[mythtv-users] another mythencode.pl error

mojo mojospam at thegeekclub.net
Tue Jul 1 03:29:37 EDT 2003


im with ya there... a minor reduction in my mpeg4 recording settings on 
myth would basically make them pretty much a perfectly good size quality 
for a straight -oac copy -ovc copy -idx which i tried on some mpeg4 
encoded files and they do play in windows but alas they are not 
de-interlaced and they hang when i try to seek.  virtualdub wouldnt 
touch em either.. something about the audio... heh.. no matter... this 
transcoding thing should be the solution we are looking for when its done...

Sean

Rob Snow wrote:

> Well, I think we agree that the entire mencoder re-encoding process is the 
> wrong way to do it, but it's all we have currently.  What I'd really like 
> someone with about 100x more A/V streams experience to do is take the MPEG4 
> frames and MP3 out of .nuv and pop them over into an .avi container with a 
> proper .avi format.  That would make me VERY happy as we'd only be encoding 
> once and I could just make my Default profile do something that was 
> watchable AND archivable/transferable with only a container switch.  
> (320x240 at 235MB/hr or 350MB/hr for 2 or 3 hr per 700MB CD) 
>  
> Imagine a mencoder -oac copy -oav copy -makeaviproper and you see what I'm 
> wishing for, 1000fps, no additional loss of quality, mostly an I/O issue 
> rather than a CPU issue. 
>  
> Alas, it's way out of my ability. 
>  
> On Mon, 30 Jun 2003 12:00:38 -0500, mojo wrote 
> 
>>i have come really close with a few of my mencoder runs... i am  
>>pretty picky... several times i've had usable files but quality was low... 
>>its all the ones with out-of-sync audio that puzzled me... 
>> 
>>my best run was i think actually a 3-pass encode i did loosely  
>>following the mencoder docs guide for 3 pass.. had to add -idx and  
>>lavcdeint of course... cant remember if i did that on both 2nd and  
>>3rd pass.. i think i did... resulting file was well-synced but i  
>>ditched it because the bit/pixel ratio came out too low and the  
>>video was grainy when full screened on my 1280x1024 desktop... i  
>>think part of my problem is that i record straight to mpeg4 not with  
>>rtjpeg which is what i think most people are doing... 
>> 
>>Sean 
>> 
>>p.s. noticed the fix in CVS ;-) 
>> 
>>Rob Snow wrote: 
>> 
>>
>>>I agree is seems counter productive to do the deinterlace twice,  
>>>but..again..I couldn't get a clear answer on if this was required or 
> 
> not, so  
> 
>>>I erred on the side of caution and did them the same.  By the way, the  
>>>.avi's produced by this don't always work with Windows.  I have no idea 
> 
> why  
> 
>>>and it seems to happen regardless of the patch version/mencoder version  
>>>used.  
>>>  
>>>I really hope we get a -real- way to transcode soon instead of my hack.  
> 
> My  
> 
>>>wife and I like to pop a few shows on our laptops for when we travel and 
> 
> my  
> 
>>>parents like to be able to schedule to record something when they are  
>>>traveling and be able to watch it when they return.  Of course all of 
> 
> this  
> 
>>>is done on Windows boxen, hence the hack of mythencode.pl.  
>>>  
>>>-Rob  
>>>  
>>>  
>>>On Sat, 28 Jun 2003 07:02:01 -0500, mOjO wrote  
>>> 
>>>
>>>>what about the de-interlace option?  is it really necessary to   
>>>>perform that twice? could it be harmful to do that twice?  
>>>> 
>>>>the mplayer documentation is nice but if it answered all my   
>>>>questions it would be 5x longer!!  haha...  
>>>> 
>>>>mOjO  
>>>> 
>>>>p.s. i guess this transcoding stuff is coming out soon so this may   
>>>>all be a moot point...  
>>>> 
>>>>Rob Snow wrote:  
>>>> 
>>>>
>>>>
>>>>>Oops, looks like thats a bug (vpass1/2).  I'll get a patch to fix it as  
>>>
>>> 
>>>soon   
>>> 
>>>
>>>>>as we get the next patch via CVS.  Robert was away this week and was  
>>>
>>> 
>>>going   
>>> 
>>>
>>>>>to apply the patch this weekend.  
>>>>> 
>>>>>vhq/v4mv: Actually, I'm not 100% certain either way.  From reading the   
>>>>>mplayer/mencoder lists it appears you should include them in the first  
>>>
>>> 
>>>pass   
>>> 
>>>
>>>>>in some questing/replies and not in others.  I don't have a definative   
>>>>>answer, but my intuition tells me (possibly wrongly :) that you want 
> 
> the   
> 
>>>>>statistics from a pass that matches what you are going to be encoding  
>>>
>>> 
>>>with.  
>>> 
>>>
>>>>> 
>>>>>-Rob  
>>>>> 




More information about the mythtv-users mailing list