Computer Audio Players

So after reading Mike Moffat's latest windows driver rant and comments about audibly better sq on linux and mac os, and feeling a little annoyed with the occasional blips and stutters that seem to be a persistent feature of win10 with usb2 audio devices, I installed elementary linux (dual boot) on my desktop last night. https://elementary.io/

I know I've heard many of you say it, but I was in denial. My god - not only is this a great looking, free, fast and stable OS, it really does sound much better as a front end for my Yggdrasil -> mainline -> hd800's than win10/foobar. I'm currently running clementine as a player (with alsa output) and looking forward to experimenting with various others. Even spotify sounds noticeably better in linux than windows.

Using a terminal window to install programs again feels a little odd in a x-windows environment in late-2015, but on the whole, with sound quality factored in, I think I am pretty much a linux convert.

 

Why use anything but MPD in Linux? Unless you really want that album art, but you can even get that working with some fiddling and the right client

 

Were you using directsound or outputting through wasapi when using windows?

 

It depends on what you want to do... I actually use DeaDBeeF which I think is an excellent audio player for Linux. I have heard good things about MPD too, but I haven't used it in a long time. Nonetheless, you still typically use some kind of client (text or GUI) to command MPD to do things. Instead of having to configure and setup multiple components (MPD and its client), I just use a regular audio player. Then, I use a VNC app on my phone to control the music player remotely or anything else on my desktop. It works really well for me anyway. Also, since my computer is a laptop, I can use the same audio player no matter what it is attached to with a minimal amount of reconfiguration. For instance, I could take it to a meet and just use the same setup attached to anyone's USB interface. The only thing I also need to do is use the alsamixer program to make sure the USB device is at the 0.0dB volume level (not all devices require this though). As a side note, if you go back through this thread I believe there are at least a few if not several of us using the DeaDBeeF audio player. MPD is typically used more in HTPC type configurations, and I usually hear various other players used in more typical desktop scenarios. In addition to just album art, you can do a variety of things on the fly if needed such as Sample Rate Conversion (maybe a DAC doesn't support a specific sample rate of your file), Equalization, display lyrics, scrobble your data, or anything else you really want including sending the data to a Jack interface for a professional audio tool to process the audio stream. Oddly enough, I rarely use those features, but they are available. I don't remember all of those things being available in MPD, but it has been a while since I looked at different Linux audio players perhaps they have added some enhancements that I was not aware of.

 

Building on Indy's comment, what's the benefit of using MPD over Clementine if I am just listening to local files? I'm interested in possible alternatives, but Clementine was the most widely recommended audio player on all the Elementary blogs and support sites I read. It takes about 6 clicks to load your library, set to alsa/bitperfect output and enjoy. Interface could be prettier, but general ease of use and the clear improvement with Yggdrasil over Foobar is what matters most to me. Also planning to give DeaDBeef and Thomahawk a go.

I've been running wasapi, ks or asio in foobar since 2008. I suffered the indignity of cplay's minimalist/nonexistent interface for a while too but decided there was no benefit. Windows with foobar was always fine for me with any number of usb1 dacs, but in win7, and to a greater degree in win10, the lack of native usb2 support and obvious issues with the Yggdrasil driver are not acceptable - particularly when compared directly to linux.

Enjoying LFF's recommended pressing of Billie Holiday's Songs for Dinstingue Lovers right now, and my rig has never sounded better.

 

I should also probably mention that DeaDBeeF for the most part is on minimalist side. It's not the most beautiful player, but having tons and tons of features can just cause problems. Why? Well, all of those features just add complexity which can introduce more bugs (read: add audible artifacts potentially). DeaDBeeF is quite extensible through plugins though, and it comes with a lot of them to do various things right out of the box. Usually on various Linux web sites that I have seen, largely the people going from Foobar2000 on Windows to Linux have gone to DeaDBeeF. The main complaint is some missing features. For me, I just want to listen to music, and I want a player that uses hardly any CPU resources (quick, efficient, and responsive), communicates directly with ALSA to bypass pulseaudio's typical sample rate conversion and sound mixing (be bit accurate), and be easy to use including minimal configuration needed. To that end, DeaDBeeF is brilliant. If you are looking for something that is very snazzy, DeaDBeeF probably isn't your first choice...

