Converting MKV to DVD compliant file

[carlmart]

It's such a long time since I don't do this that I got rusty on how to.

I have an mkv file that I wish to convert to m2v so I can do a DVD from it. The best tool for that is Avisynth, of course.

The file is 1200x720 (1.667) with 23.976fps, according to Mediainfo.

From the time I did such conversions, the script I used for such files was:

FFMpegSource2("i:\file.mkv")
Spline36resize(672,480)
addborders(24,0,24,0)

I loaded this script in AvsP and they seem to load fine.

But when I load them in HCenc the frame rate is changed to 23.8092. A warning advice says there's a resolution/framerate mismatch.

What am I doing wrong?


Carlos

[poisondeathray]

what does info() report the fps is ?

some ffms2 versions get the fps slightly off

you can correct it by adding

AssumeFPS(24000,1001)

[carlmart]

I took the 23.876fps I mentioned from Mediainfo's data.

What does AssumeFPS do? I'm not familiar with that script.

[poisondeathray]

ffms2, the source filter sometimes gets the fps wrong, but the frame count is correct

assumefps changes fps without changing framecount
http://avisynth.org/mediawiki/AssumeFPS


if you use info() and preview the script, it will tell you what ffms2 "thinks" the fps is. If it's NOT 23.976, then use AssumeFPS. If it IS 23.976, then the problem is after the script (maybe HCenc) . I suspect problem is ffms2

[carlmart]

Sorry, this is the complete text report:

General
Complete name : I:\file.mkv
Format : Matroska
File size : 4.37 GiB
Duration : 1h 51mn
Overall bit rate : 5 630 Kbps
Encoded date : UTC 2011-04-12 21:06:52
Writing application : mkvmerge v4.6.0 ('Still Crazy After All These Years') built on Mar 10 2011 02:50:32
Writing library : libebml v1.2.0 + libmatroska v1.1.0

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 5 frames
Muxing mode : Container profile=Unknown@4.1
Codec ID : V_MPEG4/ISO/AVC
Duration : 1h 51mn
Bit rate : 5 070 Kbps
Nominal bit rate : 5 182 Kbps
Width : 1 200 pixels
Height : 720 pixels
Display aspect ratio : 1.667
Frame rate : 23.976 fps
Resolution : 24 bits
Colorimetry : 4:2:0
Scan type : Progressive
Bits/(Pixel*Frame) : 0.245
Stream size : 3.93 GiB (90%)
Writing library : x264 core 114 r1924 08d04a4
Encoding settings : cabac=1 / ref=5 / deblock=1:0:0 / analyse=0x3:0x133 / me=umh / subme=7 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=0 / chroma_qp_offset=-2 / threads=18 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=1 / b_bias=0 / direct=1 / weightb=1 / open_gop=0 / weightp=2 / keyint=250 / keyint_min=23 / scenecut=40 / intra_refresh=0 / rc_lookahead=40 / rc=2pass / mbtree=1 / bitrate=5182 / ratetol=1.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / cplxblur=20.0 / qblur=0.5 / ip_ratio=1.40 / aq=1:1.00

Audio
ID : 2
Format : AC-3
Format/Info : Audio Coding 3
Codec ID : A_AC3
Duration : 1h 51mn
Bit rate mode : Constant
Bit rate : 448 Kbps
Channel(s) : 1 channel
Channel positions : C
Sampling rate : 48.0 KHz
Stream size : 356 MiB (8%)

[poisondeathray]

not mediainfo , info() in avisynth


FFMpegSource2("i:\file.mkv")
Info()


It's supposed to be 23.976 , but FFMS2 is probably getting the fps incorrect

[poisondeathray]

Also fitcd is giving different resize values (656x480)

does the 1200x720 (1.667) look correct ? or did you use those resize values to compensate for something ? Is the 1200x720 cropped or does it have borders ?

[carlmart]

Thanks Poison. Now the file worked fine.

