LAV Filters - DirectShow Media Splitter and Decoders

[huhn]

Originally Posted by Arm3nian

If a system has problems with bass when a LFE channel is not present then that system is a piece of crap...

Almost all content, besides the audio on bluray do not have an LFE channel. So when listening to music or watching videos on youtube there would no bass... but of course there is. The bass is either reproduced by the left and right speakers or fed into to the subwoofer at the crossover frequency. The purpose of an LFE channel is to isolate specific sound effects from the other sounds, so the subwoofer reproduces them instead of the other speakers (since it does a better job). LFE can contain frequencies up to 120hz. An LFE channel can allow you have a higher crossover frequency if your main speakers cannot reproduce low frequency sounds. Headphones do not need this feature as they can reproduce the audible range of tones, but some speakers might. Quality systems focus on a specific frequency response for each speaker.

Basically, any quality headphone has 2 drivers, you should know this. What hifi gear have you seen with 8 drivers that costs $80. That makes no sense when a proper headphone with 2 drivers costs more. Any decent headphone can reproduce bass without a dedicated subwoofer driver/cone. Notice how standalone subwoofers have huge cones? You need them to produce the low frequencies at audible volumes. A half inch subwoofer inside a headphone isn't going to provide much benefit compared to the main drivers.

The surround sound aspect does not make much sense either. Notice how a proper surround system has the surround sound speakers on the opposite sides of the other speakers? The auditory system is good in positioning, but it doesn't make up for the fact that the surround speakers are right next to the other ones in a headphone. Any proper headphone will have a good enough sound stage and imaging capabilities to mimic what multiple drivers do in such a confined space.

tldr: Why should software support badly designed hardware that sells purely on marketing?

i don't use hardware like this and i have personally no need for such a feature with my beyerdynamics T5p...

that doesn't change the fact that people buy logitech speaker on mass. and a system without digital processing can't do the cross over by it self so a cross over is needed.

if you run an sub with an analog connection without digital processing a crossover is needed it's that simple. and this can happen with a more expensive system too.

luckly some soundcards can do this be them self.

nevcairiel clearly said this is not going to happen so no need to discuss about this here.

[nevcairiel]

The next nightly of LAV will feature the libdcadec DTS decoder.
It supports full bitexact DTS-HD decoding of up to 96/24, as well as much broader support for DTS extensions than the ffmpeg decoder.

Most prominently (things the ffmpeg decoder didn't do):
- DTS 96/24 (96 kHz support for lossy DTS streams, and DTS-HD HRA)
- Bitexact DTS-HD MA up to 7.1, 96 kHz, 24-bit

The only missing features I'm aware of right now is support for 192kHz DTS-HD MA, which I hope will be added soon, and DTS Express, which may take a bit longer (but is also much less important).

[kalston]

Awesome! Can't wait to try it.

[FireFreak111]

Good to hear its coming along rapidly. Can't wait to ditch Arcsoft, use something that isn't a black box, and is guaranteed to work as it should.

192khz is important but thankfully its not used much, and I only heard of DTS Express last week.

Will the data of the DTS-HD MA stream be properly reported outside of the Status tab now? Is that a thing? Or will it still report 5.1 for 7.1 streams?

[nevcairiel]

The splitter will also properly report stream properties now.
One note, the current version will fail to decode 192kHz streams entirely (ie. not even the core). I thought about adding a fallback to core decoding, but I decided to simply wait until the decoder supports 192kHz. If it takes too long, I can still add the fallback later.

[kalston]

I don't think I have any 192khz DTS files anyway But really this is great, even more so knowing that the splitter will properly report it too

[nevcairiel]

They are very rare. Mostly used for a few classical music Blu-rays.

[Snowknight26]

Originally Posted by nevcairiel

Most prominently (things the ffmpeg decoder didn't do)

Which is funny because ffmpeg uses a libdcadec wrapper so it should have the same capabilities.

[nevcairiel]

You do realize that I wrote this wrapper, don't you? :P And it only exist since last week, because thats when I submitted it to ffmpeg, so that LAV (any anyone else) can use it transparently without needing extra code.
And when someone refers to "the ffmpeg decoder", it generally always refers to the native decoder present inside ffmpeg, ie. ffdca or however people like to call it, not an external library that is routed through ffmpeg. If I wanted to discuss the capabilities of an external library, I would call it by name.

[Liandri]

Originally Posted by Arm3nian

And to be clear, I'm going completely off what you're posting. I don't know how your headset performs, but it seems to me it's causing more trouble than it's worth. You're losing the benefit of the extra speakers for movies, and the bass is weak in content without a dedicated LFE stream. The configuration in your player is basically doing what ffdshow/ac3filter would do, which is either downmixing to 2.1 or upmixing 2.0 to 2.1 for the headset to use the subwoofer/manage the bass properly.

I'll try harder next time with 5.1 content.
But also I just currently watch mostly stereo content, so there is little reason to mix into 5.1, also LFE still doesn't work in LAV for that case. Seeing my messages, you can tell my request was somewhat incorrect - what I effectively wanted was LFE redirection, not just 2.1 option. I'm not sure if that can be considered upmixing.

