HC018 released

[hank315]

A new release of HCenc is out: HC018

Changelog:
- updated encoding engine
- added pulldown 23.976 --> 29.97
- added adaptive matrix
- GUI bug fixes

Get HCenc at www.bitburners.com

[video_magic]

:thanks:

[Eastermeyer]

Is it possible to remove the Resolution Limit ?

[buzzqw]

:thanks::thanks::thanks:

your encoder is simple great !

BHH

[dirio49]

thanks.
:thanks: :thanks:

[Fishman0919]

WHOA :cool:

[Adub]

Great man! Oh, I can't wait to test this out!

Thanks a bunch! You rule!

[dragongodz]

Is it possible to remove the Resolution Limit ?
this was already asked in the last thread, i think it was in the last couple of pages.

short answer is not at the moment. for more complex answer go read the other thread.

[Audionut]

Thanks!

[rhys87]

my favourite mpeg2 encoder, improves all the time and free how good can it get :)

[ljpp]

The sub-domain http://hcencoder.bitburners.com will get you to the HC homepage slightly quicker.

[FredThompson]

- updated encoding engine


Updated in what way? Would you give a brief explanation?

[AVIL]

Hi,

I've tested this new version against v017.

d2v files of newest version of DGindex are rejected (max version 12 like in 017).

Slightly slower (28 fps vs 30) althought need more tests.

---------- Ignore this part ----------------------------------
Final size less accurate than version 017. In my case final size requested of 1048424 KB, obtained file of 958264 KB.
------------------------------------------------------------

Encoding size is as accurate as with 017, simply the output file isn't complete. Sorry.

Best regards

[ljpp]

Updated in what way? Would you give a brief explanation?

Here is what Hank told me:

The major change is in the encoding engine itself, it's really a lot better.
Did a lot of tests, visual and metrics, on metrics only Mencoder is better... (at 8 fps)

So I'm curious about the comments.

[dragongodz]

Slightly slower (28 fps vs 30) althought need more tests.
that should be expected. lets just say the the changes are for extra quality and that of course generally means slower.

[hank315]

Updated in what way? Would you give a brief explanation?It has a better RD optimization for P-frames, also B-frames will benefit from that because most B-frames are predicted from the P-frames.
Also some changes to the quantization routines and a lot of other tweaks

On my PC best profile also runs a bit slower, normal and fast profile runs a bit faster than 017.

And about this:
The major change is in the encoding engine itself, it's really a lot better.
Did a lot of tests, visual and metrics, on metrics only Mencoder is better... (at 8 fps)
I will never comment about visual quality, everybody has his own opinion about it, you can't measure it...

But the metric tests I did shows an improvement over HC017, especially with lower bitrates:
In the "Challenge Low Bitrate" by Sagittaire I got the next values:
HC017, OPSNR 46.71, SSIM2 81.38
HC018, OPSNR 46.97, SSIM2 83.04

using best profile and the standard MPEG matrix

[Sounddude]

Thanks for the update. I just started a new encode, but had to stop it and use the new version.

Keep up the great work!!

[linx05]

I've added CCEAQM=1 into the [Options] of DVD-RB. Does that make much of a difference? I've read a couple threads where it has but it's still a bit hazy in my mind.

[Mug Funky]

sweet! i will try this out right away :)

thanks for your work. RD optimisation with such a small speed decrease sounds like a very good thing.

[edit]

just looked at the release date... my birthday :) damn nice present...

[Pasqui]

Thanks hank315 for your updated encoder ! I get great results using it with DVD-RB Pro.
BTW, will you provide an adaptative matrix dll in the future ? Currently for my backups I switch between MPEG matrix, AutoQ2 matrix and QLB matrix according to bitrate (DVD-RB Pro allows it)

[Ebobtron]

Well my new system likes it. Wow, real time encoding.

Sorry Hank, I do not have a bunch of "can you"s, just a bunch of "thank you"s.

:thanks: :thanks: :thanks:

This time.

[FredThompson]

I've just run some complete jobs with DVDRBPro. This encoder sure looks nice. Up next, the torture test: WWE interlaced high-motion source.

[rhys87]

yeah it does undersize a little bit, but everything else works fine in it.

[Fishman0919]

yeah it does undersize a little bit, but everything else works fine in it.


?!?!?!?, undersize how? I don't think I've ever had HC (any ver) under size anything that would be something to mention..... ex. avg bitrate 3500k and HC comes in at 3498k.

[Mug Funky]

i noticed that the matrix editor seems a bit off...

try backspace over a 1 digit value (or hit it twice on a 2-digit value) and see what happens. it crashes on the P4 Opteron (hyperthreaded) i tried it on. big visual fortran error, then closes.

i also noticed a crash when using the GUI to set up a half-d1 encode (not sure that's the problem). switching from "main" to "advanced" would bring down the program if i change the matrix (changed from the custom one that was still loaded to the internal default one).

not sure if this info helps you reproduce, but don't be worried - i mostly use the batch encoder as most of my settings don't change and i only have to set bitrates and aspect.

