Media Player Classic supports Matroska v2, *.mp4, *.mov, and MPEG-4 Timed Text

[Elias]

OGM is a dead format. I would suggest you to use MKV instead.Not to mention, a hack of OGG, which means no real support for such things as this. And yes, use MKV if you want an open source container suitable for all kinds of codecs.As some people from HydrogenAudio proved, AAC in current implementations is quite often worse then MP3.What? Are you stating that mp3 is a better audio codec than AAC as we speak?The best AAC encoder seems to be WinAMP. You can use WinAMP's filters in BeLight a you don't have to download 100MB of Nero's all-in-one, which can do everything, but nothing correctly.Last I checked, Apple's AAC encoder gave best quality. What does Nero's AAC encoder handle/do in an improper way?

[RyosukeFC]

I understand quite well the interest in trying out new things, but considering my needs (receiver-based decoding of any multi-channel material over S/PDIF and highest audio quality all around), as well as the procedure I've standardized myself upon (DVD Decrypter -> DGIndex -> GKnot (not autogk) => .AVI, then VDubMod muxes the .AVI and audio extracted from the above DGIndex step (DTS whenever available, then 5.1 AC3, and Vorbis for all 2ch material), it doesn't make sense for me to use AAC for anything (since my receiver can't decode it, can any?), or a container format other than OGM. I've never had any trouble with OGM, and VDubMod supports it fully, unlike MKV. When Matroska support in VDubMod materializes, I'd be happy to start using that. VDubMod is so useful for things above and beyond encoding and muxing, I'm not yet willing to abandon it and start creating files it can't read in.

Can anyone see any flaws in my logic? Am I missing anything? I understand the preference for formats that properly support Unicode (I'm a big supporter myself), but I don't need anything other than vanilla English support for my own purposes, so I'd like to wait for VDubMod to properly support the full Matroska spec before switching from OGM. Does this make sense to anyone? Also, why are people even discussing anything made by Apple? What could possibly be interesting about anything QT-related, in the context of the discussion of the selection of superior formats? I think there's something key to this whole thing that I'm missing.

@Vlada: I'm also curious about the role pixel shaders have in video playback. Does anyone know?

[Liisachan]

Yeah, basically I totally agree with you, I've been waiting for SV8,
but I suppose that won't come soon...

BTW before someone gets wrong idea reading vlada's post:
MPC could mean both Media Player Classic and Musepack.
Vlada said "MPC seems to be dead," but that doesn't mean Media Player Classic.
This is getting too off-topic, but there is some complication about Musepack.

[vlada]

Liisachan> I don't say that MP3 is better format then AAC, but current implementations of AAC (FAAC, Nero) give sometimes a worse result then MP3
(LAME 3.97). Have a look at this (http://www.hydrogenaudio.org/forums/index.php?showtopic=36465) thread. Of course AAC has a higher potential then MP3, but it just seems that current encoders don't use it.

Regarding Nero, I meant to say that it has many bugs in all of it's applications. Not that there is a bug in AAC encoder. But anyway the Nero AAC encoder doesen't seem to give a good quality results, but the people from Nero say they are working on improvement.

RyosukeFC> VDM is also a dead project and is not developed for more then a year, IIRC. So don't expect to have a correct support for MKV there. To work with MKV use AVI-Mux GUI or MKVToolnix. I don't say, that OGM is a bad container, but it is unsupported and you will likely run into problems with it.

I would suggest you to produce a video file using DGIndex+AviSynth+VD (or VDM) as you do now, then mux it with original DTS or AC3 sound with MKVToolnix. It easy to do and it gives you a lot of new possibilities. A perfect splitter with chapters support, a good support for multiple audio tracks and a lot more. I would SUGGEST you to switch to MKV, but if you don't, it's your choice. I can't make you ;)

P.S. Sorry for this offtopic.

[Liisachan]

RyosukeFC> VDM is also a dead project and is not developed for more then a year, IIRC. That's not true.
VirtualDubMod (http://sourceforge.net/projects/virtualdubmod/)1.5.10.2 is dated August 29, 2005, and it's one of my everyday tools for typesetting when subbing.

[videomixer9]

AAC is generally a really crappy codec, MP3 is so mature that is can easily produce similar quality nowadays, to get a real improvement you should still use OGG Vorbis Audio, it's a real free format and it has superior sound quality. He may not do wonders at low bitrates like HE-AAC but for general material it's the best codec to choose.