I learned that AC3filter can do this (tested and it works, but somehow I found it performing worse in other matters so maybe I need to tweak more settings). Found a good set of information about LFE on its homepage too. From what I understand, without a dedicated LFE stream everything will get processed through usual speakers, but if the source itself has low frequencies outside of speakers' limitations, it's a good idea to redirect those if there is another speaker that can handle those (sw).

[Arm3nian]

Originally Posted by huhn

i don't use hardware like this and i have personally no need for such a feature with my beyerdynamics T5p...

that doesn't change the fact that people buy logitech speaker on mass. and a system without digital processing can't do the cross over by it self so a cross over is needed.

if you run an sub with an analog connection without digital processing a crossover is needed it's that simple. and this can happen with a more expensive system too.

luckly some soundcards can do this be them self.

nevcairiel clearly said this is not going to happen so no need to discuss about this here.

What kind of system uses only a direct analog connection from a PC? Sound cards provide less than a watt of power, my sub requires 100+ watts. Yes there are lower powered subs but we are talking order of magnitudes. The sub has to have an amplifier, which will also drive the other 2 speakers, so just put low pass filter in there. No proper system would REQUIRE a dedicated subwoofer channel to deliver bass, even though it might improve it. Not everything you listen to is through mpchc . A software mixer would not solve the problem.

Originally Posted by Liandri

I'll try harder next time with 5.1 content.
But also I just currently watch mostly stereo content, so there is little reason to mix into 5.1, also LFE still doesn't work in LAV for that case. Seeing my messages, you can tell my request was somewhat incorrect - what I effectively wanted was LFE redirection, not just 2.1 option. I'm not sure if that can be considered upmixing.

I learned that AC3filter can do this (tested and it works, but somehow I found it performing worse in other matters so maybe I need to tweak more settings). Found a good set of information about LFE on its homepage too. From what I understand, without a dedicated LFE stream everything will get processed through usual speakers, but if the source itself has low frequencies outside of speakers' limitations, it's a good idea to redirect those if there is another speaker that can handle those (sw).

Everything without a dedicated LFE stream will be processed through the left and right IF there is no passive low pass filter and nothing is done in software. You can test if your headset has this by seing if the subwoofer driver does something when no LFE stream is present but your content has deep bass. I mentioned why LFE is useful in my previous post.

As mentioned lav is a decoder, but also provides common options that are needed for optimal playback. Requiring an LFE channel for bass is rare. If there is no bass without an LFE stream then that's obviously a problem. The actual LFE content will be hard to reproduce on such a tiny driver/cone, so what you might be experiencing is the sub working vs not working with an LFE channel as all bass is redirected there.

[huhn]

Originally Posted by Arm3nian

What kind of system uses only a direct analog connection from a PC? Sound cards provide less than a watt of power, my sub requires 100+ watts. Yes there are lower powered subs but we are talking order of magnitudes. The sub has to have an amplifier, which will also drive the other 2 speakers, so just put low pass filter in there. No proper system would REQUIRE a dedicated subwoofer channel to deliver bass, even though it might improve it. Not everything you listen to is through mpchc . A software mixer would not solve the problem.

there are analog running amps out there yep not everything is digital to day how can a system without full range speaker redirect the LFE to the sub without digital processing.
and what kind of system? maybe the these headphones that started this hole topic to start with?

not all people can buy a system with full range speaker so sorry they can't get a proper system... just ignore them i guess. all soundcards should remove there LFE redirection/LFE crossover there is no proper system that need these option any way, so why are they even there?

[Arm3nian]

Originally Posted by huhn

there are analog running amps out there yep not everything is digital to day how can a system without full range speaker redirect the LFE to the sub without digital processing.
and what kind of system? maybe the these headphones that started this hole topic to start with?

not all people can buy a system with full range speaker so sorry they can't get a proper system... just ignore them i guess. all soundcards should remove there LFE redirection/LFE crossover there is no proper system that need these option any way, so why are they even there?

That was my point, you don't need software to do the crossover and processing. Hardware does it digitally or by analog, and hardware should do it because not every program that outputs audio has the capabilities of lav or all the direct show filters. A low pass filter or simple DSP chip can do it for pennies, and all systems that have a sub should have one built in.

My home theater speakers aren't full range either, the front speakers start at 80hz and the center starts at 120hz, but my receiver has the proper hardware to send bass at the crossover frequency that is not found in the LFE stream to the sub. You don't need a receiver for this capability. What is the point of LFE redirection if the system has proper bass management by sending frequencies at a certain point to the sub. Its the same thing. Full range speakers don't need a sub, that's why they're full range, and non full range speakers already have the capability for content with no LFE stream. If there is no proper bass management to begin with, then content with no LFE stream will have the bass produced by the other speakers, making it sound terrible.

[huhn]

most soundcards can't do that that's why software is needed.

and a lot of AVR run in problem when used analog because they don't know if that input signal is 5.1 or 2.0 and not doing blindly a LFE crossover.

and true multi channel non virtual sourround 5.1 headsets can't do this at all they need a signal with an extra LFE channel to sound properly.

