Big question on music streaming

[WinterRacer]

Hi,

I read with interest the "Big Question" on music streaming in the latest mag and had a question about the set up used for the test. In the test a Macbook and an Olive 4HD are used, but how was the system set up to get those files from the Mac to the Olive?

For example, was music server software installed on the Mac and the files streamed, e.g., a DLNA server installed on the Mac, or were the music files accessed from the Mac via a shared folder, etc.?

Could WHF give a bit of detail about how the system was set up to perform this test please?

Many thanks.

 

[WinterRacer]

I guess not. Shame as I think it's good that WHF tries to tackle issues like this. Personally I would like a bit more information on the setup on tests like this to help me determine what's being tested and how.

From my knowledge, if data is being streamed using reliable network protocols (e.g., as a Squeezebox does), the use of wireless vs. ethernet can't make a difference. Hence the question.

IMO, what would be good for readers is for WHF to give advice on setups and products so that the use of wireless vs. ethernet didn't make a difference rather than getting people to worry about things that may not effect them.

 

[AnotherJoe]

look at this thread. The differences are caused by the laptop streaming to the streamer (yes it is a strange setup). On the other hand it is the way some people (Apple users in particular) will work.

http://www.whathifi.com/forum/accessories/streaming-over-ethernet-testing

 

[WinterRacer]

That thread is locked, so I doubt I'll get an answer on this one.

Olive don't give out too much detail on their site, but as far as I can tell it uses DLNA for streaming. If differences did exist, my guess would be the application protocol had to resort to interpolation as too many packets were getting lost.

DLNA streaming has mandatory support for http chunking over TCP and option support for RTP (over UDP), so packet loss could have been avoided.

I think a much better article would have been one to help readers to avoid problems.

 

[Andrew Everard]

We have already 'how to' articles on streaming set-ups, and will continue to do so. This was simply to see whether readers felt they could hear differences between different methods, and whether they'd prefer one over another. But not having been involved in to set-up of the article, that's really all I have to offer on the subject.

(Apart, that is, from the slightly daunting prospect of setting up a complete new server/streaming system tomorrow, including configuring, formatting and copy content over to a new NAS unit, as my current one is full. Could be a long day).

 

[Ravey Gravey Davy]

The set up is irrelevant- there were 3 changes- homeplug/wireless/hardwired- and can the invited people hear the difference.

Answer -all three heard differences and preferred different sounds from the 3 options.There aren't "problems",and it isn't a problem solving test.It is an individual preference test based on three people hearing the same system in three different formats.I wish people would stop trying to prove things through statistics and "proven " hardfacts and just accept that different people hear different things.

That thread may have been locked,but the early beautiful post by the prof really said it all about the Big Question approach.he could equally have done the same post about light reception through the eye and its transmission to the brain if the question had been about TVs.It is personal interpretation in this test,not a be all and end all verdict on sound and vision.

 

[WinterRacer]

Thanks for the reply. BTW, if you've got lots of data to copy over, it's far less painful and time consuming if everything is connected by ethernet with a gigabit switch. Also make sure you use a restartable copy procedure, e.g., rsync or robocopy rather than just copy and paste.

Sorry if sucking eggs springs to mind.

 

[Andrew Everard]

Everything except one laptop and iOS devices already Ethernet connected, but i fear those copy solutions may be a tad too Windowscentric for me.

 

[WinterRacer]

The set up is irrelevant- there were 3 changes- homeplug/wireless/hardwired- and can the invited people hear the difference.

Answer -all three heard differences and preferred different sounds from the 3 options.There aren't "problems",and it isn't a problem solving test.It is an individual preference test based on three people hearing the same system in three different formats.I wish people would stop trying to prove things through statistics and "proven " hardfacts and just accept that different people hear different things.

That thread may have been locked,but the early beautiful post by the prof really said it all about the Big Question approach.he could equally have done the same post about light reception through the eye and its transmission to the brain if the question had been about TVs.It is personal interpretation in this test,not a be all and end all verdict on sound and vision.

The setup may be irrelevant to you, but not to me. I guess we're coming at this from quite different angles, so no point in debating and getting another thread locked. I'm glad you enjoyed the article.

 

[AnotherJoe]

The set up is irrelevant- there were 3 changes- homeplug/wireless/hardwired- and can the invited people hear the difference.

Answer -all three heard differences and preferred different sounds from the 3 options.There aren't "problems",and it isn't a problem solving test.It is an individual preference test based on three people hearing the same system in three different formats.I wish people would stop trying to prove things through statistics and "proven " hardfacts and just accept that different people hear different things.

That thread may have been locked,but the early beautiful post by the prof really said it all about the Big Question approach.he could equally have done the same post about light reception through the eye and its transmission to the brain if the question had been about TVs.It is personal interpretation in this test,not a be all and end all verdict on sound and vision.

