VLC 2.1.0
videolan.org
videolan.org
First, this is a release that fixes some important architecture mistakes we've done in 2.0.x branch of VLC. I'm notably speaking of the lag in reactivity, notably on volume change (that was shared on the mpv thread) and seeking, but also some grave video settings propagation. I wish we could have fixed and shipped that earlier, but we couldn't (long release cycle).
Then, this is the first official release of libVLC that is LGPL for most of what you need as a developer, including the right modules. SDKs for Win32/64, MacOSX, iOS and Android are getting ready.
If you are a web developer, our VLC plugin now supports Windowless, to fill the gap between Flash and HTML5 (it should work on IE6,7,8 without too much work).
If you are on Mac OS, the interface is finally polished after the major changes of 2.0.0 :)
Finally, we decided, as a community that we will accelerate the major release cycle of VLC. The fact that we needed 1,5 year to get the fix to some critical audio core and video settings issues out is way too much. We will move towards a 6-months schedule with LTS.
Sure, there are other very good players on each platform, but we are doing our best so that you can play everything everywhere for free, using open source technologies :)
I know this feature probably wouldn't be useful to anyone but me, so I'm not really putting in a feature request. Just hoping to discover that someone else has already figured out a way of stepping backwards frame-by-frame, to complement the existing "step forwards" functionality.
See http://ffmpeg.org/ffmpeg.html#Video-and-Audio-file-format-co... for more examples.
This is, in fact, far and away my biggest complaint about VLC. You can say it's slow, but it's just a matter of caching the frames, something that QuickTime seems to have no problem with for the formats it supports.
And you can only cache a couple of frames or so.
Anyway, patches very welcome on that.
The issue is that doing a correct back-frame can mean going back up to 2 I-frame in the back, in some situations.
I guess that would break when you jump to a specific location in the video, without the benefit of having cached the previous frame(s).
There's no reason why the software can't cache frames (it's only 25fps normally) or fast forward from the last I-frame. 25 frames of 1080p HD video is 200MB, as uncompressed 32bpp bitmaps. I have 24GB of memory in my PC; that's the worst case, and it's not bad. A bunch of prior frames can be cached, and the next bunch recalculated if it looks like the user is going to continue stepping backwards.
Around 8 TB~ of content, estimating about: 666 KB/s.
It's probably higher then that. I just wanted to thank you.
Just wanted to add on a serious note, VLC is great at everything I've thrown at it. Radio and video streams work fantastically and the interface is simple and efficient.
Somehow still can't figure out how to do the one additional thing I want (above and beyond what I have already sorted out how to do), which is quit the player and re-open it later at the TS of where I left off with my video. :-)
Can you give any high level updates on the VLC for Window 8 effort? I've been following the Kickstarter updates and know it's been more difficult than expected, just wondering if there's any new news.
Thanks for all your hard work anyway. VLC has been my default media player for years now. I can't think of another piece of software that I've used so regularly for such a long time.
And yes, a bit of polish has been done on the OSX interface, but there is a bit more yet to do :)
If noone wants to work on that and shares design, it isn't done.
Edit: Nevermind, looks like this has been answered. VLC waits a week before releasing the update to the public.
The rest is volunteer work :)
If you have a million or two to spare, we could do something though :)
To be honest, we need code more than money.
I really hope all these great improvements are shared between the media player projects as much as possible.
For example, intel VA-API support was ported and integrated into master in about a week. With support for VA-API VPP (video post-processing), so deinterlacing comes free too (just like for VDPAU).
That means that on my weak intel boxes, decoding 1080p content takes about 3-5% CPU and 25MB. Which means I can play 16-32 simultaneous videos in a grid at the same time. You should've seen the look on my clients' face when they saw me pull that one off, where the former developer couldn't even do one HD clip without lagging horribly.
It is, in a word, awesome. Just like this new VLC, which I use on my mac. Also awesome. I actually looked at the new VLC plugin code to make my own browser plugin for mplayer/mpv. Though in the end I didn't decide to support a windowless interface like VLC (because I couldn't really, mplayer-likes don't support that). That's fine though, windowed mode was far easier to code.
Two issues that are very annoying to me:
1. The lack of "Remember last played position" option is very frustrating. This option is now part of many popular media players.
