New FreeEnc 0.42

[freelock7]

29-04-2005 -FreeEnc 0.42 released! The changes:
- Now using the improved Rate Control by Peter Cheat.
- New modes: CBR, and multipass VBR (also by Peter Cheat)

http://www.downloadcounter.com/cgi-bin/download.pl?username=vmesquita&account=5
[

[guada 2]

Thank you for information freelock7,

Can you tell to me if VBR mode is at the experimental stage?

[freelock7]

No, VBR works fine with the incredible KVCD template.

Some bugs report (sent to VMesquita):

-Inversion of CQ mode and CBR (If you want to use CBR, check <CQmode>)
-CQ don't work (in fact, the CBR mode)
-VBR1Pass: in fact a CQ mode based on a quantisation=2
>5000= use the best quality
<5000= use the correct average prediction
-VBR2Pass: very good work of the RC in this mode (even under 2500kbs)


Note: the max bitrate limit can be exceeded sometimes. Don't set it too high (7000 or 7500 is better).

[guada 2]

Hello Freelock7,

If you carried out test with Freenc 0.41, perhaps you will say to me why mode CQ functions better than world VBR2 Pass.

By using IIP of Didée:
CQ = 2001 kbps
VBR2 pass:average bitrate 2002, max 2005 and Min 2000

The report is as follows: mode VBR2 pass lack really of stability, the image is not beautiful whereas the CQ maitrise the source.
Can you check my statements, because I tested several times and the report does not change.

To validate my approach I have make a test with HC and Quenc, not of particular problem.

Thank you to specify.

[freelock7]

No. <CQ> don't work in this version. If you choose the <CQ> mode in FreeEnc014, it is really a CBR mode!
Peter Cheat just tried to develop the CQ mode in NuEnc0.01b but unfortunately it was inachieved.
This mode is based on a constant quantisation -used by QuEnc too- but DVD compliant -with a max bitrate limit.
Hank has just improved the CQ mode in the 014 version now DVD compliant.
But you'll still have a great quality in 2Pass!
BTW, why these strange settings for a VBR2Pass. It looks like a CBR mode, no?
Just set a correct VBR2Pass: Avg=2000*-Max=7000-Min=0 .
*Very low bitrate. At this stage, the matrix used is very important.
(f.e. the CYDVD_OEX from Amnon)

[fluffysnurgle]

Reposted from the old FreeEnc thread:

I have downloaded the new(?) FreeEnc 0.41 but I have a problem with it.

In the past using version 0.31, using the "default" template, the encoding would ignore bitrate settings and always encode at around 1000kbps. But the problem went away by using the Incredible_KDVD template.

Now, however, the "Incredible" template is the default, and the low bitrate problem occurs with it, and there is no other choice. So the new version is not usable.

Also: Upthread it is mentioned that the maximum bitrate on rate control is not perfectly accurate. Is it better than it was before? Because in the previous version I have found the average bitrate to be very accurate, but the maximum bitrate to be essentially useless. I have to set the max bitrate to 5000-6000 and the average bitrate no higher than about 3500, or the max allowable bitrate will be exceeded. It hasn't been a big problem for me but it could stand improvement.

[guada 2]

Hello Freelock,

What I noticed on the software Freeenc 0.41:

- mode "VBR" is well "VBR"?;
- mode "CQ" is really CBR;
- mode "CBR" is "CBR"?

As regards the CQ, one is of acord on the points according to:
The CQ is not exploited by this software Freeenc 0.41, and mode "CQ" is CBR.

Nevertheless, of the reports challenge me on mode VBR:
Some tests VBR carried out:
VBR 2pass:Avg = 6500 Max = 7000 Minis = 0 Size:4.14 mo
VBR 2pass:Avg = 6000 Max = 7000 Minis = 0 Size:4.42 mo
VBR 2pass:Avg = 5000 Max = 7000 Minis = 0 Size:4.14 mo
VBR 2pass:Avg = 4500 Max = 7000 Minis = 0 Size:4.22 mo
VBR 2pass:Avg = 4000 Max = 7000 Minis = 0 Size:4.44 mo
VBR 2pass:Avg = 3500 Max = 7000 Minis = 0 Size:4.19 mo

What think you of the sizes final obtained?
What is clear, the bitrate 4000 is to be recommended.

