Mainconcept Encoder How does it compare

[ZenZen]

Thanks for all your help BJ_M

A little trick I'd like to pass on to anyone using MC encoder.
Let's say that you are half way through an encode, and you suddenly realise you want to capture something on Cable. What to do?

Well, MC Encoder actually has a Pause button (undocumented )
Click on the Abort button. A dialogue will pop up asking if you want to Abort the Encoding Yes or No. Ignore This !! Load your capture program and start capturing, or whatever else you want to do. MC will sit idle in the background waiting. When you're ready to resume encoding, just click on No (I do not want to Abort the encode) and you're on your way again.
(This trick also works with TMPGenc, as long as it is not run in full-screen mode).

[mgh]

Have downloaded the 1.4 demo-haven't checked quality-But the video capture tool is superb-am capturing directly to 720x576 PAL MPEG2-audio PCM-no dropped frames-and quality is excellent. It can also capture to avi with any codec you have.

[eshangold]

Sorry if the answer to this question is obvious.

I am trying to decide the best matrix to use for my circumstances.

I am using the encoder to encode a QT reference file created in avid xpress dv to create an Mpeg-2 file for use in a DVD.

Which matrix would be best?

Any other settings I should be concerned with for my situation?

Thanks!

Ed

[sammie]

@BJ_M

yea - i had that problem also (can you say bug) , i had to enter the numbers starting in a special order to get around it and i thought nuts to that and just edited the template files directly in notepad ... you can do all the settings changes there.


Actually, it's not a bug, for some reason they have it set where you can't enter matrix numbers lower than 16 using the edit martix. You have to directly edit the template as mentioned. I remebered this after entering my last post, and haven't had a chance to get back and tell folks that.

Having someone host some templates a-la Kwag's TMPEGEnc, would be great.

[Michelle]

I'm having a problem with MCE profiles. The Profiles I save which are 1-pass are fine. If I change it to 2-pass and then save that as a profile then it doesn't work. When I select the profile it reverts to DVD. If I have a 1-pass profile selected then the 2-pass profiles don't show up. I then have to select a built-in profile to get all the saved profiles to show up.

[BJ_M]

make sure the profiles are not named the same -- you can copy your custom profile into your mc directory ...

there is a line in the profile (sometimes):

Passes=1 or Passes=2


if it isnt there --- just add it in like the following:

[MAIN]
StreamType=1
MpegType=2
VideoMode=0
AudioMode=4
Width=1920
Height=1080
DestStartX=0
DestStartY=0
DestWidth=1920
DestHeight=1080
SrcStartX=0
SrcStartY=0
SrcWidth=1920
SrcHeight=1080
FlipVertical=0
SampleRate=48000
LineFiltering=0
Deinterlacing=0
Passes=2
ID=mep-profile

[Michelle]

The profiles have different names and the 'passes' line is there.

[BJ_M]

im stumped -- you might want to email mc about that one or post the question in thier forum .. i cant duplicate the problem ..

[ronnylov]

Regarding the quantization matrices I could not input any lower values than 16 except for the upper left that should be 8 (and the help file also say that 16 is the lowest value for other values than upper left that should always be set to 8). But CCE and TMPGEnc can have values below 16 and the matrices posted in this thread had some values below 16. Is there something in the mpeg specifications that say not to use values lower than 16 or is it just maincocepts software that limits it? How did you enter values below 16? Maybe it's possible to edit the values in the template file instead of using the user interface?

I have made some testing of converting interlaced VHS tapes at high bitrates (8 mbit/s, 704x576 PAL) and I got better results from mainconcept 1.4 compared to all the other encoders I tested. Canopus Procoder was very close but did not produce such a sharp and natural looking image as mainconcept 1.4. CCE came on third place after Procoder. I tweaked the quantizer matrices to the DV matrice but used 16 where the lower values like 13 could not be entered. I set search ranges to maximum and the tweaked it even more in the advanced setting to 256 pixels horixontal and 128 pixels vertical. Not sure that it's the best settings, the encoder was as slow as Procoder but still faster than TMPGEnc.

EDIT: OK I found it in BJ_Ms post, he changed the matrice values in the template file using notepad. But I still don't understand if values below 16 is bad or not.

[Greydog]

My comments:
I have run comparisons with the MC encoder included in Premier 6.5 and TMPGEnc using the same home movie. It's true the MC encoder is much faster and does a pretty nice job, but I think TMPGEnc is the winner in quality. My movies run through TMPGEnc are 10% - 15% better in quality. I have yet to try CCE.

[BJ_M]

....................

EDIT: OK I found it in BJ_Ms post, he changed the matrice values in the template file using notepad. But I still don't understand if values below 16 is bad or not.

values below 16 are not bad but should never be used for low bit rate .. you can see the general tread in the examples as bit rate lowers -- matrice values increase ..

[xiaNaix]

What matrix/settings are recommended for Huffy compressed VHS captures?

[ronnylov]

What matrix/settings are recommended for Huffy compressed VHS captures?

I used the DV matrix (mb1 DV) on interlaced huffyuv VHS capture and it looked good! Remember to choose alternate scanning order when encoding interlaced with with this matrix.

But it also depends on your source. If it's from a VHS movie you may be able to restore the progressive frames and then you should encode non-interlaced (progressive) with zig-zag scanning order and a matrix optimized for zig-zag scan, like the "Andreas 78er" matrix. It also depends on your target bitrate.

The matrix can almost be compared with a quantization filter with low frequencies on the upper left corner and high frequencies at the bottom right corner. Some matrices with values of 99 at the bottom right corner will filter out high frequency details and make it easier to encode at low bitrates. The values are scanned in a certain order, zig-zag or the alternate scanning order. If you understand German language you may find this explanation very interesting (at least look at the pictures):
http://www.edv-tipp.de/gastbeitraege/kika001_dct.htm