Unfortunately OGG doesn't support mixing channels together for multichannel audio, thus you need higher bitrates for 6ch audio. However AAC sounds pretty much crappy on a good 5.1 sound system with almost any bitrate, AC3 or DTS are still far superior to AAC. I think the problem is the prefiltering on playback that AAC requires, it uses basically the same trick MP3Pro introduced.

OGG Vorbis should also work in MP4 so you can actually keep using and just ditch AAC. AAC is just hyped because it's prolly the next industry standard for audio and the thing you call MPEG-4 audio, however, it sucks. Industry likes it as it's just another codec that they can make lots of cash with by licensing and of course it supports DRM, just the same reason they like the mp4 container, as you can also enhance it with DRM.

Anyone still using ogm must be an idiot :devil:

[Elias]

Anyone still using ogm must be an idiot :devil:This just makes your bold statements about how much "AAC sucks" sound invalid and unprofessional. Maddox (http://maddox.xmission.com/) type of statements aren't popular around here.

mp3 can use DRM too if I remember correct, yet Apple, Nero, Winamp and other companies favour AAC over mp3. The only reason why mp3 is still being used over other superior codecs like AAC, OGG Vorbis etc, is because of its compatibility. That's it. mp3 is old and outdated (even though Lame is really good), and as much as the current implementations of AAC don't use the full potential of AAC, I seriously doubt they're all inferior to mp3.Have a look at this (http://www.hydrogenaudio.org/forums/index.php?showtopic=36465) thread.That's an old test, iTunes didn't even have VBR AAC at the time. CBR is always inferior to VBR in terms of quality and filesize.

[videomixer9]

it's obvious for anyone who got a good sound system and some working ears, AAC often suffers from crispy sound and especially with 6ch audio has a noise loaden flat sound. The decoder applies this special filter for HE-AAC. For 2ch audio it suffers mostly from the crisp sound that's not really what i'd call clear and in blind hearing test i'd bet most ppl would say an 96kbps MP3 sounds better than the same bitrate AAC file. AAC only has a better scaling on low bitrates, but with video that's not meant for low bitrate streaming you'd never use so low rates anyways, and to get 6ch AAC to sound good you're almost at rates many AC3 streams are at anyways.

AAC encoders are far from mature, just like most H264 encoders. Maturity is one thing you should make a choice by, and there is not a single mature AAC encoder unlike there are for ogg or mp3.

[bond]

Unfortunately OGG doesn't support mixing channels together for multichannel audio, thus you need higher bitrates for 6ch audio.right

However AAC sounds pretty much crappy on a good 5.1 sound system with almost any bitrate, AC3 or DTS are still far superior to AAC.you can never compare a source ac3/dts with a reencoded other audio format because the reencode will always have less quality

but yeah it would be interesting to see a codec comparison comparing ac3/dts performance on 5.1 compared to aac encoded from a real raw 5.1 source at lets say 192kbps

i never saw such a comparison but a wild guess would be you are totally wrong :p

I think the problem is the prefiltering on playback that AAC requires, it uses basically the same trick MP3Pro introduced. there is no prefiltering done in normal aac

OGG Vorbis should also work in MP4 so you can actually keep using and just ditch AAC.yeah it works, but its not standardised so it will hardly work anywhere

AAC is just hyped because it's prolly the next industry standard for audio and the thing you call MPEG-4 audio, however, it sucks. Industry likes it as it's just another codec that they can make lots of cash with by licensing and of course it supports DRM, just the same reason they like the mp4 container, as you can also enhance it with DRM.drm is not a thing of the container or audio format

[videomixer9]

sorry for the prefiltering, meant postfiltering of course ...

[bond]

sorry for the prefiltering, meant postfiltering of course ...there is also no postfiltering done in aac

[vlada]

Liisachan> I didn't know about this VDM release, I was using 1.5.10.1 from December 2003 - almost 2 years old. I couldn't find a changelog, do you know what's new? It would be great if somebody starts working on this program again.

Elias> The comparison of MP3, AAC and Vorbis, I reffered to, is from August 2005, it is the newest one I saw.

bond> What is meant by channel mixing? Is it something like joint-stereo? You create a "middle channel" and then you only store differences between the main one and the other channels? That shouldn't be so hard to implement. Why doesn't Vorbis support it?