Yes, the size looks correct. It doesn't look cropped or anything. It shows black borders on the sides, as it should being slightly smaller than 16:9 (1.777).

[poisondeathray]

Yes, the size looks correct. It doesn't look cropped or anything. It shows black borders on the sides, as it should being slightly smaller than 16:9 (1.777).

but are the black borders from the source file, or from the media player using a 16:9 display ?

If the source file looks correct (circles look like circles, not ovals, no borders in the source), then fitcd says the resize should be 654x480, not 672x480

FFMpegSource2("i:\file.mkv")
AssumeFPS(24000,1001)
Spline36resize(656,480)
AddBorders(32,0,32,0)

[carlmart]

As I said before, I am a little bit rusty with conversions to DVD. My fault because it's quite necessary for my line of work (video production).

I use HCenc 0.23, and now I see that there's even a o.26 beta release.

Should I go for the 26, stick to my 23 or get the 25?

What got me curious is that the 23 only uses just one core of my CPU for processing, and perhaps it might cut time in half by using both.

The other versions use all cores?

[carlmart]

We had a similar discussion some time ago, about the size suggested by Fitcd and the one I wanted to use, based on my calculations.

That time I was right, and I did several tests to be sure about that.

Now I in the middle of the conversion, so when it's finished I will check on the results again and even do a short test with FitCD's data on a circle at the beginning of the film.

Tomorrow I will get back on that.

[poisondeathray]

Newest HCEnc is fine, should be using more than 1 core unless you have other bottlenecks

The AR depends if you're basing calculations based on 704width +/- letterbox vs. 720width . There is lot of disagreement on which one is "correct" for DVD

The 672 value would be "correct" based on 720 width assumption (32:27)
672/480 x 32/27 = ~1.659

The 656 value would be "correct" based on 704 width assumption (40:33)
656/480 x 40/33 = ~1.657

[SingSing]

Use TMPGenc to convert to mpg2 files, then use TMPGenc DVD author to generate entire DVD file sets from the mpeg2 files.

[carlmart]

Use TMPGenc to convert to mpg2 files, then use TMPGenc DVD author to generate entire DVD file sets from the mpeg2 files.

You mean TMPGenc Xpress 4.0?

I have used it, but Avisynth is a LOT better. Final quality is much closer to original HD file than with TMPGenc.

[carlmart]

The AR depends if you're basing calculations based on 704width +/- letterbox vs. 720width . There is lot of disagreement on which one is "correct" for DVD

The 672 value would be "correct" based on 720 width assumption (32:27)
672/480 x 32/27 = ~1.659

The 656 value would be "correct" based on 704 width assumption (40:33)
656/480 x 40/33 = ~1.657

My calculations were based on the 720 width.

Just now I did the test using the 656 value and compared it with former 672 value.

It's unsaid than on the 656 value, the side borders are larger, of course.

On 656 test, the circle at the beginning (Criterion) is vertically egg-shaped. In the 672 test, it's undistorted.

I can try capturing the circles for you to see.

[Gavino]

You mean TMPGenc Xpress 4.0?
I have used it, but Avisynth is a LOT better. Final quality is much closer to original HD file than with TMPGenc.

Do you mean HCEnc is a lot better?
Avisynth is a frame server, not an encoder, and in fact you can use Avisynth scripts as input to TMPGEnc.

[carlmart]

Of course: HCEnc is a lot better.

But the results I got from making the same conversion on HCenc and TMPGenc were quite different, a lot better on the former.

[poisondeathray]

I can try capturing the circles for you to see.

not necessary - if that value works, use it

on some retail content, they use 704w, on others 720w. There is no strict consistency

on paper, the DVD specs suggest DVD's should be using 704w with borders

[carlmart]

Well, I was sure about the 720 value because I had been using it and seeing circles on my plasma, and they reproduced as they should.

Particularly Criterion's and MGM's. Paramount's is a bit oval vertically.