Hi guys!

Just wanted to inform you about this new encoder; HC.

Read more:
http://forum.doom9.org/showthread.ph...0&pagenumber=1

Download:
http://hank315.dyndns.org/HC_0.1.zip

I've tried it and the results are not bad at all!
In fact, they're great!

__________________
AudioSlave

Hi audioslave
Yes, I had noticed it since a few days ago.
It sure looks very much promissing.
It's still a few holes below tmpg for the moment being, quality and speed wise.
On the other hand it's still very young and most probably it has a lot of room to grow better and faster
And Hank sure has done a very nice job in Fortran95
I've sent him an email this morning to see if he wants to join us here at kvcd.net too
We're (me and Karl) waiting for an answer.
Let's keep our fingers crossed
Cheers

__________________
Rui

I just recognised that developement today on doom9
Yep looks like a new toy to test. But imho it could be that Hank wont participate in here as this forum is not free anymore.

I did a quick test with the GUI version on the first 1000 frames of a movie. The process has finished, all the button in the GUI were greyed except the "exit" button. I exited.

Now the GUI hangs when I open an avs

Is it a trialware limited to one attempt ?

Nope. No trialware
The X button doesn't respond when I hit it.
But I can use the exit button without any problem.
And I can re-use the tool as many times I want.
I just made sure that both my input avs path and my output m2v path are short and without any spaces on it.
But I am sure you also did the same
So I'm clueless of what's happening since I can't replicate it myself.
Cheers

__________________
Rui

Maybe a win2000 issue. I'll see on my winXP PC.

Hi,

I'm using W2K, and it works (encoding the 10th sample right now)...
Are you opening the same script?

Yes I did.

Note: I am the only one that had a finished encoding but with nothing enable inthe GUI but the ''exit' button ? Like if the tool was still working ?

Hi Phil
What do you mean by "nothing enabled besides the exit button"?
You mean that all other options/buttons are grayed when it finishes encoding?
If so, I must say that I have other options enabled on the GUI besides the exit button
Again, you are keeping your in/out paths short and without spaces aren't you?
Also don't use the Trim function as it gave me some hard work on the previous release 0.01 and I still haven't tested it on 0.1.
Instead I'm using the internal "cut" from HC.
Also be aware that HC uses DGDecode.dll to open a D2V source.
I am not sure if it also uses DGDecode.dll to open AVS scripts.
And I am using DGMPGDec package instead of DVD2AVI to create the D2V project file.
Would you care to install DGMPGDec on your PC and create the project file again to see if it works on HC?
I guess not, right?
If my above assumption is correct, then it is a big issue for kvcd.net users since we mainly use DVD2AVI 1.76 (or 1.77).
Keep us updated on the progress, Phil.
Cheers

__________________
Rui

Yes it is.

"D:\kvcd\test.avs".
I can do it shorter but...

I used the same.

I tested on an avi, I didn't have any DVD ripped on my PC...

... but I already use DGIndex 1.0.2 (when I do a DVD so )

You guess almost right because... I will return back to MPEG2DEC3 very soon !
The blindPP function in new DGIndex.dll does not work ! It give an avisynth crash on almost all the xvids I try to convert, making my V4 script useless. As I continue to have most of my activity in KVCD on xivd sources, I don't care to have old DVD2AVI and MPEG2DEC3.

[edit]I will use the new feature that allows you to call a specific function by using <nameoftheplugin>_<nameofthefunction>. This way I will continue to have DGIndex.dll and MPEG2DEC3 in the same directory

PS: now that I rebooted the PC, the GUI perhaps works again but... I removed the sources so I can't test

Hi all,

Just registered here today, kwag thanks for the very quick response .
First some words about the encoder, ATM the GUI is still pretty beta but I just put it online to get feedback to improve it. The "old" version which runs in a command prompt and uses an ini file to set all parameters is also still online and has the same encoding engine as the GUI version.
HC is written in Fortran95, it's the only language I feel comfortable with and still is one of the fastest. It also uses a lot of MMX/SSE stuff.
My goal was to create an encoder which would generate video streams which are DVD compliant (maybe not a big issue here).

The next GUI version will be ready in 1 - 2 weeks, a lot of the problems I read here and at Doom9 will be solved.
As a Fortran programmer I had some trouble to get used to the weird and twisted way a Windows program works.

- About the DGDecode issue, it's just the combination of DGIndex and DGDecode which will work properly, the latest beta of it (V7) is incompatible with the stable version (V6) so you cant' use DGIndex V7 with DGDecode V6. You could also use DVD2AVI projects if you rename the MPEG2DEC.dll to DGDecode.dll and place in the same directory as the encoder (or system32). Maybe I will implement a check by reading the first line of the d2v file and let it choose automatically the right decoder.
- Works only with Avisynth 2.5 for *.avs input, for an avi encode it will not need DGDecode.dll
- HC gets YUV planes (YV12 colorspace) from Avisynth or DGdecode and during encoding there are no colorplane conversions.

If there are new versions I will provide the link here also.

Releases so far:
first release http://hank315.dyndns.org/HC.zip
update: http://hank315.dyndns.org/HC_0.01.zip (exe only, *BATCH command added)
update: http://hank315.dyndns.org/HC.pdf (pdf of the manual)
update: http://hank315.dyndns.org/HC_0.1.zip GUI version (still pretty beta

And no, it's not trialware. it's freeware but closed source.

All questions/feedback are welcome.

Welcome Hank

-kwag

@Hank315
Great seeing you here Hank
I'm the one who sent you the emails inviting you to join us too.
And also the one who mixed up things about Karl's email account .
On to business.
I can't tell you how much I appreciate your work being freeware (even if closed source).
That only gives us more options besides libavcodec.
Having said that I must add that I am also loving Peter Cheat's & Nic's work.
As to comparisons (as the ones I have seen on D9) I gotta say that in low bitrate, that is below 2000Kbit, libav based encoders still got the lead in quality.
My tests were all done using "The Notch" (thanks for including it BTW) in 2-pass VBR mode.
I will perform them again and post you either a screenshot or the ini file.
I don't have much internet bandwidth but at least I'll try to post some pictures to illustrate my tests.
Stay tunned

@Karl
Could we maybe open a section dedicated to HC, or add it to an already made section like the Libav one?
TIA

Cheers all

__________________
Rui

Funny
I couldn't find any information regarding the Licensing you're using on HC.
Now I hate raising this issue, but it would be advisable to include it in the destributable zip files that you are linking.
And possibly include it in the nice PDF you're distributing too.
Just to prevent future issue, you know
Cheers

__________________
Rui

I will, but only with Hank's permission
If he says OK, then a "HC Encoder" forum section will be born

Your call Hank

-kwag

Just wanted to drop by and welcome Hank .

__________________
AudioSlave

Also noticed with very low bitrates Quenc is a little bit better if all quality options are on, so I still have some work to do
Yes, Nic and Peter did a real good job squeezing the last bits out using libav, also like Quenc and Nuenc a lot.

OK, why not

Haven't thought about that, ATM it's just freeware but will think about it.

Done

-kwag

Welcome Hank,
as you can see, here at KVCD are all well aware of any tool that can improve the quality/way we make our encodings. And your tool was also welcomed.
Be sure some/many of us will keep an eye/test your encoder. It looks really nice.
Just curious, what parameters does affect it when choosing a "encoding profile"?.
Congratulations, and welcome again.

The parameter which is most influenced by the profile choice is the way the motion vector search works.
The better the MV predictions are the less there's to encode for P and B frames which should result in a better general quality.
The profile choice is only important for the first pass, the second pass just uses the information read from the database from the first pass.