Experiences with HC Encoder

[TCrowe]

Originally posted by dragongodz
johnhamler1 - you should try the new HC 0.13. its both faster and has better sizing. so restore the TargetSectors back to default and see how you go.

I just upgraded to .85 RB Pro but it only has HCBatch_012.exe. Do you simply delete that and copy over the new HCBatch_013?

Thanks...

[lamster]

Originally posted by TCrowe
I just upgraded to .85 RB Pro but it only has HCBatch_012.exe. Do you simply delete that and copy over the new HCBatch_013?
No need to delete it - from Rebuilder, select Options / Setup, and change "Path to HC Encoder" to point to the version you want to use.

[TCrowe]

Originally posted by lamster
No need to delete it - from Rebuilder, select Options / Setup, and change "Path to HC Encoder" to point to the version you want to use.

Thanks lamster...

[Matt Edd]

I just used HC for the first time two days ago (after the new version) and it does as good of a job as QuEnc but is a bit faster... for me at least. Also I read somewhere that it is less buggy but I have yet to run into a bug in either encoder.

[lamster]

I had asked if the HC count-down when it finishes encoding is necessary, and dragongodz replied here (http://forum.doom9.org/showthread.php?s=&postid=642080#post642080) that "hank315 already added the *wait command in 0.12 so exiting could be faster if used."

How can I use this from Rebuilder? Shouldn't it be done by default?

[jdobbs]

It is done by default. DVD-RB Professional sets WAIT to zero. As WAIT is not a command line parameter in QuEnc, it isn't set in the freeware version. You could add it to the default INI file, though. If you use the batch version of HC with DVD-RB Pro you will get no countdown. If you are getting the countdown -- it is my guess that you aren't pointing to the encoder you think you are.

[winny]

I've tried using hcbatch instead of cce for the first time and I'm very impressed with the output. Thank you hank :D

During the encoding I experienced a Windows Dr Watson error, but after clearing this and restarting from the last successful segment it completed without further problems.

I tried a second project on a different PC and had another Dr Watson error during encoding.

Has anyone else had this? I've done a forum search but it's difficult to find the correct criteria when "hc" is too small a string.

Thanks in advance.

[DMagic1]

What is the default matrix with HC? Is it suppose to be MPEG? I thought is was HC. I did a project yesterday and it was set to MPEG.
Sorry if this was covered before.

[onesoul]

@ DMagic1

Since version HC 0.11 mpeg matrix is the default one. Check the changelog that comes with HC.

[lamster]

Originally posted by winny
I've tried using hcbatch instead of cce for the first time and I'm very impressed with the output. Thank you hank :D

During the encoding I experienced a Windows Dr Watson error, but after clearing this and restarting from the last successful segment it completed without further problems.

I tried a second project on a different PC and had another Dr Watson error during encoding.

Has anyone else had this?
Yes, I described it in a "Other MPEG1/2 encoder (http://forum.doom9.org/forumdisplay.php?s=&forumid=62)" thread here (http://forum.doom9.org/showthread.php?s=&postid=641942#post641942), and Hank replied in the new HC 0.13 encoder (http://forum.doom9.org/showthread.php?s=&threadid=93211) thread here (http://forum.doom9.org/showthread.php?s=&postid=642012#post642012).

[lamster]

Originally posted by jdobbs
It is done by default. DVD-RB Professional sets WAIT to zero.
You're absolutely correct. Once I figured out where to find the HC.INI file (in the D2VAVS directory), I saw that it had a "*WAIT 0" line.
If you are getting the countdown -- it is my guess that you aren't pointing to the encoder you think you are.
Well, I was checking to see if a crash I experienced in HCBatch_013.exe also occurred in HCBatch_012.exe, so I was using a back-level version. Perhaps support for it was added in 0.12a rather than 0.12? I guess the point is moot, since I didn't see the count-down with 0.13, but I'm not confused about what version I was running - the screen shot here (http://forum.doom9.org/showthread.php?s=&postid=641942#post641942) clearly shows "HCbatch 0.12" in the title bar, and the countdown in the "input statistics" field.

[Numer0bis]

Amazing.

I usually use DVD-RB with CCE 2.70 but tried HC013 on my Bridget Jones 2 disc and I must say the result is just amazing. I can't see any blockness and what me made wonder is why it just needed 193 minutes for the whole process. I've read that it would take normally 8 hours or so for a conversion. I don't know if this has something todo with my athlon64. I know it's a fast machine but that fast?? Well I must cleary say thank you to hank and I'll appreciate his work into making a very good mpeg -2 encoder, even if it is just version 0.13 is simply amazing for that stadium of development.

Greetz Numer0bis

[alfixdvd]

HC012 it would take probably 8 hours, but HC013 cuts the time by half.

regards
alfixdvd

[Numer0bis]

so the new version cuts the time by half and what about the quality? Is there a differnce between the versions or doesn't suffer the quality under the cut of the time?