Btw. I think we got quite off-topic. Why not split this thread and move this discussion to the proper section?

Vlada

[bond]

bond> What is meant by channel mixing? Is it something like joint-stereo? You create a "middle channel" and then you only store differences between the main one and the other channels? That shouldn't be so hard to implement. Why doesn't Vorbis support it?its called channel coupling and its afaik something similar as you wrote
vorbis doesnt support it as noone implemented it till now

[Liisachan]

A few users reported that recent versions of Media Player Classic got problems when handling streaming formats, such as .asx.
http://forum.doom9.org/showthread.php?p=735095#post735095
I use MPC mostly for internet streaming, since version 6.485 the player couldn,t
handle the asf stream included in asx files. Prior versions works perfectly.
The following are what I gathered. This is not my direct experience and may be inaccurate, but 6.4.7.8 and newer versions seem to have these problems, and older versions are generally okay:

(1) MPC doesn't understand the relative path in .asx, such as

<ENTRY>
<REF HREF="./relative/path/to/video.mpeg" />
</ENTRY>
(2) MPC adds URL in <MOREINFO> to the playlist as a file to play:

<BANNER HREF="http://sample.com/logo.png">
<ABSTRACT>About Us</ABSTRACT>
<MOREINFO HREF="http://sample.com/dont.play.me.html" />
</BANNER>
(3) MBCS in .asx may make MPC crash, whereas WCHAR in .asx is ok

(4) The same ASX file doesn't work if the extension is not .asx (This should be 'by design')

(5) Streamed .rm / .ra doesn't play just because the file extension is .ram

(6) Streamed vorbis.ogg doesn't play by default. Using File extensions|.ogg|RealMedia may fix the problem as a workaround.

(7) Streamed (? Embedded-in-HTML?) OGM doesn't play by MPC's internal splitter. An external filter like OggDS is needed.

@vlada
Just click the green book icon to see changelog in sf.
http://sourceforge.net/project/shownotes.php?group_id=65889&release_id=352709

[videomixer9]

there is also no postfiltering done in aac

I meant the filter used on HE-AAC that was already implemented similar in MP3Pro, you e.g. hear the result of this method when you used winamp on aacplus streaming media, for a short moment the sound is damp and then suddenly gets much brighter, I forgot how this filter is called, it's part of the proper decoding but i'd compare it to h264 and deblocker that's controlled at encoding time, as the real work is done by the decoder.

[dimzon]

I forgot how this filter is called, it's part of the proper decoding
This is SBR and PS decoders

[bond]

I meant the filter used on HE-AAC that was already implemented similar in MP3Pro, you e.g. hear the result of this method when you used winamp on aacplus streaming media, for a short moment the sound is damp and then suddenly gets much brighternever heard this, propably a bug in winamps decoder

[Elias]

Is anyone experiencing bugs with MPC anymore?

[SeeMoreDigital]

Is anyone experiencing bugs with MPC anymore?I'm of the opinion that when playing 1280x720 MPEG-4 files with SRT subs in AVI (or MP4), using MPC's VMR7 (renderless) mode and its own MP4 splitter, the video has a tendency to stutter quite a bit ;)

[SeeMoreDigital]

The latest build does not appear to correctly display the streams properties within the MP4 container: -

http://img44.imageshack.us/img44/2046/sinternalsplitter2xp.png


Notice how it lists the first AAC stream as being "stereo", when it is in-fact 6Ch (5.1) "surround sound".


Cheers

[Elias]

The latest build does not appear to correctly display the streams properties within the MP4 container: -

http://img44.imageshack.us/img44/2046/sinternalsplitter2xp.png


Notice how it lists the first AAC stream as being "stereo", when it is in-fact 6Ch (5.1) "surround sound".What have you been smoking? :D

[dimzon]

What have you been smoking? :D
http://img7.imageshack.us/img7/9697/sinternalsplitter2xp4qp.png

[Stebs]

I'm of the opinion that when playing 1280x720 MPEG-4 files with SRT subs in AVI (or MP4), using MPC's VMR7 (renderless) mode and its own MP4 splitter, the video has a tendency to stutter quite a bit ;)
Since some time, VMR7 (renderless) is not as fluid here as VMR9 (renderless).