[hank315]

try backspace over a 1 digit value (or hit it twice on a 2-digit value) and see what happens. it crashes on the P4 Opteron (hyperthreaded) i tried it on. big visual fortran error, then closes.Yep, that produces a nice crash...
Already fixed it for the next release.

i also noticed a crash when using the GUI to set up a half-d1 encode (not sure that's the problem). switching from "main" to "advanced" would bring down the program if i change the matrix (changed from the custom one that was still loaded to the internal default one).Can't replicate this one...

[rhys87]

?!?!?!?, undersize how? I don't think I've ever had HC (any ver) under size anything that would be something to mention..... ex. avg bitrate 3500k and HC comes in at 3498k.

I'm just being very critical because to me there isnt much else wrong with it. When i've compiled the dvd with previous versions of HC the end result is about 4.3gb in dvdlab with the latest its about 4.15-4.2gb.

[Mug Funky]

oh, and while on the subject of custom matrices (i forgot to mention this the other day, but it is important), i got black blocks in (pure 235) white areas on decoding a stream encoded with a "half mpeg standard" matrix, that is all the values from the "standard" matrix divided by 2, rounded down.

i'm not sure if this is a new thing because i actually haven't used HC with custom matrices before. there are no values below 8 in this matrix, so i don't think that's a problem (apparently that used to be a problem in xvid with trellis or something?).

the black blocks would only happen in the centre of objects, even though there were large areas of the same colour around them. it was a flat animation as source - pure CG, exported from Flash.

[Riuyin]

My favorite encoder. Any one knows if there is a way to load an kdvd matrix in HC, I saw some KVCD's matrices but not an sigle KDVD matrix.

[kevin23]

My favorite encoder. Any one knows if there is a way to load an kdvd matrix in HC, I saw some KVCD's matrices but not an sigle KDVD matrix.

The matrix is the same.Load the Kvcd Notch Matrix provided in the matrix folder.

[Riuyin]

The matrix is the same.Load the Kvcd Notch Matrix provided in the matrix folder.

Your are right both matrices are the same.

But the template in tmpgenc also change the GOP structure like this:

Number of I pictures in GOP 1
Number of P pictures in GOP 5823
Number of B pictures in GOP 2
Output of interval iof sequence header: 1
Max number of frames in a GOP: 18

Correct me if i am wrong

Hc..............................................................Tmpgenc
GOP length.........................=...............Max number of frames in a GOP
B frames............................=...............Number of B pictures in GOP

But what about Number of I and B pictures in GOP, where can I change that? or it doesn't matter.

[tamahome]

would it be possible to add a possibility of regulate the priority of the process of hcencoder?

a little like quenc low priority has high priority

[Sagittaire]

Your are right both matrices are the same.

But the template in tmpgenc also change the GOP structure like this:

Number of I pictures in GOP 1
Number of P pictures in GOP 5823
Number of B pictures in GOP 2
Output of interval iof sequence header: 1
Max number of frames in a GOP: 18

Correct me if i am wrong

Hc..............................................................Tmpgenc
GOP length.........................=...............Max number of frames in a GOP
B frames............................=...............Number of B pictures in GOP

But what about Number of I and B pictures in GOP, where can I change that? or it doesn't matter.

For MPEG2 PAL/NTSC DVD compliant stream max GOP must be 15/18 frames (can be adaptative with change scene detection). max bframe must be 2 (can be adaptative too). TMPGEnc template is not false because max GOP has the priority and choose 5823 Pframe is simply useless here.

[Chainmax]

I've just run some complete jobs with DVDRBPro. This encoder sure looks nice. Up next, the torture test: WWE interlaced high-motion source.

That sounds very interesting, please do tell how it does.


Does HC support multicore CPUs (real one like core duo or Athlon X2, not HT P4s)? If not, could it be introduced in the future?

[Mug Funky]

i just noticed that in CQ mode, if HC raises the quant to prevent underflow, it'll lower it in the next second or so to keep the average. this looks like a big step on the way to 1-pass RC... all that remains is to figure a good guess at initial quantizer and keep a running average (i'm thinking it sounds far simpler than it is :))

good stuff. i like.

also, the RDO saves huge amounts of bitrate for the same constant quant on clean sources. very cool indeed.

[Revgen]

Does HC support multicore CPUs (real one like core duo or Athlon X2, not HT P4s)? If not, could it be introduced in the future?