The setup is not irrelevant. In this case it is vital. It is all down to the fact that one device is streaming to another.

When the laptop streams to the Olive it uses RTSP + UDP as the transport protocol which is susceptible to packet loss - so the 3 may well sound different depending how many packets get lost.

If the Olive was connected directly to a NAS (as most people who buy a standalone streamer would) there would be no difference as it would use TCP/IP. One of the main points of having a streamer of course - is so need to have a pc on.

 

[WinterRacer]

AnotherJoe,

Respectfully, I think you might be focussing on the role of the NAS a little too much. Remember that a SOHO NAS is really just a low powered computer, usually running a unix variant, that allows files to be accessed by protocols such as SMB(CIFS), NFS, AFP, etc. A lot of NAS devices also will run small programs to support streaming, e.g., Minidlna, Squeezeboxserver, etc.

The use of an AppleMac vs. a NAS is not really here or there, rather how do the files get from the AppleMac, NAS, WHS, or whatever to the steaming device.

Perhaps you're suggesting that a more normal setup with an Olive, is to avoid a network all together and load files directly onto the internal HDD or perhaps to plug in an external USB drive?

IMO, more representative setups than the one used would be a computer running iTunes -> Apple AEX or computer running SqueezeboxServer to a Squeezebox. More high end alternatives would be a NAS running a DLNA server to a Nail or Linn streaming client.

The Olive device used appears to be attempting the all-in-one solution.

Perhaps a future big question could be one comparing streaming players across a wide price spectrum, e.g., Apple, Squeezebox, Naim, where they all feed into a common DAC, amp, speakers. That is, taking as many variables out of the equation as possible, to determine if the digital out of these streaming devices sound any different and is expensive better?

(Sorry if this comparison has already been done and I missed it).

 

[def lugs]

I was unsure as to what the conclusion of the article was. I read it as -
"an equal price point CD player and streamer = CD wins. Spend loads more on a streamer and you will get a better sound than the cheap CD player." Should this have been followed up by "spend the same money on a CD player and the CD will still win." Or does the streamer better the CD player once you reach the premium end?

 

[professorhat]

I was unsure as to what the conclusion of the article was. I read it as - "an equal price point CD player and streamer = CD wins. Spend loads more on a streamer and you will get a better sound than the cheap CD player." Should this have been followed up by "spend the same money on a CD player and the CD will still win." Or does the streamer better the CD player once you reach the premium end?

I think this thread was originally about the previous BQ where streaming over ethernet, wireless and powerplugs was compared.

However, your conclusion on the CD player vs streamer is over simplified. I preferred System 1 overall whereas the other guys preferred System 2 overall. So there's no simple "one = better than the other" equation, you need to listen and decide for yourself. A lot may come down to your music tastes - as you can see, when it came to the classical music track, I actually preferred System 2. However, since I don't really listen to classical at home, this wouldn't influence my decision.

And though many will doubt it, when listening through WHF's reference system, the differences were so marked it was astonishing - those who doubt and haven't tried it really should - it would be an eye opener for many I think.

 

[The_Lhc]

Prof, I'd best interested to know if at any point during the test you cottoned on to what was being changed? Or even just had a sneaking suspicion.

 

[Craig M.]

I was unsure as to what the conclusion of the article was. I read it as - "an equal price point CD player and streamer = CD wins. Spend loads more on a streamer and you will get a better sound than the cheap CD player." Should this have been followed up by "spend the same money on a CD player and the CD will still win." Or does the streamer better the CD player once you reach the premium end?

I think this thread was originally about the previous BQ where streaming over ethernet, wireless and powerplugs was compared.

However, your conclusion on the CD player vs streamer is over simplified. I preferred System 1 overall whereas the other guys preferred System 2 overall. So there's no simple "one = better than the other" equation, you need to listen and decide for yourself. A lot may come down to your music tastes - as you can see, when it came to the classical music track, I actually preferred System 2. However, since I don't really listen to classical at home, this wouldn't influence my decision.

And though many will doubt it, when listening through WHF's reference system, the differences were so marked it was astonishing - those who doubt and haven't tried it really should - it would be an eye opener for many I think.

why would it be an eye opener? you were listening to three different sources - you would hope they didn't all sound the same.

 

[professorhat]

Prof, I'd best interested to know if at any point during the test you cottoned on to what was being changed? Or even just had a sneaking suspicion.

Nope, really not a clue. The only thing you could tell was Ketan was switching something, but whether that was the amp or the source, it was impossible to tell until he revealed it.

 

[Craig M.]

The set up is irrelevant- there were 3 changes- homeplug/wireless/hardwired- and can the invited people hear the difference.

Answer -all three heard differences and preferred different sounds from the 3 options.There aren't "problems",and it isn't a problem solving test.It is an individual preference test based on three people hearing the same system in three different formats.I wish people would stop trying to prove things through statistics and "proven " hardfacts and just accept that different people hear different things.

