Announcement

Collapse
No announcement yet.

Even With AMDGPU DC, HDMI/DP Audio Isn't Working Out For All Radeon Linux Users

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    Originally posted by direx View Post
    Are you referring to the sound registers? From Takashi Iwais post on the Kernel Bugzilla (https://bugzilla.kernel.org/show_bug.cgi?id=196637) it looks like this is not an ALSA issue. Intel also has HBR support in place for many years now, so this is not a general driver infrastructure thing. What do you mean by "Anyone that wants to can take a look at it"? You mean AMD folks, right? Because I've tried helping with this on the FDO bug report with some printks, but things did not really lead anywhere...
    The code is open source. Anyone can hack on it. If there is a bug that is bothering you or a feature you want to implement, the source code is available and you have direct access to the engineers. That is not possible with other OSes. That is kind of the whole point of Linux; the community. Rising seas lift all boats and stuff.

    It should also be noted that just because you happen to be experiencing a bug, it doesn't mean it automatically affects everyone else.

    Comment


    • #22
      Originally posted by agd5f View Post
      It should also be noted that just because you happen to be experiencing a bug, it doesn't mean it automatically affects everyone else.
      So you are saying that *bitstreaming* HBR audio is actually working? I have not read a single report about that, just failure reports here and on the Kodi forums. Did anyone here succeed with bitstreaming HBR audio on a recent AMD GPU so that your AV receiver actually displays "Dolby TrueHD"/"DTS-HD Master Audio"/"Dolby Atmos"? If your AVR only shows "DTS" or "Multi-Channel In" then you are *not* bitsteaming HBR audio.

      Originally posted by agd5f View Post
      The code is open source. Anyone can hack on it. If there is a bug that is bothering you or a feature you want to implement, the source code is available and you have direct access to the engineers.
      You are basically saying "write your own driver" and I must say that I find this attitude not very customer-friendly. I understand that you guys cannot deal with everything at once but the hardware has been around for more than a year now and nothing has happened regarding HBR audio. As I pointed out earlier I actually do know a few things about software development in general and I have also tried to provide some debug messages. I am however not capable of actually fixing the AMD Linux driver.

      BTW: In the company I work for we are very happy about customers who are willing to help debug issues with our products. There are very few who are actually able to debug things for us and I would never dare to tell them "fix our product yourself, we'll hand out source". I'd rather say "thanks for helping us, let's fix this together". Please don't take this personal, but I think the "let the customers fix this" is definitely the wrong approach - whether it is open source or not.




      Comment


      • #23
        Originally posted by direx View Post
        So you are saying that *bitstreaming* HBR audio is actually working? I have not read a single report about that, just failure reports here and on the Kodi forums. Did anyone here succeed with bitstreaming HBR audio on a recent AMD GPU so that your AV receiver actually displays "Dolby TrueHD"/"DTS-HD Master Audio"/"Dolby Atmos"? If your AVR only shows "DTS" or "Multi-Channel In" then you are *not* bitsteaming HBR audio.
        I'm talking in generalities. For example, if suspend and resume doesn't work on your particular laptop, it doesn't mean suspend and resume is broken on all laptops. Some people have reported problems with HDMI audio, for others it appears to work fine. I'm not sure of your particular issue. My point is just don't assume it's broken for everyone just because something is broken for you.


        Originally posted by direx View Post
        You are basically saying "write your own driver" and I must say that I find this attitude not very customer-friendly. I understand that you guys cannot deal with everything at once but the hardware has been around for more than a year now and nothing has happened regarding HBR audio. As I pointed out earlier I actually do know a few things about software development in general and I have also tried to provide some debug messages. I am however not capable of actually fixing the AMD Linux driver.

        BTW: In the company I work for we are very happy about customers who are willing to help debug issues with our products. There are very few who are actually able to debug things for us and I would never dare to tell them "fix our product yourself, we'll hand out source". I'd rather say "thanks for helping us, let's fix this together". Please don't take this personal, but I think the "let the customers fix this" is definitely the wrong approach - whether it is open source or not.
        I was not saying you need to fix it yourself. I was replying to another post that asked what users could do to look into the issue themselves if they wanted to.

        Taking a step back, we need to keep things in perspective. My whole team wants to produce the best experience possible on Linux, but the reality is, Linux has a relatively low desktop market share so we don't have the same level of resources relative to other OSes for desktop features. Most Linux customers tend to be in the server, embedded, and workstation markets so most of the feature work focuses on those markets. One of the things that has made Linux as popular is the fact that is it open source. You don't have to wait for the vendor to implement whatever feature you want. The code and register information is there. I used to hear the argument all the time, "if the hw vendor would just release the register specs or source code, we the community could implement everything we want." In practice, I rarely see this. In general, it seems like most Linux users seem to just want the windows model. Vendor provides the drivers in a nice package.

        Comment


        • #24
          Originally posted by Michael
          Currently only the legacy HDMI audio formats appear supported of PCM and Dolby Digital/DTS. But support for modern audio formats like Dolby TrueHD / DTS-HD MA / Dolby Atmos and others are not supported
          LOL, since when is PCM legacy?

