One cause of rebuffering on the Touch

[Phil Leigh]

Who introduced religion? I never suggested the sound quality would be improved by tweaking network settings, not did anyone on this thread as far as I can see. I did say you could get transmission problems, and some people do, even with default settings.

I was interested to try and resolve problems I had, and reported what I found, thinking other people might be interested. Sure, if you don't like tinkering, don't, but some do, and some have to.

It has been suggested elsewhere that reducing the load on the Touch can improve the sound quality, and the effect of transmitting PCM rather than FLAC is measurable (using ADM). Whether one can hear a difference is a different matter. That's not the topic of this thread.

No - that's wrong. ADM reports NO difference when flac or PCM is used - at least that has always been my test result (with both Touch and SB3). If you have evidence to the contrary, I'd be interested to see it / try and reproduce it.

[pippin]

Who introduced religion? I never suggested the sound quality would be improved by tweaking network settings, not did anyone on this thread as far as I can see. I did say you could get transmission problems, and some people do, even with default settings.

I didn't introduce religion, I said it's a religious discussion and that's what it is and why I actually wanted to stay out of it. Have a look at other threads in this or the audiophile forum if you don't believe it. I wanted to stay out of this, I just say that I don't believe there is a difference and I have good reason to do so but if others believe it, they may as well tinker with it.

I just wanted to make clear that these kind of issues are what you have to expect if you do things like that.

Apart from that, there are threads on this forum which DO suggest that the network settings on the server do have an effect and since you mentioned that

My NIC was set at ‘100Mbps full duplex’ following advice given elsewhere.

I assumed you did that, too, for sake of SQ,I was probably wrong with that assumption.

I was interested to try and resolve problems I had, and reported what I found, thinking other people might be interested. Sure, if you don't like tinkering, don't, but some do, and some have to.

It was not about tinkering but I wanted to add the advice that if you run into trouble,go for the default settings first.

It has been suggested elsewhere that reducing the load on the Touch can improve the sound quality

It has also been suggested elsewhere that running your server under Windows instead of Linux or using special SATA cables in your server setup will improve SQ. There are a lot of things that are being suggested and you will always find a lot of people who hear "dramatic" differences in SQ once they apply these. Good for them. Even better for certain vendors.

[PasTim]

I didn't introduce religion, I said it's a religious discussion and that's what it is and why I actually wanted to stay out of it. Have a look at other threads in this or the audiophile forum if you don't believe it. I wanted to stay out of this, I just say that I don't believe there is a difference and I have good reason to do so but if others believe it, they may as well tinker with it.

I just wanted to make clear that these kind of issues are what you have to expect if you do things like that.

Apart from that, there are threads on this forum which DO suggest that the network settings on the server do have an effect and since you mentioned that

I assumed you did that, too, for sake of SQ,I was probably wrong with that assumption.

It was not about tinkering but I wanted to add the advice that if you run into trouble,go for the default settings first.


It has also been suggested elsewhere that running your server under Windows instead of Linux or using special SATA cables in your server setup will improve SQ. There are a lot of things that are being suggested and you will always find a lot of people who hear "dramatic" differences in SQ once they apply these. Good for them. Even better for certain vendors.

This is getting out of hand. I will leave a few comments and then ask the sound quality topic be left to rest in this thread. All I was trying to do was to help people use Windows tools to see what might be going on when they have problems.

If you say it is a 'religious' discussion, that's about 'religion'. Best to find different terms or 'religious' people may get upset.

Network settings can cause, and prevent, re-buffering, stuttering and crackling. Not sound quality exactly, but certainly real and annoying effects.

[pallfreeman]

If you say it is a 'religious' discussion, that's about 'religion'. Best to find different terms or 'religious' people may get upset.

"Religious" is being used in the technical sense: something that a person believes despite others having evidence to the contrary. It's quite a common geek use of the word, nothing to get upset about.

[PasTim]

No - that's wrong. ADM reports NO difference when flac or PCM is used - at least that has always been my test result (with both Touch and SB3). If you have evidence to the contrary, I'd be interested to see it / try and reproduce it.

My procedures were somewhat sloppy - I wasn't trying to be very methodical, just tinkering. I don't have the raw data any more, and I'm not enthusiastic enough to want to go through it all again (not least because I have to move stuff around the house to do the tests).

However, with SPDIF into my sound card my notes say I got correlations in the low 90s from a 2 minute piano track compared with the original rip using PCM and mid 70s with FLAC (all at CD quality). Maybe my card didn't like some extra noise or somesuch. Anyhow, that's what I saw, 'wrong' or otherwise.

Having seen that level of difference with SPDIF I didn't try with the analogue out.

[Phil Leigh]

My procedures were somewhat sloppy - I wasn't trying to be very methodical, just tinkering. I don't have the raw data any more, and I'm not enthusiastic enough to want to go through it all again (not least because I have to move stuff around the house to do the tests).

However, with SPDIF into my sound card my notes say I got correlations in the low 90s from a 2 minute piano track compared with the original rip using PCM and mid 70s with FLAC (all at CD quality). Maybe my card didn't like some extra noise or somesuch. Anyhow, that's what I saw, 'wrong' or otherwise.

Having seen that level of difference with SPDIF I didn't try with the analogue out.

I get a null correlation over 110 for FLAC and PCM against each other - via analogue!.
This is using an M-Audio 24/96 Audiophile card, synced by s/pdif, with ADM sample rate correction enabled on 16/44.1 source material.

Do you use Replaygain? That would lower the FLAC correlation on this particular test...

http://forums.slimdevices.com/showpost.php?p=494130&postcount=105

When I get time I'll try and repeat your test (wav vs flac via s/pdif against rip)

[Phil Leigh]

...
However, with SPDIF into my sound card my notes say I got correlations in the low 90s from a 2 minute piano track compared with the original rip using PCM and mid 70s with FLAC (all at CD quality). ...