That thread may have been locked,but the early beautiful post by the prof really said it all about the Big Question approach.he could equally have done the same post about light reception through the eye and its transmission to the brain if the question had been about TVs.It is personal interpretation in this test,not a be all and end all verdict on sound and vision.

you completely missed the point on the locked thread, and again on this one. there could be an easily avoidable problem in the setup that others would like to avoid. you wish people would stop trying to prove things with facts? i wish people would stop interfering with those who would like the facts.

 

[Andrew Everard]

why would it be an eye opener? you were listening to three different sources - you would hope they didn't all sound the same.

The only thing you could tell was Ketan was switching something, but whether that was the amp or the source, it was impossible to tell until he revealed it.

 

[Ravey Gravey Davy]

The set up is irrelevant- there were 3 changes- homeplug/wireless/hardwired- and can the invited people hear the difference.

Answer -all three heard differences and preferred different sounds from the 3 options.There aren't "problems",and it isn't a problem solving test.It is an individual preference test based on three people hearing the same system in three different formats.I wish people would stop trying to prove things through statistics and "proven " hardfacts and just accept that different people hear different things.

That thread may have been locked,but the early beautiful post by the prof really said it all about the Big Question approach.he could equally have done the same post about light reception through the eye and its transmission to the brain if the question had been about TVs.It is personal interpretation in this test,not a be all and end all verdict on sound and vision.

you completely missed the point on the locked thread, and again on this one. there could be an easily avoidable problem in the setup that others would like to avoid. you wish people would stop trying to prove things with facts? i wish people would stop interfering with those who would like the facts.

I think our lines are crossed -I do not have a problem with trying to understand more factually when one person hears differences in two systems,or if there could have been a better way of setting up. I am referring to three people listening to one set up at the same time with the same output-and they all hear different things.Packet loss is irrelevant.It applies to all three people.The search for facts to prove the differences at that juncture seems to me fairly pointless.

There is of course a completely different discussion to be had on whether the set up could have been different etc( I hesitate to say better because then the argument becomes circular)

but that is not the point of the Big Question-and as usual it generates more questions than answers but it is after all a fun day .

 

[Craig M.]

why would it be an eye opener? you were listening to three different sources - you would hope they didn't all sound the same.

The only thing you could tell was Ketan was switching something, but whether that was the amp or the source, it was impossible to tell until he revealed it.

i'm not sure i understand your point. i was commenting on the prof saying (if i read it right) that the marked differences between the systems was an eye opener. maybe if all three sources were connected via, say, optical to a decent dac, and then there were marked differences - that would be an eye opener. to me anyway. i'm not sure if the analogue outputs of three different sources sounding different, qualifies.

 

[Craig M.]

ravey, sorry, your meaning wasn't clear to me from your post. i understand now, thanks for explaining.

 

[Anonymous]

When the laptop streams to the Olive it uses RTSP + UDP as the transport protocol which is susceptible to packet loss - so the 3 may well sound different depending how many packets get lost.

As much as I do not believe the described sound variations can be caused by differences in digital transmission, I do believe you are talking with the wrong side of your body here. If we're talking packet loss, then we're talking about an information gap of typically 1500 bytes. Disregarding UDP overhead (32 bytes) and assuming the source files were uncompressed WAV, we are talking about a 1ms gap. If the source material was compressed, multiply that gap with the compression factor (6-10 for a typical mp3, 2 for flac). Such a gap can not be interpolated without audible defects.

That does not rule out packet corruption though, and you are right, TCP/IP has safeguards against such corruption that UDP does not have. But packet loss is very unlikely.

 

[WinterRacer]

As we know UDP is an unreliable transport. MTU is often set at the 1500 bytes, however this is a maximum, VOIP apps often use smaller packet sizes, no idea what packet sizes are used here, so not sure we can rule out packet loss/corruption and interpolation.

However, despite the above I totally agree that SQ differences perceived are unlikely to be caused by the choice of network connection. Which is my issue with these articles, that is, for me they don't actually answer the big question they ask.

Reading this months Big Question, I get the same sense of missed opportunities.

For me, the Big Questions seems more about reinforcing the idea that more expensive is better. 🙁

 

[Andrew Everard]

For me, the Big Questions seems more about reinforcing the idea that more expensive is better. 🙁

Whereas of course the fact of the matter is that we have no such agenda, and just report what the BQ participants say.

 

[BenLaw]

For me, the Big Questions seems more about reinforcing the idea that more expensive is better. 🙁

Whereas of course the fact of the matter is that we have no such agenda, and just report what the BQ participants say.

This edition's BQ gave us the rather uninspiring conclusion that (generally) a highly rated source at a higher price point was better than a highly rated source at a lower price point. :O Can't imagine such a test would have been performed with CDPs.