[rendez2k]

My initial test showed much quicker time and no visable loss in quality. I'm guessing the code has just been optimized?! I'd also like to say how cool the new 'one-pass' option is! Took just over 100mins to due a complete DVD and the quality again is amazing! Fantastic work.

[ernstblaauw]

Originally posted by rendez2k
My initial test showed much quicker time and no visable loss in quality. I'm guessing the code has just been optimized?! I'd also like to say how cool the new 'one-pass' option is! Took just over 100mins to due a complete DVD and the quality again is amazing! Fantastic work.
But the bitrate is not DVD compliant. So not suitable for DVD backups.

[jdobbs]

Just to clarify, the bitrates output in 2 pass mode are DVD compliant. Unfortunately right now the one pass VBR is having a bit of a problem in that area. Hank315 is working that.

[JohnG]

But the bitrate is not DVD compliant. So not suitable for DVD backups.

Is this unequivocal for every encode? How would it be non-compliant? Are you referring to bitrate spikes above the compliancy level? If so, there are ways to check the min/avg/max levels on the backup before burning.
ernst... can you please clarify.

[jdobbs]

High demand scenes can spike above the DVD maximum of 9.8Mbs. Some (like the HBO whitenoise screen) can go much higher...

[rendez2k]

So HC as issues in OPV mode? Although I haven't tested my first DVD all the way through, would I need to, to see the problems or the DVD just won't play at all? Will the problems be present on all DVDs?

[winny]

lamster, thanks for covering the Dr Watson problem so extensively in the hc_13 thread.

For anyone else getting them, don't let it put you off using it. The end results are still superb.

[dragongodz]

johnhamler1 - you still reading this ? did you try the tests again as i suggested earlier ?

[jdobbs]

Originally posted by rendez2k
So HC as issues in OPV mode? Although I haven't tested my first DVD all the way through, would I need to, to see the problems or the DVD just won't play at all? Will the problems be present on all DVDs? From what I've seen it is only a problem on very high demand areas, like the infamous HBO whitescreen. I'd be careful using OPV with the current version -- think of it as experimental.

Just as a heads up, Hank315 has modified HC and put bitrate limitations on OPV so that it can maintain DVD compliancy... I'm testing it now -- way cool. This encoder is getting better all the time.

[Vanderlow]

Ok I just installed the freeware version Rebuilder .83 and it also installed the .13 HC encoder and I see a line in the Rebuilder.ini that references the HC Encoder, but I do not see the HC encoder in the actual menu of Rebuilder, so do I go back in the Rebuilder.ini and change the QuEnc line to point to the HC encoder? What's that HC line already in the Rebuilder.ini file doing (HC=C:\Program Files\DVD-RB\HC Encoder\HCBATCH_012a.EXE)? Thanks

[jdobbs]

The Pro Versions have an HC Path.

[jsquare]

I purchased the Magnum P.I. season 1 TV series(NTSC 4:3) 4-Disc set and wanted to back it up to DVD-R, I decided to give a try to DVD-RB since disc 1 side A holds 4 episodes(200Mins) and didn't like the results with Shrink or D2O.

First tried DVD-RB with CCE 2.67, speed was fast enough around 2.25 but end results was basically the same as some other CCE encodes, the "mosquito or halo" effect in most places remind me of the ultra-low bitrate encodes trials with D2S. So I decided to try this new HC encoder with DVD-RB.

The full disc encode took 290mins(not counting the preparation and final steps of DVD-RB), that's an speed of about 0.69 on my Dell 400SC(Win2K3) with P4HT 2.8Ghz and 1.5GB Dual Channel PC3200 RAM. End results were better than expected, but still have to watch it on my 55" TV, so far only my 19" CRT and 17" LCD at work were used for preview.

All I can say is that Hank is doing a great job with this project and jdobbs by supporting on DVD-RB.

[DMagic1]

Well I get blocks using HC that don't seems as visiable with CCE. Heres a shot from Pirates of the Caribbean. Thats seems to be the only problem I have with HC and even QUENC, the blocks in flat and dark scenes.
http://img139.echo.cx/img139/4541/untitledhc5eg.jpg

[johnhamler1]

ok, i have updated the hc encoder with the new version.
I have set the target size to a value I dont remeber when typing this.
I did a DVD, "identity", US ntsc version. and it gives me 4.30G, which is perfect!

will do more tests...

(time to encode is now around 5 hours, instead of 8, goood job!!!)

[DMagic1]

Oh, btw
It was only about 75% compression and the bitrate was pretty good. I can't remember exactly but maybe about ~3500.

[onesoul]

I also had better results with cce at bitrates 3000 - 3500kbps on progressive anime source. Maybe the tweaks such as bias and Quantization characteristics could be some of the reasons for this happening.

In the test I did before with interlaced source, hc was better (at 4400kbps).

