HC encoder

[parcival]

Dear Hank315,
This is the first time I use HC (that is HC015) so I am not sure if I am doing something wrong. I am trying to encode a dvsd video (DVC/DV Video from a camcorder) and I am using the following script:
a = DirectShowSource("D:\road.avi")
b = a.TomsMoComp(-1,5,1)
c = b.Undot()
d=c.ConvertToYV12()
return d

I also tried:
a = DirectShowSource("D:\road.avi")
b=a.ConvertToYV12(interlaced=true)
return b

And even this:
a = DirectShowSource("D:\road.avi")
b=a.ConvertToYV12()
return b

I always get the same completely distorted result. What am I doing wrong?
If you want I can send you a dvsd sort video (2 sec only but 6.8MB size).

[Kika]

HC calculates a minimum bitrate using the values for max. and average bitrate.

I'm a bit curious: Why a user can't define the Min. Bitrate?
I'm mostly using TMPGEnc in CQ Mode, so i tried HC in CQ Mode too. But i was not able to prevent HC from choosing too low bitrates for some still or dark scenes. Guess that's why i have these strange results in my tests:
In some scenes HC did a better job, in some scenes TMPGEnc did. Surprisingly the low bitrate scenes came out better encoded by TMPGEnc.

[LigH]

@ parcival:

(Welcome to the doom9 board...)

You don't really need all those clip variables in your scripts. They are only necessary if the video processing path splits or joins. But your scripts are pretty linear. So e.g. just use

PHP Code:

DirectShowSource("D:\road.avi")
TomsMoComp(-1,5,1)
Undot()
ConvertToYV12() 


Read the documentation, which was installed together with AviSynth, about the implicit "last" variable.

BTW: I got distorted video, too -- as long as I ran AviSynth 2.56 betas. When I downgraded back to AviSynth 2.55, everything looked fine again.

[parcival]

Thanks LigH for your help!
I had to use clip variables for adding a lot of videos (140!) into a single one, something like:
a1 = DirectShowSource("D:\road 1.avi")
a2 = DirectShowSource("D:\road 2.avi")
a3 = DirectShowSource("D:\road 3.avi")
...
b = a1 + a2 + a3 +...
...

I tried to give a simple example but I forgot to remove the clip variables.
Yes, I had installed AviSynth 2.56 Beta 3 just a few days ago. I will try good old AviSynth 2.55 and get back to you.

[parcival]

Yes! AviSynth 2.55 did the trick!
Whoau!!! What else can I say about this excellent encoder?
Well done Hank315!

[johnhamler1]

hello,

for what I am concerned, the 0.14 Hc works perfectly, and give me 4.3 Gig as file output, what is new in the 0.15?

does it worth to install?.
Will try a test with the excorsist pal version.

Guys, how do you see a difference on your little screens, I watch movies from a beamer, and I can not tell any difference?


are you a kind of superman, do you see through walls too?

[hank315]

I'm a bit curious: Why a user can't define the Min. Bitrate?

Well, I tried to create an encoder with not that many settings.
And if the minimum bitrate setting is too low the result is just horrible.
But maybe one of the next versions will have this min. bitrate setting.

I'm mostly using TMPGEnc in CQ Mode, so i tried HC in CQ Mode too. But i was not able to prevent HC from choosing too low bitrates for some still or dark scenes. Guess that's why i have these strange results in my tests:
In some scenes HC did a better job, in some scenes TMPGEnc did. Surprisingly the low bitrate scenes came out better encoded by TMPGEnc.

Don't really understand this, running CQ mode means it runs with a constant quantizer, bitrate just comes out as a result how well the source can be compressed.
If there's not much to quantize, like in dark scenes, bitrate will be low.

[Encoder Master]

I'm mostly using TMPGEnc in CQ Mode, so i tried HC in CQ Mode too. But i was not able to prevent HC from choosing too low bitrates for some still or dark scenes. Guess that's why i have these strange results in my tests:

And I don't recommend the CQ mode because you have to predict and this don't give you the best quality. The best quality you get with 2pass und a perfect average Bitrate. With 2pass it looks better, in low bitrates, even TMPGEnc's mpeg-1 mode.

[jdobbs]

The quality in CQ mode will be just as good as you set it... it could be better or worse than 2-pass... you just don't know how big it will be until it's done (or you've done as sample pass in an attempt to predict it).

No doubt, though, if you have a target size you have to hit, you are almost always better off with 2 pass.

[rail grinder]

GGRRRRR

i hate hc .......... my uncle came over the weekend and was watching one of my backups and was so shocked at how good the quality was that i've now been made his official backup monkey

keep up the awesome work hank

hehee i feel like using dvdshrink or dvdxcopy on his backups just to get him off my back

[freelock7]

-Hank315
Problem with the shutdown option:
the cpu is still runing (but win xp is closed correctly).