Something was very wrong with your test procedure,or your sound card/drivers. Via s/pdif, both FLAC and WAV/PCM always give 100% identical (correlation null = -144) comparisons with the original rip file.

However, for this to work you do need a sound card that is not being (ahem) "managed" by the operating system... e.g. no Windows kernel sample rate conversion weirdness...

By the way, this test actually tells you nothing about what happens when you connect a DAC to the Touch s/pdif... it simply proves what we already know, namely that both wav and flac playback are "bit-perfect" via s/pdif.

Noise from the transport can IN THEORY still interfere with the operation of SOME DACs - likewise, jitter and clocking can still be an issue for SOME DACs.

[PasTim]

Something was very wrong with your test procedure,or your sound card/drivers. Via s/pdif, both FLAC and WAV/PCM always give 100% identical (correlation null = -144) comparisons with the original rip file.

However, for this to work you do need a sound card that is not being (ahem) "managed" by the operating system... e.g. no Windows kernel sample rate conversion weirdness...

By the way, this test actually tells you nothing about what happens when you connect a DAC to the Touch s/pdif... it simply proves what we already know, namely that both wav and flac playback are "bit-perf ect" via s/pdif.

Noise from the transport can IN THEORY still interfere with the operation of SOME DACs - likewise, jitter and clocking can still be an issue for SOME DACs.

I haven't got the time or energy to try again right now, but the results were consistent over a couple of runs. I was using an M-audio 24/96, external sync. The results were not exactly the same as the original. Comparing the rip to itself I got a correlation of 300 (as below - I kept that one for reference)

parameters: 0sec, 0.000dB (L), 0.000dB (R)..Corr Depth: 300.0 dB (L), 300.0 dB (R)

I don't know why I get 300 and you quote 144. Maybe a different version of ADM? Mine is 3.22.

I'm using M-audio's own Windows 7 64 bit drivers.

During the runs it said I didn't need sample rate compensations, so I didn't.

Something of a mystery then.

I don't have an external DAC (see signature below) but was running with SPDIF into the sound card since I thought it might tell me something interesting.

[Phil Leigh]

My tests were extensive and under reasonably controlled conditions. There was no difference recorded. I was using XP drivers for the M-Audio, but that shouldn't matter... YOU CANT GET A NULL BY ACCIDENT.

You can get a difference by accident.

I have no idea why you recorded a difference: it would mean that the bitstream was different - which it simply cannot be/isn't. This is easily proven by other means (e.g playing DTS or HDCD).

[PasTim]

My tests were extensive and under reasonably controlled conditions. There was no difference recorded. I was using XP drivers for the M-Audio, but that shouldn't matter... YOU CANT GET A NULL BY ACCIDENT.

You can get a difference by accident.

I have no idea why you recorded a difference: it would mean that the bitstream was different - which it simply cannot be/isn't. This is easily proven by other means (e.g playing DTS or HDCD).

My background is in (digital) computers, where bits don't usually get lost without detection (although it can happen very rarely). They do get lost/mangled, but they then get detected and corrected.

I freely admit I don't understand real-time hi-fi digital processing. As a computer man, sending data without big buffers at either end and the option to retransmit seems madness, but it seems that what audio people decided to do.

All of this is leading up to saying that I don't understand some of what you said above, especially the capitals. Things can go wrong, but a decent design would recover and correct loss.

I have no knowledge of, or facility to use, HDCD or DTS.

I do know that I did several runs, none of which came near to the ideal, and they did vary a bit.

I also have no idea why my ADM reads 300 for a rip against itself but you say 144 is perfect. Something very strange there. From what you say you get 144 every time with SPDIF. I never got close to 100. So my Touch, and/or SPDIF RCAs, and/or M-audio card, and/or drivers, and/or etc.... are playing up, and playing up reasonably consistently.

Not having a separate DAC this is only of theoretical interest to me. Trying to find the cause of what would appear to be a repeatable 'accident' is probably beyond me.

[Lestrad]

What happened to rebuffering?

This was supposed to be a thread about rebuffering and it seems to have totally abandoned that topic.

Rebuffering has nothing to do with sound quality. Rebuffering results in an absence of sound, and that sound, being absent, has no quality whatsoever. Of course silence is perfect in a way...

I'm under Windows 7 and having rebuffering problems like the original poster. I'm resurrecting this thread in the hope that somebody can help me (and others) with that.

Please?!

PS When it comes to religion, I'm a Catholic with animist tendencies.

[Apesbrain]

Is this over wireless? If so, are you able to temporarily wire the Touch and test?

[Lestrad]

Is this over wireless? If so, are you able to temporarily wire the Touch and test?

Thanks! No, Sir, I'm strictly wired. All Cat-5 or above cable, no long runs.

[Apesbrain]

I know it's a bother, but I'd temporarily install LMS (no customizations nor additional plugins) on one of your desktop PCs, index some music files on the same PC, and try to play your Touch from there. You point the Touch to the new server by: My Music > Switch Library.

If the above is successful, consider getting a Raspberry Pi for use as an LMS server. Many people on this forum can help with RPi issues. Your configuration (Arch/Wandboard) is not that common, so finding help is going to get increasingly difficult.

[bpa]

Thanks! No, Sir, I'm strictly wired. All Cat-5 or above cable, no long runs.

Plase clarify is LMS on Win7 or on Wandboard ?

If it is on Win7 then it may be worthwhile upgrading to LMS 8.*, it has many bugs fixes including a few Windows specific I/O scheduling ones.