Do media players have a sound?

[Phelonious Ponk]

and people make
Audirvana
Decibel
Max
XLD
PureMusic
Amarra
Spatial
Mach2
etc
as not everybody is happy with the constrains faith bestow on them

Even I have had my crisis of faith, Vincent. I own Max. I rarely need it, but it's here. I've auditioned Amarra and Pure Audio; for playing redbook files, I didn't hear a point. It's not a blind faith I have, but I admit, I read these threads and it looks so unnecessarily complicated over there in the other church. It's as if someone didn't design the layout of the sanctuary properly in the first place and the faithful are constantly moving walls around to get from pew to altar. Meanwhile, I'm drinking the wine.

And that's as far as I'd like to stretch that metaphor, as I was mostly just amused. Here. Have another:

Tim

 

[Vincent Kars]

Cheers

 

[Phelonious Ponk]

Cheers

Salute. By the way, I agree with this post completely...

I do think it is possible that different media players do sound different.
The explanation is probably along the same lines as the difference between WAV and FLAC.
http://www.whatsbestforum.com/showth...LAC-versus-WAV

Tim

 

[richr]

...Most streaming audio-video related applictions and services would be classified as connectionless netwok...

Not always. If you watch the network traffic stream for Pandora (for example) you can see that it sends audio over HTTP/TCP (mostly to get through firewalls I suspect) . It is possible that once in a long while a packet might be dropped, but almost all the time there will be an error free stream.

In fact since Pandora requests megabytes at a time in advance, it could even handle dropped packets at the application layer in an error-free manner (the TCP layer will inform the application that something went wrong).

 

[FrantzM]

Not always. If you watch the network traffic stream for Pandora (for example) you can see that it sends audio over HTTP/TCP (mostly to get through firewalls I suspect) . It is possible that once in a long while a packet might be dropped, but almost all the time there will be an error free stream.

In fact since Pandora requests megabytes at a time in advance, it could even handle dropped packets at the application layer in an error-free manner (the TCP layer will inform the application that something went wrong).

Pandora way of "streaming" is mostly for security purposes. So that the user cannot (very simply) save the songs, firewall traversal is not much of a problem these days for udp and tcp alike ... It remains however as Orb pointed out that for streamin it is more efficient to use udp rather than tcp .. as he wrote, (emphasis is mine)

...Most streaming audio-video related applictions and services would be classified as connectionless netwok...

 

[Orb]

Just to be clear, we're talking here about streaming music services such as Pandora, Rhapsody, or Last.FM, correct? In my limited understanding, you're right about this, although I think that, in terms of audio quality, the issue with these services is the level of compression (bitrate) rather than data transmission per se.

Inside a PC, failure to accurately transfer data to the soundcard bus would be the conceptual equivalent of failure to accurately transfer data from memory or hard drive, i.e. indicative of catastrophic hardware failure.

Just want to comment again as I noticed another query - Thanks Frantz for picking up on it.
Scott I am using your response more as a reference rather than actually responding to anything you said.

The UDP streaming of packets will be of consideration more in future than now due to QoS-weighting of packets in carrier networks, and that they are not guaranteed to travel the same path.
Another aspect that affect VoIP is latency and order packets received.
I am not suggesting this is a major problem, just that it should not be considered comparable to a TCP connected session that are "most" connections to web browser-site as an example.

UDP is specifically designed to be efficient and fire and forget packets that do not rely on the same level of information protection, application based information is nearly always TCP to protect the information and provides transmission resiliance and correction as any incorrect data can be a problem, this though can add unexpected delay-latency.
And I think I mentioned before (more in response to richr), that these audio streaming services combine multiple sessions, there will be TCP for management and application connection to the service, then there will be the UDP for the streaming of the actual music.
It is highly likely that there are at least 4 sessions related to audio streaming and the associated application and site server.

Thanks
Orb

 

[Othersongs]

This is something that has puzzled me as well. Once identically set up, how can two different players sound different? But I think that they do.

