HCEnc: Quick Questions about HC Enc

Blubear · #1 05-17-2006, 11:53 AM

Hello,

I have a couple of quick questions about HCEnc. Firstly does HC Enc use its own internal Q factor or CQ finding strategy and does this presumably work of the input of min and max bit rates etc. If this is the case what happens if the min figure is omitted? would HC Enc find its own minimum default setting perhaps? Secondly i am interested in performing a few batch encoding using this encoder since i have had some really good ones so far. However after reading the manual about modifying the .ini file i have to say i am a little confused as to which .ini this is or do i create one etc?

Lastly can someone please post an example of a batch encoding .ini setting for use of a couple to three movies using the encoder.

Thanks for the time,
Blubear

Dialhot · #2 05-17-2006, 05:16 PM

Quote:

Originally Posted by Blubear

Firstly does HC Enc use its own internal Q factor or CQ finding strategy

Yes it does. And that "Q" has nothing to do with the one you use to have in CCE (the greater the Q, the lesser the quality)

Quote:

and does this presumably work of the input of min and max bit rates etc. If this is the case what happens if the min figure is omitted?

There is no min parameter. Only the max (and the average if you want to work in 2-pass mode).

Quote:

would HC Enc find its own minimum default setting perhaps?

Kinda. The min will be the minimal bitrate needed by the less complex scene. That's how I see it but if Hank reads your post perhaps he will tell us more about it.

Quote:

However after reading the manual about modifying the .ini file i have to say i am a little confused as to which .ini this is or do i create one etc?

That is the ini that I use :

Code:

*MAXBITRATE       2500
*PROFILE          best
*ASPECT           4:3
*GOP              15 2
*CQ_MAXBITRATE    5.300
*CQ_BFACTOR       2.00
*NOSCD
*MATRIX           notch
*CPU              sse2

After this I can batch encode just by calling HC with the parameters "-i", "-o" and "-log". This suppose that all the movies I need to encode are done with the same Q factor, and the same maxbitrate. It is possible also to give the max_bitrate on the command line, but unfortunally not the Q factor (calld CQ_MAXBITRATE in the infile).

Quote:

Lastly can someone please post an example of a batch encoding .ini setting for use of a couple to three movies using the encoder.

The ini that I gave was for a 13h DVD in VCD resolution (this is to explain the max set to 2500 only).

supermule · #3 05-18-2006, 01:37 AM

I would suggest you try using DVD Rebuilder free, it has all that you are trying to do, but in a more automated and structured way.

Dialhot · #4 05-18-2006, 06:27 AM

Quote:

Originally Posted by supermule

I would suggest you try using DVD Rebuilder free, it has all that you are trying to do, but in a more automated and structured way.

Except that it works in 2-pass mode with HC, isn't it ? Or is RB-OPT adapted to HC now ?

In 2-pass mode, the results for my very low bitrate DVD (average 660 Kbit/s) was really bad compared to the CQ mode. I did not test for "normal" bitrates.

Boulder · #5 05-18-2006, 07:12 AM

The Pro version does have CQ mode for HC. I don't know how accurate the prediction is though.

supermule · #6 05-19-2006, 12:27 AM

Quote:

Originally Posted by Dialhot

Except that it works in 2-pass mode with HC, isn't it ? Or is RB-OPT adapted to HC now ?

CQ mode doesnt give an accurate DVD fill and is available in the pro ver for HC Enc.

Quote:

Originally Posted by Dialhot

In 2-pass mode, the results for my very low bitrate DVD (average 660 Kbit/s) was really bad compared to the CQ mode. I did not test for "normal" bitrates.

Which means you might not have tried AutoQmatenc by SAPSTAR, it can go where no encoder has gone before, with extremely low bitrates and has a host of other new features like matrix optimization, multi-threading and now comes bundled with the DVD-RB pro ver.

Dialhot · #7 05-19-2006, 04:28 AM

Quote:

Originally Posted by supermule

CQ mode doesnt give an accurate DVD fill and is available in the pro ver for HC Enc.

CQ is far better in quality than 2-pass and manual prediction gave me a DVD oversized by... 3 MBytes ! For a 13 hours long material predicted with a 1% sample, I have to say that people should learn how to predict first

May be a beginner luck (this was my first usage of HC), but I did 2 DVDs with such bitrate (second was undersized by 50 MB) and a third more regular one, with 2 movies (4h36 long) that came 100 Mb undersized.