Mpeg2 Decoding done with Nvidia Pure Video Decoder (has hardware acceleration + hardware pixel adaptive interlacing on my GF 6800GT), VMR7 (renderless) does only do Software accel., VMR9 (renderless) does hardw. accel. but deinterlacing does not work. VMR 7+9 (windowed) work great (accel+ deinterlacing), but do not support Subtitles :(
So for now, I have to constantly change the Output mode, it would be great to have at least some easy way for switching between Output modes (in Context Menu etc.)
BTW, it would also be great if you can load external Audio-files in File-Menu (just as you can do with Subtitles), as currently only external Audio-files with the same name as the Video-file are automatically loaded.

[SeeMoreDigital]

What have you been smoking? :DI think I'll have whatever the guy in dimzon's avatar is having :)

[Elias]

@SeeMoreDigital: I'm sorry, I didn't read the audio channel name, just the number. My fault :)

[Tima]

MPC (at least 6.4.8.4 and higher) hangs up when I try to open tab with stream properties when watching mkv. It hangs using either internal matroska splitter or Haali Media Splitter (all versions).

[celtic_druid]

Stereo doesn't have to refer to 2 channels.

[Rash]

So for now, I have to constantly change the Output mode, it would be great to have at least some easy way for switching between Output modes (in Context Menu etc.)
I know that what I am going to say next is asking a lot. I thought about it before but never posted because I didn't want to push Gabest into even more work. Afterall MPC is great already. :)

But since you mentioned, I think it would be great if we could have an option to make MPC render on renderless mode when it finds a subtitle and on windowed mode when it doesn't.

I am actually using Pixel Shaders for resizing, so I have to use renderless anyway. But I guess that would make a nice option for some people. ;)

[Elias]

Stereo doesn't have to refer to 2 channels.What do you mean? Like Joint-Stereo, Dual Channel and stuff like that? Anyway, I sent Gabest some mp4 samples, and he's working on this issue.

[celtic_druid]

Refering to 2 channels as stereo is something that has become common and we are all used to, but technically stereo doesn't have to be 2 channels, 2 speakers, etc. That is what I mean. So refering to 6 channels as stereo whilst confusing isn't strickly speaking wrong.

Although maybe it does the same for 1 channel? 'Cause that wouldn't be right.

[bond]

according to the iso media specs
ChannelCount is either 1 (mono) or 2 (stereo)

[SeeMoreDigital]

according to the iso media specsDamn those specs again!

"2 channel" audio could refer to two totally unrelated streams. Whereas "Stereo" is the ability to observe a given image/sound from (two) different perspectives or reference points.

[Elias]

New MPC CVS added in the first post. Check it out! :)

[Tima]

New MPC CVS added in the first post. Check it out! :)
My problem with hanging is solved now. Thanks! :)

[Elias]

My problem with hanging is solved now. Thanks! :)How did MPC it hang itself? I wonder because I want to check if it's reproducible.

[lrms]

For me this new version (2005-11-13) still has the .mov regression I found with previous 6.4.8.6 builds. By the way, for those unable to reproduce it:
1) I don't have QuickTime installed
2) The problem only happens with CloneWars chapters 1 to 20 (21 to 26 play the video, as did version 6.4.8.4)

[Elias]

What mov regression? Explain please.

[lrms]

What mov regression? Explain please.

http://forum.doom9.org/showthread.php?p=734371#post734371

[Elias]

Changelog for the latest CVS added in the first post, I will update it more throughout the day if Gabest gives me more information. Checkout the new features.

[bond]

Overall the new version works for me, but with a somewhat higher memory usage.
But I've found a regression regarding MOV playback. With MPC version 6.4.8.4 I could play at least the video track (SQV3) of some MOV files (handled by ffdshow 2005-08-22). With MPC version 6.4.8.6 it will not play the audio nor the video. The MOV files I'm talking about can be found here: http://www.starwars.com/clonewars/microseries/chapter/archive.htmlthe question is what mov parser you use!
do you have qt alternative installed? do you have qt wrapping of mpc enabled? do you have any other mov parser installed (like haali, 3ivx...)? what filters are listed under "filters" when you try to play such a mov? what filters are used when you try to play such a mov in graphedit?

[Pug Crydee]

Playing an 5.1 AC3 file in MPC, downmixed with internal ac3-decoder, the volume is good. Converting this file to 5.1 AAC (inkl pregain), the volume is very low when downmixing to stereo is enabled. Both files are at the same volume when playing back without downmixing.
I guess downmixing means a loss of volume and that the AC3-Decoders applies some gain to the audio afterwards.
Question: Any way to get this for AAC as well??