To my surprise, I agree that two different players on the same PC sometimes do sound slightly different.

I set up J River and Foobar on my music server and they sound different. Foobar was noticeably cleaner - even the Foobar website says that Foobar should not sound better than other correctly set-up players. I thought that it might be jitter, but then I added a sound spectrum display in Foobar - that should eat up CPU resources and make it sound different, but it didn't. I hoping that somebody else here has some ideas.

For me the players are J River Media Center ver. 16 and Foobar2000 ver. 1.1.1, on a recent Gigabyte/AMD PC. Whereas I suspect that you're using an Intel setup?

FWIW I doubt that choosing either Intel or AMD makes any sound difference. It's the player. Given AMD's lackluster outing of their new FX bulldozer CPUs, maybe I'll spend the extra money for an Intel setup.

I'm using the on-board mobo audio (Realtek ALC889) via the mobo coax S/PDIF output (mobo only has back optical connector, had to fab wiring from the mobo's 2 pin out connector) into my external hifi system's Bel Canto DAC3 which handles 24/96 via coax S/PDIF input. On the Win7/64 driver I only have 96 checked, and not 32/44.1/48/88.2/192. Meaning that the digital format is likely being kept intact at it's original 24/96 format.

Anyway, J River Media Center ver. 16 sounds slightly better; this is with 24/96 music (stereo .flac siles) that I downloaded from http://www.2l.no/ ( Click on 2L brand store, then select Test Bench HD Audio Files ). Subtle stuff like slightly better air and occasionally less abrasive.

I only setup with J River Media Center ver. 16 a few days ago, so haven't done any listening with standard redbook (16/44.1) music.

 

[fas42]

All part of the great mystery, the great journey of unravelling all the myriads of influences on sound quality. As Gary once said, the colour of the cable can have an impact, Believe It or Not!! Once you kick the objectivists out of the room, with their $100,000 bits of measuring kit, which they don't use properly to really check for the subtleties of component behaviour, we all might get somewhere ...

Frank

 

[Vincent Kars]

On the Win7/64 driver I only have 96 checked, and not 32/44.1/48/88.2/192. Meaning that the digital format is likely being kept intact at it's original 24/96 format.

If you do “nothing”, using DS (Direct Sound, Win default audio driver) all audio is converted to 32 float by the mixer and back to integer to feed the DAC and dithered.

As you have JRiver you might try the WASAPI driver to bypass the Win mixer.
http://thewelltemperedcomputer.com/SW/Players/MC14/MC_Wasapi.htm

 

[Othersongs]

On the Win7/64 driver I only have 96 checked, and not 32/44.1/48/88.2/192. Meaning that the digital format is likely being kept intact at it's original 24/96 format.

If you do “nothing”, using DS (Direct Sound, Win default audio driver) all audio is converted to 32 float by the mixer and back to integer to feed the DAC and dithered.

As you have JRiver you might try the WASAPI driver to bypass the Win mixer.
http://thewelltemperedcomputer.com/SW/Players/MC14/MC_Wasapi.htm

I saw Gary's post in this thread where he provided a graphic of foobar2000 Preferences list (post#50) which is why I thought that he's likely using an Intel based PC. Meaning that my AMD based PC shows no WASAPI named driver to choose from within my foobar2000 player driver choices. So I'm fairly certain that I had/have foobar2000 player and also mobo Windows driver set up properly.

Meaning one not only has to change the player for driver choices, but *also* the mobo audio driver (Realtek in my case).

It's hard to do these posts and be accurate and complete without writing a dreary book.

