[Ffmpeg-devel-irc] ffmpeg-devel.log.20121006

[burek]

[00:01] <cone-429> ffmpeg.git 3Stefano Sabatini 7Changelog doc/ffprobe.texi ffprobe.c: ffprobe: add -select_streams option
[00:01] <cone-429> ffmpeg.git 3Stefano Sabatini 7ffprobe.c: ffprobe: reindent after previous commit
[00:06] <cone-429> ffmpeg.git 3Stefano Sabatini 7Changelog: Changelog: add empty line after version lines
[00:20] <Daemon404> durandal_1707, pelase refrain from giving such useless reviews
[00:21] <durandal_1707> Daemon404: not useless
[00:21] <Daemon404> "I do not like this."
[00:21] <Daemon404> useless
[00:21] <Daemon404> as
[00:21] <Daemon404> fuck
[00:21] <durandal_1707> _not_ useless
[00:22] <Daemon404> how
[00:22] <Daemon404> on earth is that useful
[00:22] <Daemon404> do explain
[00:22] <Daemon404> im all ears
[00:22] <Daemon404> maybe he should use telepathy
[00:22] <Daemon404> to figure out whats wrong
[00:22] <Daemon404> and why you dont like it
[00:23] <JEEB> "I do not like this." <- always try to give at least some kind of a reasoning
[00:23] <JEEB> otherwise it's as useless as a spit on a face
[00:23] <JEEB> and leads to "OK, so he doesn't like it. How can I fix it?"
[00:24] <JEEB> instead of "OK, X could be done better."
[00:24] <durandal_1707> people are just ignorant
[00:24] <Daemon404> it soudns mroe liek youre jus ta ragain asshole
[00:24] <Daemon404> let me restate that without typos:
[00:24] <Daemon404> youre an asshole.
[00:24] <burek> it's never boring here :)
[00:24] <JEEB> durandal_1707, it might be clear to you why something's bad
[00:25] <JEEB> but it might not be for the other person
[00:25] <JEEB> tl;dr explain it
[00:25] <durandal_1707> Daemon404: thank you for so kind words
[00:25] <Daemon404> im doing my paul b mahol impression
[00:25] <Daemon404> i hope you like it
[00:26] <Daemon404> (but if you dont, feel free to say so. but not why.)
[00:28] <saste> Daemon404, moderate terms
[00:28] <saste> you made your point, no need to give names
[00:28] <saste> blame the attitude, not the person
[00:28] <Daemon404> saste, youre right
[00:34] <gnafu> "<@Daemon404> it soudns mroe liek youre jus ta ragain asshole" <-- I like how you only spelled one word right, and it was the most important word.  It kinds adds emphasis in its own, "he must be drunk or something" kinda way ;D.
[00:34] <nevcairiel> "it" is also spelled correctly!
[00:34] Action: Daemon404 looks at the beer on his desk
[00:34] <gnafu> nevcairiel: Yeah, but that's too short to count :-D.
[00:35] <Daemon404> i spell 'it' wrong all the time
[00:35] <nevcairiel> hey he screwed up shorter words in that sentence
[00:35] <Daemon404> as 'i'
[00:35] <gnafu> Daemon404: I had a friend from Canada who only typed correctly when he /was/ drunk.
[00:35] <Daemon404> we call them newfies
[00:35] <gnafu> XD
[00:38] <saste> michaelni, I need something like 576 words for xface bigint
[00:38] <saste> lavu/integer supports 8 words
[00:39] <saste> should I extend the number, make it configurable or use a custom biginter implementation?
[00:39] <saste> second option will break compatibility, but not that this would be a problem, since possibly nobody is (still) using it
[00:40] <saste> ah and the author of the code (~1990) agreed to relicense the code
[00:41] <Daemon404> this reminds me
[00:41] <Daemon404> we can ditch x11grab
[00:41] <Daemon404> vlc has xcbgrab
[00:41] <Daemon404> whic his lgpl
[00:41] <Daemon404> and also uses xck
[00:41] <Daemon404> xdc*
[00:41] <saste> Daemon404, patches welcome :)
[00:41] <Daemon404> saste, that requires me to touch stuff related to X
[00:41] <Daemon404> never. again.