2. The "Disable screensaver" option doesn't work properly. When you pause a video, VLC will stop the computer from going into sleep mode. This is pretty awful! VLC should only disable the screensaver while the video is being played.
And one last thing that would be nice: the ability to move the main VLC window by clicking and dragging on the video area, and not just by the titlebar.
2. yes, pause is bad in VLC, because it still displays...
I still think there's some weirdness going on with using F11 vs double-clicking to switch to full screen view for video playback.
Overall great job with this release.
BTW, I'm also loving the fact that I can use VLC on Android. I compiled it from source back when there were no binaries, currently am using the nightlies, and am looking forward to being able to get the stable releases in the app store eventually.
Even when I launch a file from Finder, I get a split-second blink of the Playlist. And when the clip stops, I see Playlist instead of the starting screen and can't drag and drop to play files to it anymore.
When I disable the Playlist by pressing its button on the interface, the expanding transition of the window when opening a file is oddly jumpy — hopefully an easy fix in future releases (I'm on OSX 10.8.5). Playlist still appears at times.
The standalone Controller module from the interface… I miss it, any chance of it ever returning?
Back to the two years old VLC 1.1.12 for me, it was much better thought-out interface-wise (Playlist is just a functionality, not the driving feature and Controller is still there) — and it still plays every file I need it to.
I will of course keep checking for updates.
In the time since,
* I have never used the playlist on purpose once
* I kept wanting to put the video always on top in a tiny window while I do something else, and always visible controls/strange resize limit make it worse than before. Compare: http://imgur.com/c4V9k12
So I wish the VLC guys the best, but I have moved on to MplayerX and only keep VLC.app around for some hypothetical case in which the other player won't work.
I'll keep hoping for 3.x interface changes.
We finally have H264 hardware decoding on Mac. That's the single largest missing feature that prevented me from completely switching to VLC. I used to open H264 videos with QuickTime.
Good job guys!
Thank you guys. You are true FOSS heroes.
This problem makes VLC useless as my default music player :'(
I tested on python a music playlist and randomly arranged the numbers, and also tried the same on vlc. I found in python, numbers were really random (done maybe a month or two ago, so they implemented the new algo). However, in vlc, this was the issue:
Songs would come in as [6, 7, 8, 3, 4, 1, 2, 5]
So the "random" or "shuffle" was kind of half-done. You would have the same artist's songs being played because if you sorted by album they were right next to each other.
I went on #vlc once and I tried submitting a bug report. I seem to recall one or two was submitted online, but many claim it is simply not a bug. Random or shuffle, either way the definition is skewed as there is some linearity.
Luckily, I have cmus, and I connected that with cmus-remote to my android. ;) Trust me, once you have a nice linux ncurses player for terminal, it is very hard to seriously consider any competitors. Vlc is still nice though to tell my friends to install on windows machines for anime. And because they have a music library of ~30 songs, they will probably not notice the shuffle issue.
What I mean to say is imagine a deck of cards, to be "shuffled".
Most players will shuffle by rearranging the cards, and you'll get the occasional adjacent cards next to each other, or you may not, it is just random all around.
In vlc, you get more of a beginner's shuffle (say, from a child). Cards are mixed, but not shuffled fully. This results in play which prefers groupings of cards (Ace of spades, 2 of spades, 3 of spades), but at a common rate. Therefore, we can say it is "sort of" shuffled.
I preformed about five tests, and found as little as ten songs will show the effect. My test was install the newest vlc (windows 7 64-bit build), add 10 songs to playlist once it opens, then hit shuffle and record how the 9 songs are shuffled (assuming you are on song 1 in the playlist).
To have comparable random data, I did "import random" in python and then random.shuffle with an array with values in the range of ten.
From this, I gathered vlc was not random "enough", which for a music library of around 10,000 (11,000 now, but 10 at the time), this became irritating enough that I view vlc more as a starter multimedia platform. Were this issue to be fixed (or even seen as an issue!), I would be quite surprised.
Just because something seems intuitive doesn't mean all developers will implement it the same way.
VLC shuffles the playlist upon hitting the shuffle button and does not alter the order until shuffle is disabled and reenabled again. I'd like VLC to go through the whole playlist in a random order once, and when every song has been played instead of reusing the same pattern again, reshuffle.