Thanks
Pug

[lrms]

the question is what mov parser you use!
do you have qt alternative installed? do you have qt wrapping of mpc enabled? do you have any other mov parser installed (like haali, 3ivx...)? what filters are listed under "filters" when you try to play such a mov? what filters are used when you try to play such a mov in graphedit?

Yes, good question! :)
About 3ivx, no, I don't have it.
I do have the Haali Media Splitter installed, but I can say it's not being used in this case (no tray icon, also not listed in Filters).
Loading graphedit only showed me a movfile splitter that doens't support the "Filter Properties...". :(

But using Process Explorer I could see the filters called "NeQTDec.ax" (Nero - QuickTime(tm) Decoder Wrapper) and "NDParser.ax" (Nero Digital / mp4 / avi / mov parser) were loaded when redering one of that files in graphedit.

The strange thing is that using the "Overrides" option in MPC I wasn't able to really block these filters... :mad: But completely unregistering "NDParser.ax" I confirmed it was being used - after that the files wouldn't play at all.

So the regression seems to be about some incompatibility with the Nero QuickTimer parser (from Nero Express OEM - 6.0.0.16 - but file version is filters is 3.02.0000.0020).

[Kurtnoise]

It seems that mpc crashes with the slideshows mp4 samples from IBM ToolKit (http://www.alphaworks.ibm.com/tech/tk4mpeg4/download)...

[thuan]

I have problem with MPC 20051113 and official 6.8.4.6 (don't keep other ver so can't test) with this file http://weedy.1.vg:8396/torrents/hishoku_no_sora_SHANA_version_(DVDRip_Karaoke_H264_720x405).torrent?info_hash=ea984eb2cc7a206b8852b6e86bf11617f22ea2e0
It seem to drop frame at many specific times with any VMR mode. It's not happen with Overlay and Zoom Player in Overlay and VMR so I don't think my com is too slow to play it (I can even play it in Zoom VMR9 and RGB32 output).

By the way I use the newest ffshow from Milan 1115 and my CPU is a Sempron 2200+.

[Stebs]

Tried that file and I had no single frame dropped...
Maybe it IS CPU-related afterall?

[Elias]

By the way I use the newest ffshow from Milan 1115 and my CPU is a Sempron 2200+.http://www.aziendeassociate.it/cd.asp?dir=/ffdshow

Update ffdshow. Milan's latest compiled build is from March or something. It could very well be ffdshow that is the problem here.

[Sharktooth]

milan's latest compile is 20051115... it's not outdated...

[Elias]

milan's latest compile is 20051115... it's not outdated...Well it seems he has been busy the past week :)

[thuan]

Well it seem it's because of my CPU.

[awx]

Here's my list of bugs or improper behaviour from MPC. Maybe I have misunderstood something, maybe these are by design or maybe there is something outside of Gabest's control that prevents him from fixing them.


DVD screen grab gives .BMP with improper aspect ratio when DirectShow Output is set to anything other than VMR9 (windowed). I think the height of the image is correct but the width is off.
command line "mplayerc.exe /dvd" won't work for me without a drive letter. I have two DVD drives, a real DVD-Rom drive and a daemontools drive. I would like this feature to work the same as launching MPC and selecting "Open DVD..." from the File menu. That is to say, I'd prefer it so that if no drive is specified, then MPC searches for the first available DVD disc. This works well from the File menu but not at all from the command line for me. I need this because I launch MPC with a remote in a HTPC environment and so I have to have an extra button on the remote call the "Open DVD..." function after MPC is launched.
In full screen mode for video playback, the Seek/Control/etc. bars will not auto hide if they are activated by keyboard press when the option "Show controls in fullscreen for" zero seconds (auto-hide) is set. Suppose this option is set to zero and you are playing a video in full screen mode. Using the keyboard (or a remote control), you tell MPC to toggle one or more of these bars so that they appear on screen. They should then auto-hide themselves but they do not. They stay on screen until the mouse is moved. Perhaps this is by design, I don't know. It did not seem like expected behaviour to me. Setting the option to a fixed number such as 4 works as expected. EDIT: what I'm really after here is a function to briefly display the currently active bars on screen for a few seconds then have them disappear again. If I should explain this more, please let me know.
MPC won't launch as a maximized window.