[00:42] <saste> i somewhat envy mplayer for x11 output
[00:42] <saste> from time to time, people tries to reimplement X from scratch
[00:42] <saste> i recall a nice project, it was called "Y"
[00:43] <Daemon404> wayland sounds like it may become something
[00:43] <Daemon404> its got people like keithp on it
[00:46] <gnafu> It definitely has a lot of people and groups behind it.
[00:46] <gnafu> It'll be interesting to see what happens.
[00:46] <gnafu> I probably won't start using it until Fedora switches to it as the default (if that ever happens, which I would think it will).
[00:47] <gnafu> I'm too lazy to do otherwise.
[00:56] <michaelni> saste, what exactly is "bigint" needed for ?
[00:56] <saste> it's used to store the image as a bit integer
[00:56] <saste> image is a 48x48 bitmap
[00:57] <saste> uhm no, wait when i'll be awake
[00:57] <saste> but anyway, i may need to extend the number
[00:58] <saste> but seems i can't count now
[03:21] <cone-532> ffmpeg.git 3Michael Niedermayer 7libavcodec/ffv1.c: ffv1: remove commented asserts
[03:21] <cone-532> ffmpeg.git 3Michael Niedermayer 7libavcodec/ffv1.c: ffv1: change w/h asserts to check as the condition can likely happen
[03:33] <cone-532> ffmpeg.git 3Michael Niedermayer 7libavcodec/ffv1.c: ffv1enc: fix assert in put_vlc_symbol() and update to av_assert2()
[05:08] <cone-532> ffmpeg.git 3Michael Niedermayer 7libavcodec/arm/mpegvideo_armv5te.c: mpegvideo_armv5te: change asserts to av_asserts
[06:21] <cone-532> ffmpeg.git 3Carl Eugen Hoyos 7Changelog doc/general.texi libavcodec/Makefile libavcodec/allcodecs.c libavcodec/avcodec.h libavcodec/codec_desc.c libavcodec/targa_y216dec.c libavcodec/version.h libavformat/isom.c: Pinnacle TARGA CineWave YUV16 decoder (fourcc Y216).
[12:15] <cone-191> ffmpeg.git 3Michael Niedermayer 7libavcodec/libvorbisenc.c: libvorbisenc: fix afq delay setting
[12:15] <cone-191> ffmpeg.git 3Michael Niedermayer 7libavformat/segment.c: segment: reverse pass avoid_negative_ts from the muxer
[12:53] <cone-191> ffmpeg.git 3Stefano Sabatini 7doc/filters.texi: doc/filters: fix typo
[13:57] <cone-191> ffmpeg.git 3Anton Khirnov 7libavcodec/avcodec.h libavcodec/resample.c libavcodec/resample2.c libavcodec/version.h: lavc: deprecate the audio resampling API.
[13:57] <cone-191> ffmpeg.git 3Anton Khirnov 7avplay.c libavfilter/af_asyncts.c libavfilter/af_resample.c libavresample/audio_data.c libavresample/audio_data.h libavresample/avresample-test.c libavresample/avresample.h libavresample/utils.c: lavr: change the type of the data buffers to uint8_t**.
[13:57] <cone-191> ffmpeg.git 3Anton Khirnov 7doc/APIchanges libavresample/version.h: lavr: bump major to 1 and declare it stable.
[13:57] <cone-191> ffmpeg.git 3Anton Khirnov 7doc/APIchanges: doc/APIchanges: fill in missing dates and hashes.
[13:57] <cone-191> ffmpeg.git 3Dmitry Samonenko 7libavcodec/libspeexenc.c: libspeexenc: Add an option for enabling DTX
[13:57] <cone-191> ffmpeg.git 3Dmitry Samonenko 7libavcodec/libspeexenc.c: libspeexenc: Updated commentary to reflect recent changes
[13:57] <cone-191> ffmpeg.git 3Mans Rullgard 7libavcodec/x86/dsputil_mmx.c: x86: dsputil: kill VLA in gmc_mmx()
[13:57] <cone-191> ffmpeg.git 3Mans Rullgard 7libavcodec/ppc/fmtconvert_altivec.c: ppc: fmtconvert: kill VLA in float_to_int16_interleave_altivec()
[13:57] <cone-191> ffmpeg.git 3Mans Rullgard 7libswscale/ppc/swscale_altivec.c: ppc: swscale: rework yuv2planeX_altivec()
[13:57] <cone-191> ffmpeg.git 3Mans Rullgard 7configure: build: error on variable-length arrays
[13:57] <cone-191> ffmpeg.git 3Anton Khirnov 7libavfilter/vf_pad.c libavfilter/vf_scale.c: vf_pad/scale: use double precision for aspect ratios.