 

I've been using MusicBee as a replacement for foobar for a few weeks now and I think I'm going to be keeping it. It has all of the features I used to use in foobar without installing additional components, and the UI is much slicker IMO.

 

If you're going to go to the trouble of installing *nix just for audio, why wreck it with a GUI? That sounds like quality-degrading cruft if you ask me. Stick with a lightweight CLI player.

 

It's extremely unlikely that a GUI could degrade the audio.

For example: mpd configured to output direct to hardware (no mixer, no resampler). What possible difference can there be between controlling the mpd server via local GUI, local commands i.e. mpc, local console interface i.e. ncmpc, keyboard shortcuts, or via remote client?

For the local GUI to make any difference you would have to be running your music server on something with CPU speed that hasn't been seen in personal computers for well over a decade. My iRiver H140 has a 120 MHz with 32Mb RAM and it decodes FLAC, MP3, Ogg Vorbis etc. and processes the audio through a BS2B/Meier type crossfeed without any issues. No skips, no pauses. While doing so it also keeps its database auto updated. I could also play games on it but 120x240 monochrome screens just don't do it for me these days. The idea that any kind of modern PC hardware suffers an audio crisis by rendering a GUI is just laughable, especially when one considers this is likely to be handled by a separate GPU.

My desktop is just an old Dell Core i3 with 8GB RAM but it can run a full graphical desktop, ffmpeg converting HD video with x264 on all logical cores not in use, a web browser, email client, rss reader, wine, and mpd outputting bit perfect audio to an external dac, and all this simultaneously without missing a beat.

People thinking they need special realtime kernels or some audiophile OS should probably think again. All you really need is to get the audio decoded and out to the amp without it being resampled or passed through the OS mixer. It's actually easy to do on any OS that lets you make a few simple choices, whether that is Linux based or Microsoft based or something else.

MPD is what I use on my server and my desktop because it lets me have bit perfect audio and a desktop GUI client (GMPC) that can control any accessible client and which displays *all* the metadata such as the composer and comment fields. Also I can both control it from my smartphone and listen to its streams when away from home (FLAC if I have wifi or reliable 4G, otherwise Ogg Vorbis).

The presence, or not, of a GUI is really not an audio issue.

 

Welcome fellow greybeard.

 

I can't say I disagree. If anything I expect a realtime kernel would cause more glitches rather than less, assuming it had any effect at all.

That said, less is more. I find the idea of an ultra efficient environment for a dedicated audio box appealing. But my favorite editor is Vim, so my perspective is probably skewed.

I'm neither gray nor bearded, but I appreciate the sentiment.

 

Because who wants to run just a CLI anymore? All of my Linux boxes are multi-purpose to some degree, and Linux is really very lightweight including the GUI especially in comparison to almost anything Windows these days. In fact, I have a full Gnome GUI based Linux running on a 9 year old laptop with a Celeron processor without any problems, and playing music through a GUI based music player on it only uses 2-5% CPU utilization. Even Ubuntu with the standard GUI which has gotten a bit bloated these days which I don't run any more, it still is a lot faster than Windows. Nonetheless, the kernel and ALSA is going to be the same regardless if you use a GUI. You still can run MPD or whatever CLI based player that floats your boat even as a daemon without the GUI doing much of anything if you want, but you can than still use a browser to surf the web, edit documents, and etc. in a modern way when you need to. I still do plenty in the CLI, and I even still develop CLI only applications from time to time. Nonetheless, I just can't go back to CLI only as it provides very little flexibility with today's applications.

 

Me, if I'm running a dedicating audio playing box thing. We live in a world where audio players sound different, after all.

I agree, running CLI only for a machine required to handle multiple roles would be annoying.