          Comment


          • #25
            Originally posted by direx View Post
            So you are saying that *bitstreaming* HBR audio is actually working? I have not read a single report about that, just failure reports here and on the Kodi forums. Did anyone here succeed with bitstreaming HBR audio on a recent AMD GPU so that your AV receiver actually displays "Dolby TrueHD"/"DTS-HD Master Audio"/"Dolby Atmos"? If your AVR only shows "DTS" or "Multi-Channel In" then you are *not* bitsteaming HBR audio.
            [snip]
            As I pointed out earlier I actually do know a few things about software development in general and I have also tried to provide some debug messages.
            Well, as someone who is also not so familiar with Linux kernel drivers (but I've written commercial NT kernel drivers), and someone familiar with quite a few media stream formats (but not intimately familiar with the ones you mention)... are you sure your source is even decoding them and presenting them to the driver? Because there's a whole chain of responsibility and a little format negotiation dance that goes on in the guts of every media playback stack, be it Kodi or whatever. Does the hardware advertise the right features in the right order? Does the playback negotiate correctly for the supported features? Does the stream decoder pull them out properly...

            None of that _except_ advertising the connected hardware features and transporting the particular stream format has anything to do with AMDs driver. You being the person with the right hardware (AV receiver, graphics card) and media (with the streams) are in exactly the right place to offer useful diagnostic information to the appropriate software team by using your expertise to pin down who's responsibility it is... squeaky wheel gets the oil and all that.

            Comment


            • #26
              Originally posted by pq1930562 View Post
              LOL, since when is PCM legacy?
              Probably ever since more advanced formats like Atmos are available, which offer much superior possibilites (like object-based sound).

              Originally posted by Happy Heyoka View Post
              None of that _except_ advertising the connected hardware features and transporting the particular stream format has anything to do with AMDs driver. You being the person with the right hardware (AV receiver, graphics card) and media (with the streams) are in exactly the right place to offer useful diagnostic information to the appropriate software team by using your expertise to pin down who's responsibility it is... squeaky wheel gets the oil and all that.
              That's exactly what I have tried. I am perfectly fine with testing stuff, adding debug messages to the right places and compiling my own kernel. But I cannot do this entirely on my own. See the bug report on FDO.

              And by the way: There is also a bug report about this on the ALSA-side. Takashi Iwai claimed that this is an AMD issue. While it is not 100% certain yet where the actual problem is I do hope that we are not playing ping-pong here.


              Comment


              • #27
                Originally posted by agd5f View Post

                One of the things that has made Linux as popular is the fact that is it open source. You don't have to wait for the vendor to implement whatever feature you want. The code and register information is there. I used to hear the argument all the time, "if the hw vendor would just release the register specs or source code, we the community could implement everything we want." In practice, I rarely see this. In general, it seems like most Linux users seem to just want the windows model. Vendor provides the drivers in a nice package.
                Well, what did you expect?

                Heck, i see fools and idiots out here asking for PPAs. Of upstream kernels. Which is like, one of the easiest damn things to build and install from source in Linux. And their common excuse? "I'm not a developer! Don't you dare expect me to compile my own software just to use it or I'll drop Linux!"

                Guess what? I'm no developer but I build my own drivers and kernels from upstream sources.

                I bet 90% of Linux users today are just hipster wannabes who think its cool to hate Windows and use Linux to 'stick it to Microsoft and Apple'.

                Comment


                • #28
                  Originally posted by endrebjorsvik View Post
                  Those speakers were good stuff 15-20 years ago. Today the rubber has probably become brittle from age, and is visibly cracking when you move the elements. At least my DM603 S2 are. And technology has evolved as well. I have transitioned to active studio monitors and I am very happy with M-Audio BX5 carbon.
                  Some people are happy with 30+ year old speakers, I don't know if stuff ages bad but at least it doesn't really get obsolete, the same way a 100-year-old table or chair is fine.
                  The bigger (and possibly unfortunate) issue is often room accoustics or size.

                  Comment


                  • #29
                    Just for the record: somebody came up with a patch for this sound issue just yesterday. I have tried it and even the current/more advanced audio formats are finally working now!

                    @debianxfce: I am not going to feed any trolls

                    Comment


                    • #30
                      Originally posted by debianxfce View Post
                      A couple of notes inspired by your writing. Do not shout in the internet, do not blame different opinions as trolls, show the WWW link the magical patch so that others can use it, list the titles that have multi channel audio sound effect tracks.
                      The patch is in the bug report that this Phoronix article is about:



                      And I cannot list thousands of titles here. To give you an overview, this is a list of movies with 7.1 tracks:

                      http://www.blu-ray.com/movies/search...rtby=relevance

                      -> 2393 MATCHES

                      Movies with 5.1 tracks:

                      -> 17241 MATCHES

                      Of course these numbers only include the bigger international movie productions, which are in that database.

                      Originally posted by debianxfce
                      Very few movies use multi channel audio effects, only star wars and a couple other action movies uses multi channel effect mixing.
                      [...]
                      So multi channel audio system is useless at home.
                      When saying that the numbers above are only very few movies and that multi channel audio is useless then this is pure trolling. Nuff said.
                      Last edited by direx; 06 March 2018, 01:20 PM.

                      Comment

                      Working...
                      X