DMagic1 Could you try doing the same encode at 4000-4500?
I am also interested in the OPV result from HC, haven't tried it myself. I suspect it will be a more close match at 1 pass vbr between HC and CCE (at least hope so).

Cheers

[jdobbs]

Wait for the next version to test OPV and HC. Version 013 can create non-compliant bitrates on high-demand scenes. Hank315 has made some changes and I just finished testing it -- it works exceptionally well... it will be supported in HC v014 and DVD-RB v0.86. Both will be out soon.

[onesoul]

Originally posted by jdobbs
From what I've seen it is only a problem on very high demand areas, like the infamous HBO whitescreen. I'd be careful using OPV with the current version -- think of it as experimental. Is the HBO white noise the same I used in my test (also interlaced)? :)

[johnhamler1]

did a second test , skycaptain & bla bla with a.jolly and her big lips, PAL.

right on the spot, 4.35 GIG like the other test I did.target size is working...

[Rockas]

Ok... I admit... just made my first experience with HC using "One Pass CQ VBR" and I'm completly "Horrorized" with the final result :D

I never thought that it would be possible to get such a great Quality from a "one pass encode" in one Free Encoder.

I have the same movie done using CCE 2.50 with 2 passes (VAF+1) and I can't see any differences at first sight...
And the final size was PERFECT = 4.583.812KB (but this was just my first test... more to come) :D

Way to go Hank!

[DK]

the 120 min title i used for the speed comparison between 012 and 013 resultet in a 4.1gb conversion with opv

using NORMAL as quality/speed setting the result is visible inferior to a 4.5gb conversion done with hc 013 and BEST setting

[Rockas]

Oh! I forgot to mention... I was using HC 0.14.

[quantum]

Originally posted by onesoul
I also had better results with cce at bitrates 3000 - 3500kbps on progressive anime source. Maybe the tweaks such as bias and Quantization characteristics could be some of the reasons for this happening.

In the test I did before with interlaced source, hc was better (at 4400kbps). I'm also finding different results depending on the source and bitrate. For me, comparing HC to CCE is getting more complicated.

[jdobbs]

Originally posted by DK
the 120 min title i used for the speed comparison between 012 and 013 resultet in a 4.1gb conversion with opv

using NORMAL as quality/speed setting the result is visible inferior to a 4.5gb conversion done with hc 013 and BEST setting You have to use v014... there were several problems with using OPV and v013...

[DK]

You have to use v014..

i acutally DID but forget to point out i used 014 in my previous post

[jdobbs]

Interesting, because every job I've done using HC v014 has been right on the money. If course it is expected for an OPV encode to be a little off -- but that seems like a lot.

By the way... a DVD-5 will only hold 4.37GB.

[hank315]

Also did some tests with DVD-RB 0.86 and HC 014, OPV mode:

LOTR - Fellowship of the Ring - PAL
final size: 4.23 GB, quantizer used: 5.5

Matrix - Revolutions - PAL
final size: 4.20 GB, quantizer used: 5.1

[DK]

By the way... a DVD-5 will only hold 4.37GB

sure thing


my results:

=> hc 014 opv: 4.166.104k
=> hc 013: 4.522.843k

[archaeo]

my results with a 100 minute title:

=> hc_14 OPV: 4.14 Gb .......260 min encoding time
=> cce_2.70 OPV: 4.17 Gb ....74 min encoding time



speed has improved in HC...

[jdobbs]

You have to divide by 1024 to get GB...

[jdobbs]

I just got an oversized test too. I'll tweak the prediction algorithm a little.

[ernstblaauw]

Hi jdobbs,

In the thread 'HC Encoder' in the forum 'Other MPEG1/2 encoder', I posted an error:

http://forum.doom9.org/showthread.php?s=&postid=645386#post645386

Hopefully you can address it (or hank315). I do not know the origin of the fault, so I posted the problem in that thread.

[jdobbs]

I think you've nailed it. The output string to HC's INI file should always have a "." --- not a ",". I used a formatting function that adjusts for different notation standards, and that isn't appropriate for that file. This will probably make playing havoc in non-English speaking countries that use the comma as the decimal notation rather than a period...

I'll fix it and release a new version today.

[Rockas]

That's funny... In the test I made (see a few replies above) it gave me a final result that was Perfect (Lucky I guess) :D

[johnhamler1]

i have not tried yet the 0.14 will do this later.


this software works fine now, much better than a month ago. so my question now is: is it possible to make more than 2 pass, 3 or even 4 pass???.
with 2 pass, it takes 4-5 hours on my 1.7pentium M.
most people I guess are working or sleeping when encoding, so 1-2 more pass, will not change my sleep time!


will encode tomb rider PAL version tonight....u dont care? I do :)
(added: encoded tom rider in 4.5 hours best quality 2 pass, 4.31 gig...perfect)

BTW,someone can explain me in 2-3 lines what a matrix is? how it works?

I guess there is a matrix for cartoon like "Incredibles" or "searching Nemo", what do u suggest?

[johnhamler1]

???