Closed GoogleCodeExporter closed 8 years ago
On Youtube the web page contains a value called 'fmt_url_map' which contains
three
(encoded) urls
Each of those urls point to an actual file on YouTube's file servers in either
HD, HQ
or Regular quality
The first one is usually the highest quality and this is the one that EMM takes.
It will then check the type of content reported back from the server for the
url -
flv, mp4, shockwave, flash and save the file accordingly.
Can you post a link to the Youtube trailer you were trying to download?
Original comment by rogue.da...@gmail.com
on 15 Dec 2009 at 5:10
example:
http://www.youtube.com/watch?v=a1I7QQcXokY&hd=1
EMM downloads a ~5.8MB .flv, as opposed to a full 38.6MB HD trailer given by a
ripping service such as: http://keephd.com/watch?v=a1I7QQcXokY&hd=1
Original comment by shutchin...@gmail.com
on 15 Dec 2009 at 5:50
Yes, this is a known issue.
Rogue, this is a big pain for me too, and that makes trailer downloader almost
useless. (Actually EMM almost never downloads HD quality trailer, so there
should be
an other trick to find the right URL, than simlpy taking the first reported)
There was a change some month ago to prevent scraping and they did a really
good job.
I was not even able to find a site which can scrape HD trailers by that time.
Hmmmm,
time has passed and it seems there is already at minimum one site.
Nice find shutchinson!
However, I am not sure how can we resolve this. Simpliest (but dirtiest) would
be to
have a separate download field for HD trailers and use keephd to download...
Rogue, or can you figure out the trick needed?
Original comment by nadas.bence@gmail.com
on 15 Dec 2009 at 6:14
Well one thing keephd can do is simply replace youtube in the URL. So
technically,
you could parse the youtube link and replace with keephd.
There was a site I came across last week with an API.
Original comment by shutchin...@gmail.com
on 15 Dec 2009 at 6:37
keeptube.com is another one whereby you just throw the youtube link at the end
e.g.
http://keep-tube.com/?url=http%3A//www.youtube.com/watch%3Fv%3Da1I7QQcXokY%26hd%
3D1
Original comment by shutchin...@gmail.com
on 15 Dec 2009 at 6:45
Fixed and verified! Excellent! It is perfect as it is. Thank you Rogue!
Could you please investigate if you can use your new class in autoscrapers as
well?
Original comment by nadas.bence@gmail.com
on 19 Dec 2009 at 11:26
Good job guys. ETA for release?
Also, will EMMs auto search youtube for trailer only search for HD and download
the mp4?
Original comment by shutchin...@gmail.com
on 19 Dec 2009 at 11:58
No, EMM is looking at IMDB and TheMovieDB for trailers (according to your
choice in
settings). It is only downloading from youtube, if there is a youtube url
linked in
TheMovieDB. And this only going to be HD in case the linked url in TheMovieDB is
pointing to a HD version.
We cannot scrape youtube directly. It would be possible technically, but EMM
couldn't
identify the right movie from the search list.
The only real solution would be to find a HD trailer site, which can searched
by IMDB
id and we can scrape it. However I think there is no such site exist today.
Original comment by nadas.bence@gmail.com
on 20 Dec 2009 at 8:34
Right, I understand.
Would be a good idea for a site - an HD movie trailer database. I'm not sure of
the
legal issues involved with that though. Apple Trailers is pretty much the
closest
that you could get.
How about if I build a database that contains 2 fields:
IMDB ID
Youtube HD Link
We can populate manually and also allow crowdsourcing. Thoughts?
Original comment by shutchin...@gmail.com
on 20 Dec 2009 at 7:17
rogue added youtube class to scrapers in r1084.
Original comment by jason.schnitzler
on 22 Dec 2009 at 3:24
Thanks a ton rogue! When can we expect a compiled version if r1084+
Original comment by shutchin...@gmail.com
on 22 Dec 2009 at 3:32
Original issue reported on code.google.com by
shutchin...@gmail.com
on 14 Dec 2009 at 11:45