BTW, I have to look at AutoQMatenc.

supermule · #8 05-19-2006, 05:37 AM

Quote:

Originally Posted by Dialhot

BTW, I have to look at AutoQMatenc.

Use the latest Beta ver since they are more updated and stabler and test it for extremely low bitrates, you'll be surprised. Quality is better with Qmatop enabled but then time taken is also more.

Oh!!! I forgot to mention one important feature, set your passes to 4 with AQMenc and it will only use as many passes as are required for best quality, skipping the next ones, intuitive

gamma · #9 05-19-2006, 11:32 AM

Quote:

Originally Posted by supermule

Quote:

Originally Posted by Dialhot

BTW, I have to look at AutoQMatenc.

Use the latest Beta ver since they are more updated and stabler and test it for extremely low bitrates, you'll be surprised. Quality is better with Qmatop enabled but then time taken is also more.

Oh!!! I forgot to mention one important feature, set your passes to 4 with AQMenc and it will only use as many passes as are required for best quality, skipping the next ones, intuitive

Hmm, don't want to be a partycrasher, but autoqmatenc isn't all that great IMO. Encodings with the KVCD notch are MUCH better at low rates. We've disussed this encoder already, the qmat function scans the video material and generates an "optimized" matrix. Sounds very good! But in my experience the notch is better

Ow an did i mention that AQE is based on libavcodec? Well, it is as far as i know. Party is over I'm afraid

Dialhot · #10 05-19-2006, 12:48 PM

Quote:

Originally Posted by gamma

Ow an did i mention that AQE is based on libavcodec? Well, it is as far as i know. Party is over I'm afraid

To be honnest, I wrote a message like this few hours ago but removed it after few minutes because I prefer to see before to judge. But I'm close to agree with you.

gamma · #11 05-19-2006, 02:09 PM

Quote:

Originally Posted by Dialhot

Quote:

Originally Posted by gamma

Ow an did i mention that AQE is based on libavcodec? Well, it is as far as i know. Party is over I'm afraid

To be honnest, I wrote a message like this few hours ago but removed it after few minutes because I prefer to see before to judge. But I'm close to agree with you.

I tested it also, along with nuenc. The release I tested of AQE wasn't all that good on low bitrates, with or without qmatop enabled.

The notch with HC was always better.

rds_correia · #12 05-20-2006, 06:53 AM

Quote:

Originally Posted by gamma

I tested it also, along with nuenc. The release I tested of AQE wasn't all that good on low bitrates, with or without qmatop enabled.

Exactly.
I ran the same tests here and even nuenc was slightly better in some scenes (mostly the lbr ones) than AutoQMat.
But my comments could be biased ones...

Quote:

Originally Posted by gamma

The notch with HC was always better.

So far this is my winner.
Plus it is free of libavcodec meaning that if something goes wrong we have to wait for Hank.
But if something goes wrong in libavcodec you could be waiting a couple of years and still they don't solve the issues...

danpos · #13 05-20-2006, 10:46 PM

@ALL

Just a commentary: it's almost pointless uses AQE (AutoQMatEnc) without uses QMatOp too because in this way any libavcodec based encoder could to do it. The differential of AQE is just the QMatOp feature. I'll quote the SAPSTAR's words from another forum:

Quote:

Originally Posted by SAPSTAR

Moreover, as you're not using QMatOp, you may as well use HCEnc for your encode....Well, that's my opinion as the little "plus" in AQE is mainly QMatOp.

With regard on quality of image output by AQE, I've to say the doing 1:1 DVD backup through DVD-RB, anyone can to get something like this. And for yours information, jdobbs has added support for AQE in the most new release of DVD-RB PRO because of the positive feedback reported by the AQE users and of tests from himself as well.

Cheers,

Dialhot · #14 05-21-2006, 07:41 AM

Quote:

Originally Posted by danpos

With regard on quality of image output by AQE, I've to say the doing 1:1 DVD backup through DVD-RB, anyone can to get something like this.

Honnestly, do you know one encoder that is not abble to obtain a good result on a 1:1 backup ? I don't.

supermule · #15 05-21-2006, 09:56 AM

Quote:

Originally Posted by gamma

Ow an did i mention that AQE is based on libavcodec? Well, it is as far as i know. Party is over I'm afraid

SAPSTAR uses certain functions in the libavcodec and not everything.Now he has been able to overcome some RC difficulties in the newer version of AQMenc(latest beta) on very high BR.

