I have some new headphones, they really are bobby dazzlers

I have been listening to music on my two setups VERY carefully and I reckon that there are differences in the sound quality/richness/feel between the two engines I use, GStreamer and xine.

Which engines do you think gives the richest/best sound?

xine
GStreamer
aRts
none

Interest and opinion only

Even thoug i didn't write or even take a look at the code of any of these programs, it doesn't make sense that they sound differently.

MP3, for example, is decoded the same way, no matter wich program you use (the codec *should* be the same.) The rest is only about how the sound is transferred to the sound card, which the system's drivers should take care of, not the playback software. Maybe the equalizer settings are somewhat set differently....

Choose the best software on the most features, ease of use, or simply on how well it looks, as the sound is the same.

There are many factors that come into play here that can make audio sound different between engines. Wait for foo_bar_foo to come along, I've seen some interesting posts from him on this subject that should help you understand a bit better, sorry I can't help more.

Really??? Sounds odd to me. I bet it will be highly intersting and educating, though!

xine is not a "sound engine".

True true.

I took some care before writing the post and did a search for 'sound engines' on synaptic and those 3 came up.

I have also been thinking about the subject a little more.

I have xmms, rhythmbox and amarok as my media players. Rhythmbox definitely has the richest sound to my ears. It could be (most probably is) the settings in the actual players that are making the difference, something that this thread may address. (There is little option to configure the sound in rhythmbox as far as I am aware though).

The whole question of sound quality and music playback is important for many users in my opinion and will become more so as computers become more integrated into households especially. There is some great stuff out there currently and whilst there may or may not be differences between engines the listening experience can certainly be different.

someone mentioned my name which is a very scarry thing.
i am a musician and yes i hear the differences too though i'm not sure of the reasons.
those old mixing daemons like arts and esd are totally useless and of no value since alsa came along
IMHO. (jack is another story)
For me the alsa oss emulation layer sometimes does a little better that straight alsa with some software, i think because the OSS compatibility devices respect POSIX rules and allow only exclusive opens on the device. This might mean software mixing layer is being skipped since it's not needed. alsa for me (with xmms only) gets all screwed up and the sound buffer starts recursing on itself. This kind of stuff might be different for different card drivers and even different versions of alsa. Different chipsets have different fragmentation limits if not respected by the software somehow or adjusted by the user can cause sound fragment truncation also.

My personal favorite is jackd -> alsa using xmms
which is very simiar to just aoss using xmms
gstreamer -> alsa also sounds good in a very different way (rhythmbox)
i don't do remote connections or anything but that's a consideration
but this still doesn't really adress the OPs question of "engine" which is something i think of as an API for generating midi sounds like for game programming or something but i'm sure it can mean alot of things. Sound is certainly a complex thing.

music file playback in Linux is all about the code in the player combined with the kernel latency and throughput and junk and the output layer. What's different about the various players is a mystery. possibly some use gain restraints or something and other issues.
Some could be using complex EQ controlls but i imagine most don't. Obviously some use different output APIs or use some better than others. Different output layers must have different levels of controll for various sound factors.
with all software it's all about what sucks less.

alsa is a HAL or hardware extraction layer that is itself software with it's own level of software glitches and insanity then it's used by other daemons and whatnot that are being written to by other programs that are reading decoding and outputting encoding formats all with their own different level of software weirdness and design problems it's certainly no wonder it all sounds different. All getting processed by a kernel that is either quick to respond or not then getting run to hardware through various burst size or latency settings accross the bus that can't be known by the software designer.

Someone who knows their stuff I think.

Its good to hear that another person thinks there is a difference and it was good to read the myriad of possible causes and factors. Fair enough to say that some of the more modern daemons handle the job better but after that it is down very much to your own system. If the poll gets a number of votes then some preference may emerge.

Of course, someone could pull down funding for such a question and all the possible permutations.

Be intersesting to hear what any sound engineers have to say......

Project for self:

Try them all as we dont have to pay for the players

Keep on listening!