<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
<br>
<br>
Chris Pinkham wrote:
<blockquote cite="mid200602271928.k1RJSUdP024015@panther.bc2va.org"
type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">VERBOSE(VB_IMPORTANT, QString("Original Aspect Ratio:
%1").arg(video_aspect));
</pre>
</blockquote>
</blockquote>
<pre wrap=""><!---->
</pre>
<blockquote type="cite">
<pre wrap="">I've tested this by making the modification you suggest, and when
GetVideoAspect() and SetVideoAspect() are called, 'video_aspect' is
correct, at 1.77778.
However, the resulting nuv file is still produced in 4:3 format.
</pre>
</blockquote>
<pre wrap=""><!---->
The aspect ratio is written to the new file in NuppelVideoRecorder::WriteHeader()
which is called only 5 lines below where the call to NVR::SetVideoAspect()
and nothing in between those two modifies the aspect so it should be writing out
the correct aspect ratio. You could try adding a similar VERBOSE line inside
NuppelVideoRecorder::WriteHeader() to verify the aspect ratio being written to
the file is the right one, but I don't see how it could be changed if you are
seeing it correct when we call NVR::SetVideoAspect().
</pre>
</blockquote>
<br>
I did, and it also says 1.77778, but the resulting file is still
1.33333. I also put in '2.0000' into SetVideoAspect(2.000), and it
prints 2.000, but again, the file is 1.3333.<br>
<blockquote cite="mid200602271928.k1RJSUdP024015@panther.bc2va.org"
type="cite">
<pre wrap="">
</pre>
<blockquote type="cite">
<pre wrap="">I use Media Player Classic or VideoLAN Client to view the files, and both
are capable of forcing 16:9, but detect it as 4:3.
</pre>
</blockquote>
<pre wrap=""><!---->
Are you sure that's not the dsmyth or whatever decoder you are using since
these .nuv files are Myth's enhanced nuppel format?
</pre>
</blockquote>
<br>
Well, as sure as I can be -- the problem occurs when opening through
<a class="moz-txt-link-freetext" href="myth://">myth://</a> which filters through dsmyth, and from the file directly, which
bypasses the dsmyth filter (but still uses the MythTV nuv splitter, I
think).<br>
Also, the nuvinfo tool result would seem to support my problem being
somewhere in encoding too.<br>
<blockquote cite="mid200602271928.k1RJSUdP024015@panther.bc2va.org"
type="cite">
<pre wrap="">
</pre>
<blockquote type="cite">
<pre wrap="">I also used nuvinfo to verify the aspect ratio in the file; for the
recorded mpg, it is 1.77778, for the transcoded nuv, it is 1.33333.
</pre>
</blockquote>
<pre wrap=""><!---->
Can you past the nuvinfo output here?
</pre>
</blockquote>
transcoded: info for: test.nuv<br>
<br>
aspect: 4:3<br>
aspect_f: 1.33333333333333<br>
audio_bits_per_sample: 16<br>
audio_channels: 2<br>
audio_compression_ratio: 11<br>
audio_quality: 7<br>
audio_sample_rate: 48000<br>
audio_type: LAME<br>
audioblocks: -1<br>
desiredheight: 0<br>
desiredwidth: 0<br>
finfo: MythTVVideo<br>
fps: 25<br>
height: 576<br>
keyframeadjust_offset: 0<br>
keyframedist: 30<br>
lavc_bitrate: 2500000<br>
lavc_maxqdiff: 3<br>
lavc_qmax: 15<br>
lavc_qmin: 2<br>
pimode: P<br>
rtjpeg_chroma_filter: 0<br>
rtjpeg_luma_filter: 0<br>
rtjpeg_quality: 0<br>
seektable_offset: 0<br>
textsblocks: -1<br>
version: 0.07<br>
video_type: DIVX<br>
videoblocks: -1<br>
width: 720<br>
<br>
<br>
original: info for: /video/mythtv/recordings/1001_20060219235500.mpg<br>
<br>
aspect: 16:9<br>
aspect_f: 1.77777777777778<br>
audio_bitrate: 256000<br>
audio_bits_per_sample: 16<br>
audio_channels: 2<br>
audio_sample_rate: 48000<br>
fps: 25.000<br>
height: 576<br>
is_mpeg: 1<br>
mpeg_stream_type: ts<br>
video_type: MPEG2<br>
width: 720<br>
<br>
I've also discovered some more information: if I use lossless
transcoding (MPEG2->MPEG2), the output file is okay (1.77778), but
if I use MPEG4 (what I want) or RTJpeg, it's 1.3333. Resizing or not,
with and without 'preserve aspect ratio' (I know this just means
automatically determine height), it's still 1.3333.<br>
<br>
Any further tests I can do? Sorry I'm not able to figure it out myself,
but I can continue to compile and test any changes you suggest.<br>
<br>
<br>
Ian.<br>
<br>
<br>
</body>
</html>