[e-t172]

Originally Posted by huhn

most soundcards can't do that that's why software is needed.

The problem is, what nev is trying to do is to avoid feature bloat. Part of that is making sure that the only features that are added are features that make sense to have in a decoder. Bass redirection is not the decoder's job - in fact, one might argue that even the mixing options in LAV Audio are already out of scope. Normally, bass redirection should be done as close to the speakers as possible - preferably in the AVR, or in the sound card if there's no AVR, or in a system-wide audio effect (sAPO) if the hardware doesn't support it, or at the extreme, in the DirectShow audio renderer (not decoder).

The reason why it's not a good idea to move this processing upstream is because the more upstream you move it, the fewer sources benefit from it. If you do it in the AVR, then it's enabled for all sources connected to the AVR, not just the PC. If it's done in the soundcard, it's enabled for all applications, including WASAPI/ASIO etc. If it's done in an sAPO, it's enabled for all "normal" applications. But if you do it in DirectShow, then only your player benefits. That's why it doesn't make sense to have this stuff in the DirectShow pipeline. This functionality should be provided by the audio equipment downstream of the DirectShow player, and if it doesn't, then you should change your equipment for something that does.

[huhn]

Originally Posted by e-t172

The problem is, what nev is trying to do is to avoid feature bloat. Part of that is making sure that the only features that are added are features that make sense to have in a decoder. Bass redirection is not the decoder's job - in fact, one might argue that even the mixing options in LAV Audio are already out of scope. Normally, bass redirection should be done as close to the speakers as possible - preferably in the AVR, or in the sound card if there's no AVR, or in a system-wide audio effect (sAPO) if the hardware doesn't support it, or at the extreme, in the DirectShow audio renderer (not decoder).

The reason why it's not a good idea to move this processing upstream is because the more upstream you move it, the fewer sources benefit from it. If you do it in the AVR, then it's enabled for all sources connected to the AVR, not just the PC. If it's done in the soundcard, it's enabled for all applications, including WASAPI/ASIO etc. If it's done in an sAPO, it's enabled for all "normal" applications. But if you do it in DirectShow, then only your player benefits. That's why it doesn't make sense to have this stuff in the DirectShow pipeline. This functionality should be provided by the audio equipment downstream of the DirectShow player, and if it doesn't, then you should change your equipment for something that does.

i'm not trying to change nev mind at all. there are audio processor that can do this and this is fine. i don't even have a use for this personally.

replacing hardware is not an option for everyone this still cost money.

[STaRGaZeR]

Originally Posted by e-t172

That's why it doesn't make sense to have this stuff in the DirectShow pipeline. This functionality should be provided by the audio equipment downstream of the DirectShow player, and if it doesn't, then you should change your equipment for something that does.

And you have the guts to tell other people they are ridiculous? Talk about being delusional

[Arm3nian]

He is correct that the equipment should do the bass management, for two reasons.

1. Your computer is not the only thing connected to speakers. Cable/satellite box, bluray player, basically anything that isn't a computer or professional gear doesn't have capabilities like down mixing or up mixing to 2.1, only traditional formats like stereo or 5.1.
2. You can't use lav or directshow filters with every program. Watching videos on the internet, games, applications, might not support 2.1.

Down mixing to stereo (headphones) exists in pretty much everything, and all standalone speakers or surround configurations require an amplifier which usually also has bass management through passive or active filters. If you want to design a system that doesn't follow the common configurations, you better make sure it works on its own, and does not rely on 3rd party help to make it function.

If you can't get a dedicated LFE stream working, then just set the crossover to 120hz. If you can't set it then hope the default filter is set to a figure that allows your sub to reproduce most of the LFE content, it most likely is, and the system itself is most likely optimized to reproduce the LFE content without a dedicated LFE stream. If your subwoofer doesn't even work without an LFE stream and your other speakers can't deliver bass then you've made the wrong purchase. Even if software supports upmixing or downmixing to 2.1, everything else will sound terrible if the system doesn't already have bass management.

[JohnLai]

@nevcairiel

Recently, I have been using LAVvf as Directshow source filter for Staxrip.
However, whenever DXVA Native is selected, it will fall back to software decoding. DXVA Copy-back is working, but there is performance penalty in memory copyback.
Any possible solution to make the dxva native decoding works with avisynth/Staxrip?

Note: NikoD and stax76 suggested me to ask in this thread XD.

[Asmodian]

Originally Posted by JohnLai

@nevcairiel

Recently, I have been using LAVvf as Directshow source filter for Staxrip.
However, whenever DXVA Native is selected, it will fall back to software decoding. DXVA Copy-back is working, but there is performance penalty in memory copyback.
Any possible solution to make the dxva native decoding works with avisynth/Staxrip?

Note: NikoD and stax76 suggested me to ask in this thread XD.

DXVA native leaves the decoded video in the GPU's memory. Getting it into the system memory is the "copy-back". DXVA Native cannot work with Avisynth/Staxrip.