[13:57] <cone-191> ffmpeg.git 3Diego Biurrun 7libavutil/parseutils.c: parseutils-test: Drop random colors from parsing test
[13:57] <cone-191> ffmpeg.git 3Diego Biurrun 7tests/fate/libavutil.mak tests/ref/fate/parseutils: fate: Add parseutils test
[13:57] <cone-191> ffmpeg.git 3Diego Biurrun 7: Give all anonymously typedeffed structs in headers a name
[13:57] <cone-191> ffmpeg.git 3Anton Khirnov 7libavformat/yuv4mpeg.c: yuv4mpeg: return proper error codes.
[13:57] <cone-191> ffmpeg.git 3Michael Niedermayer 7: Merge remote-tracking branch 'qatar/master'
[13:57] <cone-191> ffmpeg.git 3Michael Niedermayer 7libswresample/swresample-test.c: swr-test: avoid VLA
[14:07] <cone-191> ffmpeg.git 3Michael Niedermayer 7libavcodec/snowenc.c: snowenc: get rid of VLA (well it wasnt really variable anyway)
[14:07] <cone-191> ffmpeg.git 3Michael Niedermayer 7libavutil/pca.c: pca: get rid of VLA
[14:07] <cone-191> ffmpeg.git 3Michael Niedermayer 7configure: configure: enable  -Werror=vla
[15:02] <siretart> ubitux: yes, debian/stable still ships 0.5
[18:38] <cone-191> ffmpeg.git 3Bobby Bingham 7libavfilter/libmpcodecs/vf_tile.c: vf_tile: fix typos/grammar in comments
[18:42] <ubitux> do we still need this filter?
[20:23] <cone-191> ffmpeg.git 3Michael Niedermayer 7libavformat/movenc.c: movenc: fix regression with yuyv caused by c5f23d
[21:32] <saste> what's so awesome about sed??
[21:33] <ubitux> tetris.sed?
[21:33] <ubitux> arkanoid.sed?
[21:33] <Daemon404> nothing
[21:33] <Daemon404> perl > sed
[21:34] <saste> i mean what's so awesome about s/PIX_FMT/AV_PIX_FMT/
[21:35] <saste> apart from annoying michael at the next merge
[21:35] <Daemon404> <insert rehtoric about poluting the namespace)
[21:35] <Daemon404> ironically, when we added AV_ to the codec ids
[21:35] <Daemon404> they added a #define CodecID
[21:36] <Daemon404> which i had to udnefine in my downstream project
[21:36] <saste> yes, but what's so awesome about cleanup?
[21:36] <Daemon404> ... because they pollutedt he namespace
[21:36] <saste> it's necessary, but is boring
[21:36] <Daemon404> [15:36] <@saste> yes, but what's so awesome about cleanup? <-- as a downstream user, clean up is very welcome and i appreciate it
[21:36] <Daemon404> since libav* libs have always been catastrophic messes
[21:37] <saste> Daemon404, libav* mess is overrated
[21:37] <Daemon404> i disagree
[21:37] <saste> cleaner than most libs i know of
[21:37] <Daemon404> if i didnt work on the libs themselves, using it work be even more horrible
[21:37] <Daemon404> some stuff is still impossible with public api only
[21:37] <saste> finding a lib which is really clean is hard, i challenge you
[21:37] <Daemon404> its really fun
[21:37] <Daemon404> [15:37] <@saste> finding a lib which is really clean is hard, i challenge you <-- ffms2
[21:37] <Daemon404> which i in fact use in favour of libav* many times
[21:38] <ubitux> will we sed s/AV/AVS/ when the project has an honorable subtitles support?