 

Non sequitur, even if true. We live in a world which is no longer the world of Windows 3.1, or even W95. We live in a world where even Windows will do more than one file copy at a time!

My hangover from those early days is that I still like a music player to be simple, and to avoid doing anything more graphic than, say, display a thumbnail of an album cover --- but I'm sort-of aware that it is a hangover, rather than a true reflection of what today's computers are capable of.

I am not saying that nothing can ever go wrong (sound of mouse moving across screen, anyone? No thanks!) even in the world of *nix, which was fully able to multi-task for multiple users before Gates even started. Which might not be quite true, but a quick bit of hyperbole never hurt :lol: so let be say before Gates was even born

Audio is relatively trivial. You do not need a dedicated machine, let alone the two-machine setup some audiophiles insist on.

(This is in no way telling anybody else what they should or shouldn't do. We do things for all sorts of reasons, including... just because we want to. No problems.)

 

I am not entirely sure I agree that audio players "sound different." If you don't setup things correctly or know what to setup, then yes they can sound drastically different. The biggest problem with audio in the Linux realm is pulseaudio. It does the same things that Windows does in that it resamples and mixes all audio at a specific sample rate so that multiple applications can all do audio at the same time. In Windows, you typically use an ASIO driver to bypass all of that. In Linux, the easiest way to bypass that is to make sure you use an audio player that can communicate directly with the underlying sound system which is ALSA. Typically, the music player defaults for ALSA are ok, but I usually disable resampling and have the output go to a SPDIF software interface. In addition, some DACs use an USB receiver chip that allows you to adjust sound levels. If that is the case, you need to go into the CLI and run the alsamixer application to set the level to -0dB for that specific USB device. If you setup those things, then all audio players should sound the same under Linux. The only way they would not is if you are using other software layers such as EQ or there are bugs in the software that makes it not bit perfect out the interface. From an audio player perspective, I agree that simpler is better in that you have a much better chance of fewer bugs that might effect the audio stream.

 

Certainly good advice. I agree with your thoughts re: Linux. My forays led me to Alsa and Deadbeef with quite good results.

What baffles me, however, is the numerous players I have used with OSX that all sounded just slightly different. One could say that if they all go through Core Audio they should all sound the same. Yet, many do and many also run another program like itunes in the background. Perhaps these programs add to the electrical noise on the ground plane and affect what I hear.

That is my explanation for it. Even the dacs we use have some choices, like the bridge that allows me to choose between a standard USB compliant driver (no install necessary) labeled "Core Audio" And another one that needs to be loaded labeled "bit Perfect". I have no clue at what level these devices are talking to the computer or bridge but they do sound slightly different.

At the end of the day, I am happy with the players I use in both my Mac and Win setups. Linux got the boot, even though it sounded good, just too much tinkering.

 

I'm still having a lot of fun playing with elementary linux and, upon further back and forth between OS's, remain convinced that it sounds better than win10. Deadbeef and Clementine, among other players that allow you to select a direct alsa output, work nicely and light up the appropriate sample rates on the front of my Yggdrasil. Deadbeef is particularly nice in this regard because, like foobar, it also displays the bit depth and sample rate in the bottom of the window.

What has me a bit perplexed though is that some of the prettier players I've tried (which are clearly focusing on UI/UX rather than the audiophile market and don't allow you to specify an alsa output - e.g. tomahawk and lollypop), still appear to be outputting at 44.1k (based on the Yggdrasil's front panel). Shouldn't I be seeing 48k if the sound is going through pulseaudio?

 

It should, but here is the thing as bixby alluded to which is there is a lot of tinkering when Linux is involved. I have not used the Elementary Linux distribution, but there are settings and defaults for the player, pulseaudio, and ALSA. In some distributions, they bundle defaults that are different than the developers. Debian, Ubuntu, and other distributions do this depending on the package which means they might have different settings. You would need to look at these settings. Perhaps the pulseaudio in your distribution resamples everything at 44.1kHz. This is quite possible, and you would have to look at your specific configuration to know.