[Kika]

@jdobbs, Encoder Master, hank315

I know how CQ works... use it since years because i get much better results with CQ than with 2pass (in TMPGEnc).

running CQ mode means it runs with a constant quantizer, bitrate just comes out as a result how well the source can be compressed.

True, but: The choice is done by the Program based of it's routines. And i think, in some cases, the choosen Bitrates are too low to get a real good result. That's why i asked for the possibility to define a Min. Bitrate. Not to set it to a very low state - it's to set it to a higher Value than HC will choose in some cases.
In TMPGEnc, there's the possibility to do this, and a well choosen Min./Max./CQ-Setting will result in a very high quality in all scenes of a Movie. With HC, i get very good results in high motions scenes, but bad results on still scenes or scenes with low contrast.

And, as i said: I don't think 2pass will give you better results, it's just time conuming.

@Encoder Master
MPEG1? This old way to encode? Oh no...

[Xeno86]

hi

I haven't been here since hc012 . Hank315 I must say that you did much progress with your encoder . The new version has better quality than 014 and 012 and is faster. However I've found some new bugs:

1. If I select "Constant Q" and "check max br" I'm not able to set max bitrate to lower than 4000, unless I unselect "Constant Q", enter a lower avg bitrate value and select "const. Q" again.

2. If I Stop the encoding in the middle using "stop" button m2v file remains locked so if I want to see the encoding I have to close HC.

3. Resource leak issue in Windows 98. In fact it only occurs if you have encoding preview turned on.

[Encoder Master]

Problem with the shutdown option:

Yes same here. Windows say that I can turn off my PC.


@Kika

For ONE-CD-Encodes encoded with TMPGEnc, MPEG-1 gives you much better quality then TMPGEnc's MPEG-2 mode. But now we have HC.

[LigH]

Originally Posted by Encoder Master

Windows say that I can turn off my PC.

This would be the usual reaction for a "Standard PC" without ACPI functionality.

Due to severe IRQ assignment problems with ACPI (6 devices sharing 1 IRQ is nonsense), I chose this mode for my system, so I was not able to tell any difference in my beta tests.

[Kika]

For ONE-CD-Encodes encoded with TMPGEnc, MPEG-1 gives you much better quality then TMPGEnc's MPEG-2 mode.

Hehe, you've never seen my MPEG2-1CD-Encodings...

OK, back to topic.

3. Resource leak issue in Windows 98. In fact it only occurs if you have encoding preview turned on.

Yes, same here. That's a bit funny. It's not realy a problem, but looks very strange.

[Encoder Master]

Hehe, you've never seen my MPEG2-1CD-Encodings...

But not with TMPGEnc, or which encoder do you use for your encodes.

[Kika]

TMPGEnc is my "Baby". I know a couple of tricks.
It's no big thing to beat MPEG1 with MPEG2 - even with TMPGEnc. But you need completely different settings to do that. If you use the same settings for MPEG1 and MPEG2 in TMPGEnc, MPEG1 might be a bit better. But only in this case.

My 1CD-Encodings are done in 352x288 anamorphic or 352x576 anamorphic (13th Warrior on a MiniDVD, 698 MByte, 192 kbps AC3 2.0 Audio from DVB-S). I never tried it with 4:3 source.

But this is a thread about HC, right? If you like, we can discuss this in an other Thread.

OK, here's something more on my wish-list for HC: Realtime display for the size of the I-, P- and B-Frames, similar to TMPGEncs log.

[hank315]

If I select "Constant Q" and "check max br" I'm not able to set max bitrate to lower than 4000, unless I unselect "Constant Q", enter a lower avg bitrate value and select "const. Q" again.

Max. bitrate just can't be smaller than average in general, will correct this for CQ encodes.

If I Stop the encoding in the middle using "stop" button m2v file remains locked so if I want to see the encoding I have to close HC.

Next version will close the m2v file if STOP is pressed.

Resource leak issue in Windows 98. In fact it only occurs if you have encoding preview turned on.

Another funny Windows issue...
Seems it doesn't appear in XP/W2K
Was this introduced in HC015?

The choice is done by the Program based of it's routines. And i think, in some cases, the choosen Bitrates are too low to get a real good result.

Sorry but I still don't get it...
Running with a constant quantizer there's nothing to choose for bitrate.
The only thing i could change is to code every P and B frame completely intra and even then it's possible the min. bitrate isn't respected.

[LigH]

A CQ encode with a minimum bitrate can be achieved - for example - by stuffing the video stream with 00 bytes up to the minimum, if the fixed quantizer results in frames way smaller than the desired minimum bitrate.

There are DVD players which have VBV overflows or similar problems if the bitrate drops e.g. below 300 kbps.