[21:38] <saste> Daemon404, but it's another layer of complexity
[21:38] <saste> the classical carpet over the dirt
[21:38] <Daemon404> libav* api is too generic
[21:38] <Daemon404> and low level
[21:38] <Daemon404> it doesnt even have "get me a frame"
[21:38] <Daemon404> most people do NOT want that
[21:39] <Daemon404> also it has fraem accuarte indexed seeking
[21:39] <Daemon404> i dare you to try and implement that in libav*
[21:39] <saste> yes i wonder if there is a reason for *not* to implement that cleanly in libav*
[21:39] <saste> sometimes people try, and disappear
[21:40] <saste> no i never tried that
[21:40] <ubitux> i often hear "vfr" when talking about that
[21:41] <Daemon404> ffms2 handles vfr properly
[21:41] <Daemon404> so whats your excuse?
[21:42] <saste> "less than straightforward and less than perfectly documented libav API"
[21:42] <ubitux> Daemon404: what do you think is preventing that in libav*?
[21:42] <ubitux> except "lavf is a mess"
[21:42] <Daemon404> saste, thats entirely accurate
[21:42] <Daemon404> especially when it was written, eyars ago
[21:42] <Daemon404> ;)
[21:42] <saste> this is the usual argument: ouch this is a mess, i'll build another layer of complexity rather than fix it properly at the right level
[21:42] <saste> then the next guy comes and see the wrapper
[21:43] <Daemon404> saste, libav* will never ever have that simple of an api
[21:43] <Daemon404> ever
[21:43] <saste> ouch this is a mess, i'll build *another* layer to cover this bloody mss
[21:43] <saste> and again and again
[21:43] <Daemon404> because certain people have moral objections
[21:43] <Daemon404> to simplicity
[21:43] <Daemon404> and non-genericism
[21:43] <saste> this is the story of audio in linux
[21:43] <Daemon404> "it should only decode packets!"
[21:43] <saste> but people is free to do whatever they seem fit them
[21:44] <Daemon404> i would really like a lot of the logic from ffmpeg.c 
[21:44] <Daemon404> moved to api
[21:44] <Daemon404> it would be -really- useful
[21:44] <saste> that's what freedom is about after all
[21:44] <Daemon404> freedom doesnt mean its practical
[21:44] <saste> Daemon404, my Great Evil Plan is to move all ffmpeg functionality to libavfilter
[21:44] <Daemon404> eh
[21:44] <Daemon404> tahts just not feasible
[21:44] <saste> then you basically need a movie source, plus some filters
[21:44] <Daemon404> for a lot of libavcodec specific things
[21:45] <saste> like?
[21:46] <Daemon404> things that would make useful higher level api for lavc
[21:46] <Daemon404> without mixing in lavfi
[21:46] <Daemon404> which
[21:46] <Daemon404> shock
[21:46] <Daemon404> adds a layer of compelxity
[21:46] <saste> yes that's true
[21:47] <saste> i realized it soon after
[21:47] <saste> but at least it is in the same low level project
[21:47] <Daemon404> that is irrelevant
[21:47] <Daemon404> (fwiw i am very soon disabling libavfilter entirely for us)
[21:47] <Daemon404> and moving to vapoursynth
[21:48] <saste> ah windows specific stuff?
[21:48] <Daemon404> no
[21:49] <Daemon404> its cross platform
[21:49] <Daemon404> avisynth replacement
[21:49] <Daemon404> i plan on writign lua bindings and makign an input module for lavf
[21:50] <JEEB> supports video atm, and after that gets more or less finished there probably will be some kind of audio support
[21:50] <JEEB> official bindings are for python
[22:16] <ubitux> (did i miss something?)
[22:50] <michaelni> ubitux, no, just <Daemon404> (fwiw i am very soon disabling libavfilter entirely for us)
[22:50] <ubitux> ok :(
[22:50] <ubitux> thx :)
[22:50] <Daemon404> keep in mind i am keepign essential ones for ffmpeg.c
[22:51] <Daemon404> audio mostly.
[22:51] <michaelni> Daemon404, why ?
[22:51] <michaelni> you should remove them too, really :)
[22:51] <Daemon404> ?