Anyhow, I checked my J River Media Center player and was surprised that it lists WASAPI as one of my driver choices. (Still learning about JRMC)

FWIW kicking the JRMC WASAPI driver in seems a mixed bag? The sound again changes subtly. A quick listen it may be still less abrasive at some points but it's not a clear win. There's a 2nd window with WASAPI settings which is now set to "default" whereas with "Direct Sound" I had to choose one of two Digital Audio (S/PDIF) settings to get it to work. Not only that, with WASAPI as driver there's now an unchecked choice of "Present 24 bit data in a 32 bit package (required by some hardware)" Wow I wonder if my B.C. DAC3 needs that checked? Talk about confusing choices!

After listening for an hour (to only JRMC, no foobar), I went back to "Direct Sound" and may like it better? They clearly sound different. Anyhow I've hit the wall for listening for today.

And with regard to Scott who seemed to think that if it does exist it is a hardware matter, I don't see that at all; the hardware is the same for both players; to me the sound difference is the players or drivers (software).

(Those without fear can insert software/hardware joke here. Be aware that like Scott I'm a software guy. )

 

[Othersongs]

If you do “nothing”, using DS (Direct Sound, Win default audio driver) all audio is converted to 32 float by the mixer and back to integer to feed the DAC and dithered.

Within JRMC with "Direct Sound" as driver choice, under "Settings" click on "DSP and output format" then on the DSP Studio window that comes up, when Output Format is clicked (on the left column), there's an interesting set of choices, but it does appear that it process the input stereo 24/96 at 64/96 internally. Is that what you mean by "32 float"?

 

[Vincent Kars]

The DSP in JRiver is done in 64 bit float regardless of the sample rate.
If you choose “no resampling” no processing will take place.

If you use DS, the audio is send to the Win mixer. There the integer > float > integer+dither conversion happens. This audio engine does its processing in 32 bit float.
If the sample rate of the audio doesn’t match the sample rate as set in the sound panel, the audio engine will also resample.

If you use WASAPI in JRiver then you bypass this conversion.

 

[Othersongs]

The DSP in JRiver is done in 64 bit float regardless of the sample rate.
If you choose “no resampling” no processing will take place.

If you use DS, the audio is send to the Win mixer. There the integer > float > integer+dither conversion happens. This audio engine does its processing in 32 bit float.
If the sample rate of the audio doesn’t match the sample rate as set in the sound panel, the audio engine will also resample.

If you use WASAPI in JRiver then you bypass this conversion.

Thank you for this thread, and your recent help
to me in this thread on player setup.

Perhaps a bit OT but related, the threat to
CD music disc existence continues to be from
low-rez music and not from from high-rez music,
either downloads of it via broadband internet
nor BD music discs.

Given the complexity of proper high-rez playback
setup with players on personal computers with
Windows OS, I'll be continuing with setting up
my PCs with standard redbook (16/44.1).

My key reason is that CD discs offer the best
combo of low cost and high quality. So if I get
a CD that I'm not fond of the music, which
happens way too often, then it hasn't cost
me much, whereas high-rez costs a minimum of
3 times more in HDD space and money spent.

FWIW I rather like the JRiver player and will
likely buy it when my 30 day trial expires,
especially given that I can set it up on several
of my home PCs for the price of one.

BTW while I've got one decent/recent PC sound card
(Asus Xonar STX), I'm finding that 2 channel mobo
sound is just as good if not better.

 

[garylkoh]

If you use DS, the audio is send to the Win mixer. There the integer > float > integer+dither conversion happens. This audio engine does its processing in 32 bit float.
If the sample rate of the audio doesn’t match the sample rate as set in the sound panel, the audio engine will also resample.

BTW while I've got one decent/recent PC sound card
(Asus Xonar STX), I'm finding that 2 channel mobo
sound is just as good if not better.

May be it does have to do with the hardware you are using. If you are using the mobo sound and/or a PC sound card the combination with DS could possibly sound better.

I was told recently that the latest version of JRiver sounds quite a bit better than the last few versions. I haven't had time to upgrade my JRiver installation and do a comparison - I'm already very comfortable with the Foobar interface that we developed for the Pacific Northwest Audio Society.