Which ver did you try, dont base your opinion on older versions since they definitely had problem. The new versions (still in Beta) are better in every sense, try them before making an opinion now.

Remember AQMenc is still in Beta, so the newer versions have bugfixes and improvements.

Dialhot · #16 05-31-2006, 06:27 PM

Quote:

Originally Posted by Boulder

The Pro version does have CQ mode for HC. I don't know how accurate the prediction is though.

I just tried and unfortunally RB does not handle the matrix setting when using HC. Prediction is done using MPEG standard matrix

(if you can submit that to jdobbs)

Edit: And that's the same with CCE !

Boulder · #17 06-01-2006, 02:07 AM

I'll check the behaviour and report it if necessary

Dialhot · #18 06-01-2006, 03:45 AM

Quote:

Originally Posted by Boulder

I'll check the behaviour and report it if necessary

Actually I should have a problem because the regular encoding (I mean, not the prediction) with CCE is also done without taking into account the matrix. Perhaps this is because I did not use one of the delivered matrix, but added a "KVCD Notch.mtx" file in the matrix directory (and selected it in the options of course). But I find this weird

Boulder · #19 06-01-2006, 03:49 AM

Now that is weird. I have the notch matrix in the matrices directory as well, it looks exactly like this (and it does work):

Code:

08 09 12 22 26 27 29 34
09 10 14 26 27 29 34 37
12 14 18 27 29 34 37 38
22 26 27 31 36 37 38 40
26 27 29 36 39 38 40 48
27 29 34 37 38 40 48 58
29 34 37 38 40 48 58 69
34 37 38 40 48 58 69 79
16 18 20 22 24 26 28 30
18 20 22 24 26 28 30 32
20 22 24 26 28 30 32 34
22 24 26 30 32 32 34 36
24 26 28 32 34 34 36 38
26 28 30 32 34 36 38 40
28 30 32 34 36 38 42 42
30 32 34 36 38 40 42 44

Dialhot · #20 06-01-2006, 04:07 AM

Same thing.
Look at my rebuilder.ini :

Code:

MainMatrix=KVCD Notch
LowMatrix=KVCD Notch
VLowMatrix=KVCD Notch
ExtraMatrix=KVCD Notch

But this is the "item.ecl" file :

Code:

; CCE SP Trial Version -- Encoder Control List
; Created by DVD Rebuilder

[item]
title=V01002800003011
aud_out=0
vaf_file=D:\DvdLab\D2VAVS\V01002800003011.vaf
aud_file=D:\DvdLab\D2VAVS\V01002800003011.mpa
file_focused=0
packet_size=2048
width=720
height=576
frame_rate_idx=3
cbr_brate=6000
vbr_brate_avg=2880
vbr_brate_min=300
vbr_brate_max=7938
seq_endcode=0
dvd=0
half_width=0
half_height=0
lum_level=0
adjust_q_matrix=0
aspect_ratio=3
gop_m=3
gop_nm=5
gop_hdr=15
seq_hdr=1
all_closed_gop=0
fix_gop_length=0
samples_per_sec=44100
stereo=2
brate_idx=7
crc=1
progressive=1
alternate_scan=0
intra_dc_prec=1
intra_dc_precision_9_max=113
intra_dc_precision_10_max=1
top_first=1
offset_line=0
aud_mode=0
tc_ref_frm=0
drop_frame=0
fix_vbv_delay=0
letter_box=0
pulldown_detect=0
create_new_vaf=1
credits_tweak=0
credits_start=0x00000
credits_brate=1000
h_filter=0
h_filter_idx=8
dither=0
dither_max=8
quality_prec=25
timecode=0x0000000
q_char_f=25
tc_offset=0
v_filter=0
v_filter_val=6
pict_name=
pict_type=3
pict_level=255
video_type=4
vid_file0=D:\DvdLab\D2VAVS\V01002800003011.m2v
vid_file1=D:\DvdLab\D2VAVS\V01002800003011.m2v
vid_out=1
vaf_out=1
opv_q_factor=17
opv_brate_min=0
opv_brate_max=7938
vbr_bias=25
vbr_pass=1
use_filter=0
filter_val=6
non_linear=1
mpeg1=0
mpeg1_cps=1
qmat_idx=0

[file]
name=D:\DvdLab\D2VAVS\V01002800003011.avs
frame_first=0
frame_last=10941
encode_first=0
encode_last=10941

No matrix there

.

I just wanted to see what PRO could do with HC. I will stick to the free version

(I'm going to test the new DVDREasy now

)