It is not hard to fix, but it needs someone who wants to spend one week-end on that part.
While VLC is excellent and we would like and want to use VLC as a plugin in our corporate environment, we sadly cannot, because there's currently no way to configure VLC plugin to use a proxy. In order to get to the internet, everything must go through the proxy. It would be very welcome if You could implement this little functionality. VLC plugin should just get the proxy settings from the browser and it would be done. Ofcourse, bonus points for SPNEGO.
I have just made my HN accounts only to post this request.
Keep up the excellent work!
Edit: minor typos.
It seems to take a long time: http://git.chromium.org/gitweb/?p=webm/libvpx.git;a=tags
It's great to see significant improvements to the software you use everyday.
In any case, this bit can be found in the Copyrights Tab of the About section:
"- This program contains the video decoder FFmpeg (http://ffmpeg.mplayerhq.hu) was used, FFmpeg subject to the LGPL license. LGPL license, refer to the installation directory LGPL.TXT. The modified FFmpeg sources can be downloaded from the http://potplayer.daum.net/devel/ffmpeg.zip."
I don't know if that resolves the situation though.
The reason is rather simple. VLC on Windows is just plain ugly. You could tell this is a Linux software ported to Windows. It doesn't even need to complex and fancy. Take a look at MPC-BE, plain simple and stylish.
And it isn't all just about the looks. The settings, menu placement, icons, etc.
I really wish something could be done about it.
For example, VLC's dithering of 10-bit content still seems to be worse than CCCP out of the box, leading to banding where there shouldn't be any, and it seems to be doing something weird to the colors in all the things I tested it with. I'm not exactly sure what it is, though, as it's not a color matrix or luma level issue - I'd post some screenshots but I'm at work now. Also, Nvidia users still seem to have luma level issues out of the box that requires a trip to GPU settings to fix.
Congrats on the release anyway, VLC has come a long way from the 0.8.6 days as far as high-end media playback is concerned.
Our goal is not to match the crazy setup people do and spend hours to tweak.
For the 10bits dithering, use the OpenGL output, and it is done.
Installing CCCP is as easy as installing VLC and gives you better quality out of the box. Installing and configuring madVR is an extra five-ten minutes of effort. Hardly a "crazy setup" that would require "hours of tweaking". There's also KCP which comes with madVR preconfigured, though CCCP offers better compatibility with a wider range of setups (which is why I personally recommend CCCP with madVR as a separate optional component on top of it).
Also, at least in the previous version, 10-bit dithering was still lacking even with OpenGL output, as demonstrated here: http://screenshotcomparison.com/comparison/32677
I haven't been able to test if the situation has improved in the new version yet, though.
Last version, the matrix colors were wrong, as clearly explained on the release page, so it's soooo great to compare with it.
And maybe, if you were not an asshole to us, we might be interested in listening to what you have to say. Until then, not so much.
CCCP requires no configuration. It just works. And I said that I recommend madVR as an optional component, since CCCP already provides high quality playback out of the box. I've been saying this for a good while now but you keep talking like I said madVR would be a must and use that as a basis to claim that VLC is "just simpler".
Installation and basic configuration (the average user really doesn't need any more than that) madVR on top of that with a guide (and there are plenty) is five-ten minutes of effort, but as I said, it's not a necessity by any means.
>Not to mention the non-free licences...
All the playback components in CCCP are FOSS with the exception of Haali Media Splitter, which is more of a fallback at this point. I'm not really sure about the non-component parts (the installer and settings panel), though. madVR is proprietary, yes, but it also happens to be the most advanced video renderer in existence. Unless you subscribe to RMS' worldviews about proprietary software (in which case you wouldn't be using Windows to begin with), I don't see a reason not to use it if you have enough interest in high quality playback.
>Last version, the matrix colors were wrong, as clearly explained on the release page, so it's soooo great to compare with it.
We were talking about dithering though (which that comparison also demonstrates), so unless the matrix issue was directly linked to the dithering issues it's not really relevant (and I didn't mention it either). And as I already said, I haven't had the chance to compare the latest version properly yet.