Darksoul71 just released a frontend for Quenc that uses Dual-Core and HT procs. I was able to record a Hi-Q interlaced basketball game (interlaced from a 720p source) at near full-speed for each pass. Hopefully he can create one for HC in the future. You can read about it here. (http://forum.doom9.org/showthread.php?t=110938)

[ricoman]

I've used the combo of HC with DVDRebuilder (free versions) a couple of times with good results. However, I just tried to backup Munich and ran into trouble. I ripped the entire uncompressed dvd to my HD and I can play it like that, but when I processed it using HC/rebuilder, it seemed to process normally but when I play the processed file from the computer or new dvd the video is screwed up. It comes out as a bunch of diagonal lines though you can tell the movie is in there between those lines somewhere and the audio is fine. Anyone have this problem? and what can I do about it?
Is there a way to get a message to Hank? This was my first try with 18.

[EpheMeroN]

DC Precision of 9 = recommended for normal bitrate

DC Precision of 10 = recommended for high bitrate

What falls under the category of normal / high bitrates?

Right now I have some videos that I am using 4,000 as avg bitrate and 9,000 as max.

[AVIL]

@ricoman

Your problem sound equal to those of planar alignment (method to arrange planes in the YV12 format). Try to install last version of avisynth.

Good luck

[ricoman]

I'm using avisyth 2.56. isn't that the latest? Any other Ideas? Hank?

[feedback]

Does this problem occur on just this movie "Munich" or does it occur on any other movie you backup with DVD-RB & HC18?

Btw, this line of questioning should be broken off by a Mod and moved to the Rebuilder forum IMO as it is not strictly an HC encoder problem.

Regards,

[ricoman]

Munich is the first I tried with HC18. I've been successful with the previous version on King Kong and Narnia. I believe it is an HC problem because I just finished processing the same Munich file on my HD with Quenc, and it is fine.

[Chainmax]

Darksoul71 just released a frontend for Quenc that uses Dual-Core and HT procs. I was able to record a Hi-Q interlaced basketball game (interlaced from a 720p source) at near full-speed for each pass. Hopefully he can create one for HC in the future. You can read about it here. (http://forum.doom9.org/showthread.php?t=110938)

That sounds great :).

[Mug Funky]

@ ricoman:

try add "turnleft().turnright()" to the end of your script.

if that doesn't work, regress your avisynth to 2.5.5 or lower.

i've had the alignment prob before, but only when using avs plugins that don't support the correct chroma alignment (old avisynth wasn't as picky about it, but don't worry - it's all for the best, we just have to wait for plugin devs to change their code a little).

[hank315]

@ricoman
Please could you try the following:
Restart the encoding with DVD-RB, turn on the preview window and report if the video is already messed up during the first pass.
If so then it's getting wrong input, if only the second pass looks bad it's a HC error.

[EpheMeroN]

I have another question about the matrices in HC Encoder.

Do any of them handle macroblocking better than the others? And are any of them better-suited for high motion video such as sporting events?

I've only been using the default MPEG Matrix.

[Darksoul71]

@revgen and Chainmax:
Darksoul71 just released a frontend for Quenc that uses Dual-Core and HT procs. I was able to record a Hi-Q interlaced basketball game (interlaced from a 720p source) at near full-speed for each pass. Hopefully he can create one for HC in the future.
Honestly I´ve been already thinking about such a "middleware" for HCEnc. Esp. because of the results you´ve posted revgen. It shouldn´t take too much coding to adopt the approach from QuEnc to HCEnc.

Currently I still have to dish out some bugs on LAN_AQE-RB (network distributed encoding with DVD-RB and AutoMatQEnc) but once this beast is working, I could focus on HCEnc^n :)

[ricoman]

@ricoman
Please could you try the following:
Restart the encoding with DVD-RB, turn on the preview window and report if the video is already messed up during the first pass.
If so then it's getting wrong input, if only the second pass looks bad it's a HC error.
The preview window looked OK on both the 1st & 2nd pass. But, I noticed that the gui said version 17, even though I have downloaded 18, and 17 had worked before.

[feedback]

Well which is it v.17 or v.18 you are using? You obviously have your wires crossed somewhere.
You, also need to, early on, give more relevent info. like indicating that Munich encodes correctly with Quenc. from the start. If you had shared that little tidbit I wouldn't have suggested moving the thread.

People here will help you but, we aren't mind readers.

@EpheMeroN
Have you tried the BDVD Matrix the QLB Matrix or the MainConcept Matrix? Give those a try...are you using HC in conjunction with DVD-RB? If you are using RB you can try the different Matrices on a short segment with Rebuilder's Matrix Editor. You may also want to drop your DC precision to 8.

Regards,

[Revgen]

@revgen and Chainmax:

Honestly I´ve been already thinking about such a "middleware" for HCEnc. Esp. because of the results you´ve posted revgen. It shouldn´t take too much coding to adopt the approach from QuEnc to HCEnc.

Currently I still have to dish out some bugs on LAN_AQE-RB (network distributed encoding with DVD-RB and AutoMatQEnc) but once this beast is working, I could focus on HCEnc^n :)

Sounds great. Good luck with it.:thanks: