AviSynth just for trim and letterbox (no filters)

3. encode audio and video with CCE (i found it better for video, no problems with audio)
res=704x576 q=40 max_bitrate=9656 audio_mono=192kbps all_filters=OFF interlaced_encode

4. Author with TMPGenc DVD author


It's not so complicate for me. I'm having troubles only with a too high volume from the VCR to the line-in connector. I have to set the volume to the very min setting both in audio control panel and in fly200tv (just one step over 0!!!). And some times I have clipping.
I have to find some regulation on my VCR or on the pC

[dplaton]

My DVD standalone want 224 audio.
The audio mpa made by CEE sound to me bad.
The WM9Cap override the TVtuner and the video is better,
using other programs including the TVtuner one the video is folding in the upper side, and that is why are unusable.
Other capture solution for me is Ulead.

All the best,
Dan

[dplaton]

I don' know what to belive.
Deinterlace smooth is supose to duble the frame rate. Right?
But please read this topic
http://forum.doom9.org/showthread.php?s=&threadid=85072

[zilog jones]

Hey, why didn't you include QuEnc in your test? I've found it to be very good - in some tests I did it proved better than TMPGEnc, especially when encoding iterlaced video. And it's FREE!!!

Just remember it only takes avs scripts and YV12 colour.

[dplaton]

I used QuEnc but it produced artifacts or (depending on version) color jumpins.

Again CEE is very good for photo or speed, if you want to make a homevideo forget about it. At a formidable price offers too little.
If you encode interlaced (there is no reasonable reason to deinterlace, but if you want to increase the encoding duration and worst the quality is reasonable to deinterlace. MPEG2 interlaced played on TV or on PC with a DVD player is the true solution. AVI you just speed the movie 2xfps and it is also perfect. If you want to take still from the movie/clip is no need to deinterlace, just separate fields. This my opinion, and I think that doom9 guide sould explain this in plain english)
So if you want to encode interlaced and you capture is bottom field first CEE will fail. That can be corrected by various technics, but at the end of the day you will find that it was better to use TMPEGenc.
Once again for home movies.
The comparation is frecvently done, just on a photographic base,
but the movie flow, and in this flowing the perfect photographic aspect dosen't matter.
A true comparation must to take in consideration more aspects than the photo quality.
Buy some better hardware ($$CCE-$$TMPEGenc) and at the same price you will have a better PC and probably the best MPEG2 encoder.

[cweb]

Please include FreeEnc too in your test.

[dplaton]

First of all I must to say that I was too critic to CEE.
Please forgive this over criticism.

I think that this topic is very usefull because someone can learn a lot about encoding, so thank you crespo80.

What I can't find in the encoding test conditions is the matrix used.

You used the standard matrix of each encoder or the standard MPEG2 matrix? And why you did that?

I did myself some modest tests and I found that there is a significant difference for 8 and 10 DC precision.

Why you used 9 as DC precision?

Nice to read your topic.

[iwod]

Why isn't TMPGenc Xpress included?

And even though all previous post i read point to ProCoder this test point to CCE.......

And it is fully that there is no TMPGenc supporters here....

Any more conclusive information?

[crespo80]

Originally posted by cweb
Please include FreeEnc too in your test.
Originally posted by zilog jones
Hey, why didn't you include QuEnc in your test?
I thought they both are not on the same levels of the other encoders and now I lost test captures and havo no time to remake the entire test.
Originally posted by dplaton
The comparation is frecvently done, just on a photographic base,
but the movie flow, and in this flowing the perfect photographic aspect dosen't matter.
A true comparation must to take in consideration more aspects than the photo quality.
What do you mean with movie flow? Which other aspect have to be taken in consideration?
Originally posted by dplaton
You used the standard matrix of each encoder or the standard MPEG2 matrix? And why you did that?
Why you used 9 as DC precision?
I used tha standard matrix of each encoder because I noticed that with MPEG2 matrix there was a big loss of quality at the choosen bitrate.
I did not find any difference with different DC precisions, so I choose 9 without a specific reason
Originally posted by iwod
Why isn't TMPGenc Xpress included?
And even though all previous post i read point to ProCoder this test point to CCE.......
And it is fully that there is no TMPGenc supporters here....
Any more conclusive information?
I thik commercial version is quite similar to the last free one so I choosed the "economic" version.
About Procoder, I think it-s good but not the best.
My conclusive idea is
CCE FOREVER
:D