Finally for mode "CBR", I do not include/understand very well.
What is sure:
mode "CQ": Avg = 6500 Max = 7000 Minis = 0 Size:12.1 mo
mode "CBR":Avg = 6500 Max = 7000 Minis = 0 Size:5.11 mo

I would like to have a little more precision to include/understand his operation.
I thank you for your effort.

[guada 2]

I noticed another thing concerning the encoding of the video one by script in mode VD:that relates to Quenc, HC, and Freenc, always with the script of Didée.
After captured a film of my video camera VD, I have decompress with Huffyuv to obtain a compression without loss (30,4 go);
Thus I separated using VDM the video audio and it for as well as possible exploited it video.
Then, by chance I wanted tested the script of Didée with one of the 3 software, problem:that does not function.
It is necessary to include in Converttoyv12 script at the end of the script, since the coder accepts only YV12.

CCE,Canopus procoder… accepts without condition Huffyuv avs.
Can one cure that it dear developer (Just by curiosity).

[Archimedes]

Hmm… I get better results with version 0.31 than with 0.42. I’ve tried both VBR and CBR settings with average bitrate of 6500. Source was interlaced (dv).

Josef

[guada 2]

Thank you for this new version,

There is not great change, if it is not that mode "CBR" finds its place as envisaged.

Concerning mode "VBR" even constation, in the scale of the encoding "2 pass" (progressive):
- 4000/6000/3000/2500/4500/5500/3500/5000/2000/6500 classified by size and quality pictorial.
Thus the bitrate 4000 is incontestably led the most according to "Template Incredible"

Concerning the multipass one, the things become complicated.
When one makes the choice of one "3 pass" with 3000 kbps in mode "VBR" is lower to it than the "2 pass"?
Whereas an optimized multipass with 9 is seen raised of almost 2 times the size of one "2 pass".

Which conclusion to draw from the mutipass mode, total interrogation.

Moreover mode "VBR" is quite present this time, and remains higher than mode "CBR" with final size obtained.
FreeEnc 0.42 is led definitely more in term of performance than the last version of Quenc:
with equivalent size/mode "VBR" "2 pass"/in speed and quality of image.

[freelock7]

-Gada2
Forget the freeEnc041 (wrong file size in VBR2Pass).
The new update seems to not correct it anymore.:confused:
I think it was a bad idea to use the last codes of NuEnc0.01b very instable.
I'm waiting to use a correct version of FreeEnc before to post again.

[guada 2]

Freelock7,

My tests were indeed carried out by version 0.42:

" Thank you for this New Version,
There is not great change, if it is not that mode "CBR" finds its place as envisaged. " (answer of the precedent post).

Yes, Qualities of mode VBR are significant:
- mode VBR is higher than the CBR.
- more clearness and of definition than Quenc an HC.

But there remains still unstable.

I did not know this software, I just tested it and it liked to me well by its swiftness and its qualities posted compared to other software of same order HC and Quenc.
In short, here. I await your next version to compare it with this version 0.42
Good luck and with soon. ;)

[guada 2]

Good morning to all,

News:

30-04-2005 -FreeEnc 0.43 released! The changes:
- Fixed minbitrate incompatibility with DIKO
- Fixed icon not being shown in Window
- Added a comprehensive HELP with template explanation
- Added a Peter Cheat template with NuEnc default settings

http://www.downloadcounter.com/cgi-bin/download.pl?username=vmesquita&account=5

To soon

[Pookie]

Thank you for this Very Fast encoder. Quality looks great as well. Good stuff. Keep it up.