[22:51] <Daemon404> im not sure i follow
[22:52] <michaelni> well, i think if there are problems in libavfilter, they should be fixed
[22:52] <michaelni> if people want a higher level interface that is simpler
[22:52] <michaelni> one should be added
[22:52] <Daemon404> the problems cannot be fixed
[22:52] <Daemon404> its the very design that is wrong
[22:52] <Daemon404> i.e. a monolithic library
[22:53] <ubitux> Daemon404: we're almost to the point were we could have a 150-line tools feeding libavfilter with a filtergraph supporting the whole chain; like: ./ffstreamer "movie=in.avi,...,outmovie=out.avi"
[22:53] <Daemon404> it's "scripting" rapes my eyes
[22:53] <Daemon404> more or less
[22:53] <Daemon404> id rather have real langauge bindings, and many newer plugins
[22:54] <Daemon404> not to mention easy frame accurate seeking for input
[22:54] <ubitux> well libav closed the api recently so it doesn't help supporting additionnal users filters or stuff like that :(
[22:54] <Daemon404> its still wrong
[22:54] <ubitux> Daemon404: did you see the [a]sendcmd from saste?
[22:54] <Daemon404> it would still require a recompile
[22:54] <Daemon404> (of the user app)
[22:54] <michaelni> Daemon404, this is not hard to fix
[22:54] <Daemon404> ubitux, tl;dr it's no repalcement for avs or vs and its use cases
[22:54] <Daemon404> ive said this so many times
[22:57] <michaelni> Daemon404, if it fails for some use case, please open a ticket on trac and explain the use case as a feature request
[22:58] <michaelni> i think libavfilters low level API is quite well designed, whats missing is as you say / i understand you dynamic loading a nice easy API for such plugins
[22:58] <Daemon404> i dont feel liek opening several hundred tickets
[22:58] <Daemon404> and a real way to bind languages
[22:59] <Daemon404> i.e. lua or python
[22:59] <michaelni> Daemon404, you seem to feel like complaining alot though (not just about libavfilter)
[22:59] <michaelni> and no offense intended, this is IMHO not helping anyone
[22:59] <michaelni> we should try to improve what doesnt work as it should
[22:59] <Daemon404> you are indeed correct
[23:00] <Daemon404> you must understand though, many times when i mention things, "why arent you usign libav* for this?" and similar rhetoric shoved down my neck
[23:00] <Daemon404> thats in fact what suprred this (in a different network/channel with a different dev)
[23:01] <Daemon404> apologies for carrying it over.
[23:01] <ubitux> Daemon404: we would really appreciate a well written mail/blogpost/article/whatever about the needs and expectations you and some other people have about a filtering library, and how libavfilter fail to achieve that, as well as what kind of things would be nice
[23:01] <ubitux> because we can't do much except listening to your incomplete complains here
[23:02] <Daemon404> i summed it up once before for libav
[23:02] <Daemon404> general response i got was "we're nto gstreamer"
[23:02] <Daemon404> or
[23:02] <Daemon404> "why would you want to do that?"
[23:03] <Daemon404> and at least one lecture on why dlopen is bad
[23:03] <Daemon404> i will write one now though and send
[23:03] <ubitux> thank you
[23:04] <Daemon404> warning: i know ill forget some
[23:04] <cone-191> ffmpeg.git 3Michael Niedermayer 7libavcodec/get_bits.h: get_bits: get_bits_long() support n=0 as the docs allow it
[23:04] <cone-191> ffmpeg.git 3Michael Niedermayer 7libavcodec/x86/dsputil_mmx.c: dsputil_mmx: put optimized gmc code back and avoid a VLA without loosing features.
[23:05] <ubitux> Daemon404: take your time :)
[23:06] <Daemon404> its the public api for lavfi documented?
[23:06] <Daemon404> is*
[23:06] <Daemon404> iirc there were complains -- i dont know if it's been rectified
[23:07] <ubitux> there are filtering examples
[23:07] <ubitux> http://git.videolan.org/?p=ffmpeg.git;a=tree;f=doc/examples;hb=HEAD
[23:07] <Daemon404> exampels are not documentation
[23:07] <ubitux> i personally still find the public api a bit complex
[23:08] <ubitux> aren't they?
[23:08] <Daemon404> they are not a substitute.
[23:08] <ubitux> the functions are likely doxified
[23:08] <Daemon404> oh
[23:09] <Daemon404> i should also sent an email about why print_options is not a good solution
[23:09] <Daemon404> (separate issue)
[23:09] <Daemon404> also a tricky one.
[23:20] <kierank> "Greek harmony" = golden ratio?
[23:22] <ubitux> saste: didn't you have a patch to remove the '-' in the ffmpeg help for the filters?