Also, chill down a bit, will you? While VLC likes to advertise itself to anime watchers, it's an audience that on average cares a lot more about high quality playback than many others. And if your "goal" is "not to match the crazy setup people do and spend hours to tweak", then why would you get upset when I point out that these "crazy" setups (that at their very basic level are as easy to install and use as VLC is) give better results?
> FOSS with the exception of No comment...
> We were talking about dithering though (which that comparison also demonstrates), so unless the matrix issue was directly linked to the dithering issues
How do you do the dithering if it is not during the YUV->RGB conversions, aka the matrix? If it is not done there, it is done in swscale like all the other players...
> Also, chill down a bit, will you?
I am usually quite calm, as you can see on this thread or the reddit one. But I'm not to people like you, who are at the limits of insults (and not on this thread, I know). Once you plan to speak correctly, maybe we can talk in a calm down way.
I thought you were referring solely to the fact that in the comparison I posted, the video was incorrectly converted with BT.601 instead of BT.709 on VLC's end.
Anyway, I'm home now, which means I can properly compare stuff (at last). Dithering certainly seems to have improved since the last version and BT.709 is used as well. There's something weird going on, however - the colors in VLC appear slightly off for some reason (and it's clearly not the usual suspects of TV/PC levels or BT.601/BT.709): http://screenshotcomparison.com/comparison/42453
Still, a notable improvement for sure.
Sorry, but the color difference is very very slight (if even noticeable), and are you sure the issue is on the VLC side? Did you try mplayer2 and mpv?
Uhh, what? I have not written something like that about the new version anywhere on the internet.
But for the record, I do think marketing VLC as "4K-ready" is silly - there really shouldn't be anything special to 4K besides the bigger resolution, meaning that if you have the resources to decode it, you should be able to play it, and it has been very much possible with other playback solution (and likely even with older versions of VLC? I haven't checked). In a way, it's similar to this[1] - what it's advertising is actually nothing new but because 4K is the hot new marketing term of the day they're going to plaster it all over the damn thing.
That's rich, considering the majority of anime today is garbage in production quality compared to anime of the past (in animation and story). Anime fans care deeply about playback quality? Yeah right. The minority do. The majority watch from CrunchyRoll or some free site like gogoanime. People such as yourself are such a rare breed, it would be an utter waste of time and resources for VLC to cater to you, and I'm not even sure why they make any attempts.
So I don't see a problem with VLC.
I agree that 10-bit playback needs to be addressed. However it seems either CCCP or mplayer2 work great with no noticeable difference in quality. Only VLC seems to have an issue with decoding 10-bit video streams.
Works fantastically well for everything else though.
(Win64 VLC build on Nvidia hardware)
Give me a note if you would like me to do more testing.
BTW: After the crashes... what kind of information sends the bug report function to you? Anything privacy related (file name, ip address, etc.)?
It took me 5 min to do the same using VLC.
Could you maybe share some news on the status of the web plugin for the Mac?
Is there any way to actually get VLC on iOS anymore?
http://i.imgur.com/oLHiQgR.jpg
so when I'm watching a movie, I will switch to Windows and play it on MPC. For everything else VLC is fine. Annoying, I've tried filters and other fixes, but nothing worked.
Media Player Classic uses codec decoders you have installed on your PC, VLC ships with everything it needs to play your videos. In other words, you should not be comparing VLC to MPC, but rather VLC's underlying codec library (from ffmpeg, I believe) to the active codec decoder being used on your Windows machine.
For example, I too use MPC on Windows to get better video - but that's because I have CoreAVC installed and MPC uses it to play h264 content. CoreAVC blows everything else out of the water in terms of performance and quality.
What codec are you using on Windows w/ MPC?
The washed out colour in that screenshot can probably be fixed by changing the video output method (to OpenGL/Direct2D/Direct3D) or changing the default luma range in the graphics card settings from 16-235 to 0-255.
The differences you have captured here indicate that at least one of them is applying non-standard filters of some sort. Perhaps it's just a matter of the right settings (though I cannot offer help in determining what they are)
>New 6.1 downmixer to 5.1 and Stereo from MKV/Flac 6.1.
>Correct YUV->RGB color matrix in the OpenGL shaders.
I like how they put a separate entry for pirates.