 

[Othersongs]

May be it does have to do with the hardware you are using. If you are using the mobo sound and/or a PC sound card the combination with DS could possibly sound better.

I was told recently that the latest version of JRiver sounds quite a bit better than the last few versions. I haven't had time to upgrade my JRiver installation and do a comparison - I'm already very comfortable with the Foobar interface that we developed for the Pacific Northwest Audio Society.

Sorry that I wasn't more clear.

My Asus Xonar STX sound board is presently not installed.

I wanted to run it on one of my two Win7/64 Pro PCs but at that time (sometime several months ago) the install disc completely screwed up the PC I tried it on. Twice no less, so the 1st time wasn't a fluke. I called up Asus and got nowhere on getting a more recent install disc, with their telephone support. It really soured me on buying any more Asus products.

Fortunately I do full HDD clone backups so it was one hour down the drain each time, and more than an hour begging Asus for help on the phone. I even offered to buy the disc but got no sympathy.

I used dial-up for internet at that time, and the large Win7 download/update files were just too over the top to bother with it. The board works OK with Win XP which is where I'll try using it with a motherboard that I don't like the on-board sound of.

BTW is your PC mobo Intel based?

 

[garylkoh]

The full disclosure of my music server build is here: http://www.whatsbestforum.com/showthread.php?1201-Music-Server-Project&p=17881&viewfull=1#post17881

Yes, it is Intel based.

I also have a newsletter that describes the build - with pictures - that you can get here: http://www.genesisloudspeakers.com/support_newsletters.html

 

[Othersongs]

The full disclosure of my music server build is here: http://www.whatsbestforum.com/showthread.php?1201-Music-Server-Project&p=17881&viewfull=1#post17881

Yes, it is Intel based.

I also have a newsletter that describes the build - with pictures - that you can get here: http://www.genesisloudspeakers.com/support_newsletters.html

Interesting refs.

Anyway, I upgraded my 1.1.1 foobar2000 to 1.1.9 (both free) on both my 2 Win7/64/Pro PCs, and also got the WASAPI DLL at http://www.foobar2000.org/

Sound quality of foobar2000 ver. 1.1.9 w/WASAPI seems equal to that of JRiver Media Center ver.16 w/WASAPI; i.e. my best guess assessment after doing the update on both PCs and listening to some of my favorite stereo redbook .flac files.

Some plus/minus for both on less important stuff; e.g. JRiver permits keyboard shortcut for mute of audio playback (Cntrl-M).

 

[Othersongs]

The full disclosure of my music server build is here: http://www.whatsbestforum.com/showthread.php?1201-Music-Server-Project&p=17881&viewfull=1#post17881

Wow, only post #14 of 114 total in a thread (Music Server Project ) started by PNWAS.

I need to settle in and read that in detail.

One key question is how much have you noticed that SQ (Sound Quality) varies with on-board sound (via coax S/PDIF output) on recent mobos?

Since you've been at this for more than 5 years, was on-board sound more iffy back then?

All of mine (5 different ones) are all Gigabyte, all AMD based. Of the 3 that I've seriously listened to the on-board sound, 2 are excellent with 1 lacking; not sure how much lacking the one laggard is but definitely lacking. I'll soon return to that one to set up dual boot with XP/Linux, listen for a bit, then likely install Asus Xonar STX on it.

Anyhow, don't get distracted by the Xonar; my key question to you and others is how much variety of SQ have you noticed with on-board sound of different recent mobos?

 

[garylkoh]

I haven't used the on-board sound much. Taking Firewire out with the mobo of my build last year into a good DAC was far, far better than optical out. What I did find in a previous build was that the USB port on the mobo itself to sound better than the USB ports in the back of the mobo. There were few mobos with a USB port directly on the mobo.

However, with the new async USB DACs, all this makes less of a difference. The computer acts simply as a data pump, and the DAC re-clocks everything internally, making jitter much, much less of an issue. Unfortunately, with USB, galvanic isolation then needs to be done right, and many aren't.