[23:23] <ubitux> and pretty-printing related patches
[23:23] <saste> kierank, yes, join the phi lovers club
[23:24] <saste> ubitux, yes, if i remember where i put it
[23:27] <ubitux> i wonder if we shouldn't remove the "AVOptions" thing too
[23:27] <ubitux> anyway, adding some more padding, remove the '-' and put some little indent would really help
[23:27] <ubitux> iirc your patch(es) was/were improving this
[23:28] <ubitux> btw, kind of related, but the "Getting help" message should appear with no input
[23:29] <Compn> michaelni : i have a complaint about ffmpeg. the whole thing is monolithic, and gigantic! and i dont have any idea how to make it smaller, but its all rubbish ... :)
[23:29] <Compn> :P
[23:30] <Daemon404> ffmpeg's libraries have deps on the otehr libraries in ffmpeg in a bad way
[23:30] <Daemon404> i.e. non oublic symbols
[23:30] <Daemon404> nonpublic*
[23:30] <Daemon404> so yes, youre sort of right
[23:30] <Daemon404> :)
[23:30] <Compn> we should make every symbol public and that way mplayer wont have trouble.... ;)
[23:31] <Daemon404> lol
[23:31] <Compn> or we could just rename every symbol every three months so it looks like we are developing code...
[23:31] <Compn> breaks api? who cares!
[23:32] <Daemon404> eh
[23:32] <Compn> already renamed the main programs, that was important...
[23:32] <Daemon404> api breaks more than every 3 months
[23:33] <Compn> ffmpeg will be depreciated so use avconv! 
[23:33] <Compn> wonder when it will finally be depreciated...
[23:33] <Compn> in the year 2022
[23:34] <Compn> wait
[23:34] <Compn> did you just complain about ffmpeg's libavfilter options being like a script
[23:35] <Compn> and then immediately say it should be like avisynth's .. options ... being like a script that you have to make a script to use?
[23:35] <Compn> Daemon404 : do you want feature request to be able to use ffmpeg via script like avisynth ?
[23:36] <Daemon404> Compn, either that or people stop pushing lavfi on me as an avs replacemeny so zealot-like
[23:36] <ubitux> Compn: give him a chance to express him properly, let's not ask him to much for trolling here :)
[23:36] <Compn> ubitux : lol ok , we troll later :)
[23:37] <ubitux> we'll troll after the post explaining deeply the issue
[23:37] <Compn> k
[23:37] <ubitux> which will be far more interesting and constructive :)
[23:37] <ubitux> (hopefully)
[23:37] <Compn> Daemon404 : who said to replace avisynth with libavfilter ? thats dumb. libavfilter doesnt have half of the filters that avisynth has
[23:37] <Compn> its not ready to be a replacement
[23:37] <Daemon404> Compn, response:
[23:37] <ubitux> i said stop it Compn! :(
[23:37] <Daemon404> PORT IT ALL
[23:37] <Daemon404> PATCHES WELCOME
[23:38] <ubitux> see... :()
[23:38] <ubitux> :(
[23:46] Action: Compn realizes that avisynth is based on vfw, remembers getting scolded about using ffdshow because it was too 'ancient' and lav filters wat hte future
[23:46] Action: Compn brain explodes
[23:47] <Compn> ah avisynth 3 was going to use a real language for scripts, instead of crazy-reinventing windows batch files
[23:47] <Compn> ubitux : so talk to me about something so i stop trolling
[23:47] <Compn> multimedia.cx is back up
[23:47] <ubitux> do you love pokémons?
[23:47] <Compn> yay
[23:48] <Compn> no not really
[23:48] <ubitux> :(
[23:48] <Compn> like power rangers, their novelty has died on me
[23:48] <Compn> and pogs
[23:48] <ubitux> pogs... i almost forgot about that :D
[23:49] <kierank> 22:23:31 <saste> kierank, yes, join the phi lovers club --> can you perhaps call it the golden ratio instead
[23:50] <kierank> i don't think i've ever heard it called greek harmony in english
[23:51] <saste> kierank, it was subtly humorous
[23:51] <kierank> oh
[23:51] <kierank> whoosh
[23:51] <saste> but i don't mind if you want to change it (PAWE)
[23:59] <cone-191> ffmpeg.git 3Michael Niedermayer 7libavcodec/x86/dsputil_mmx.c: gmc_mmx: enable also for large pictures when emu edge isnt needed.
[00:00] --- Sun Oct  7 2012