PDA

View Full Version : Sound quality compromized when using TinySC ?



rotho
2009-11-05, 00:46
I have read in another thread ("Output to USB DAC" http://forums.slimdevices.com/showthread.php?p=481661#post481661)
that there is a loss in sound quality when using the Touch with its internal server rather than with a networked server, perhaps because of the server load.

Since this is quite a let down for the usability of the Touch in standalone mode (at least for the audiophiles like me), I thought I should start this thread and have the Touch beta-testers and developers share their thoughts and experience on that matter.

So : What is the difference in sound quality of the Touch when used with TinySC and when used with an external server (with the analog and also the digital outputs) ?

And if there is a difference, could that be solved when the Touch is released, either with hardware or software upgrades ?

audiomuze
2009-11-05, 08:27
Not sure that's a plausible explanation...surely the bits either make it to the dac or they don't?

bpa
2009-11-05, 10:14
I have read in another thread ("Output to USB DAC" http://forums.slimdevices.com/showth...661#post481661)
that there is a loss in sound quality when using the Touch with its internal server rather than with a networked server, perhaps because of the server load.


That thread was referring to using an additional (and unsupported by Logitech) external DAC attached via USB and on the same USB port there is a USB hard drive.

There is no comment about an external USB hard drive affecting audio of the internal DAC, phono or digital interfaces of the Touch.

funkstar
2009-11-05, 10:51
There is no comment about an external USB hard drive affecting audio of the internal DAC, phono or digital interfaces of the Touch.
Yes there is, in the last post.


I'm pretty sure its the server load rather than both drive and DAC on the USB port because I can here the same degradation on the analog outs as well.
I'm not doubting JohnSweson, but I am a sceptic when it comes to things like this unless it can be proved.

Phil Leigh
2009-11-05, 11:29
Yes there is, in the last post.


I'm not doubting JohnSweson, but I am a sceptic when it comes to things like this unless it can be proved.

I'm working on a test to prove there is a difference (if there is one). I just need a day free of DIY, work etc... :-(

Themis
2009-11-05, 11:39
I'm working on a test to prove there is a difference (if there is one). I just need a day free of DIY, work etc... :-(
That's very interesting, Phil.
It's nearly impossible to do an A/B test on this, without having two boxes, so, would be nice to have a test with a single box. ;)

Phil Leigh
2009-11-05, 11:42
That's very interesting, Phil.
It's nearly impossible to do an A/B test on this, without having two boxes, so, would be nice to have a test with a single box. ;)

AudioDiffMaker strikes again (or will strike when I have a spare hour or two to do the test - probably Sunday...)

JohnSwenson
2009-11-05, 19:18
I said I do hear a difference between networked SBS or TinySC, at this point I don't know for sure why. It MAY be different data, but I don't think so. Phil's test will prove that one way or the other.

If the bits are identical then "proving" it gets hard. At this level of quality (I DID say the touch's outputs are very good didn't I?) finding iron clad causal relationships between measured signal parameters and perceived sound gets difficult, the differences in the signals are very small. By the time you get sensitive enough instruments to measure the small differences you have a lot of noise sources riding on top of things. It seems the human perception system can be affected by very small amounts of certain types of distortions even when they seem to be swamped by other distortions and noise. Unfortunately we don't know what these are so its hard to come up with filters in the measuring systems to match.

The point of all that was that if you want hard measurement data that proves it, you are probably not going to get it. That leaves what people hear. Thats much more difficult to "prove".

What brought all this up was a change in the firmware several weeks ago that at least for me brought an improvement in sound quality. This had nothing to do with the USB interface, I was listening on headphones plugged in directly to the Touch. The developers said they had done nothing to change the audio processing. But they did say they had done some major re-writes improving the overall efficiency and "speed" of the code.

Since TinySC adds a lot of extra processing I decided to try it and see if I could hear a difference, and yes I could, some of the extra WOW that happened with the firmware change was gone. I then tried it with the USB DAC connection and I heard the same thing. The network server did sound better.

The primary areas of difference are in soundstaging and ambiance, reverberation. So this is somewhat recording dependent. The biggest differences are in recordings recorded live in reverberant spaces. The sense of space of the performance, the "you are thereness" is decreased somewhat with the local server.

I have heard this before in other low power linux systems. Two audio players, both outputting the same bits, can sound significantly different. In most cases the simpler one with smaller, tighter loops winds up sounding better. Before someone asks me to come up with measurements to prove it, I'll reiterate, I can't. And this is not just me, there are many other people hearing these same sorts of things with different software and identical bits. I can't give you a proven mechanism for this. Does this mean everybody who hears these things is deluding themselves? Maybe, maybe not. We don't know yet.

Remember that either way the sound form the Touch is still way better than from an SB3 or SBR, so even if you do go with a local server its still a good thing. If what I'm hearing holds true its a choice you have to make if getting the best sound possible out of the Touch is important for you. Its just another trade off people obsessed with getting the best sound have to go through.

Also remember that the firmware is still changing significantly and since all this started with what I heard at a firmware change, things may be different at ship date. I hope they manage to preserve the improvements I heard, but who knows what will happen.

John S.

peterw
2009-11-05, 20:36
Remember that either way the sound form the Touch is still way better than from an SB3 or SBR

I don't have the ears, training, space, gear, or source material to challenge such a notion, but I find it amusing how all these products over the years are said to sound both excellent and much better than their predecessors. I especially like reading about the superiority of the newer gear's digital outputs. Deluded? Obsesssed? Entertaining, you audiophile types certainly are that. :-)

Phil Leigh
2009-11-05, 23:59
Just to be clear, the test I am planning will see if there is a difference between the same flac being played from a network PC running sbs or a local usb disk runing under tinysc (all via the analogue outputs).
It won't categorically prove the difference (if any) is audible, but it should provide a guide to that.

Themis
2009-11-06, 02:05
Just to be clear, the test I am planning will see if there is a difference between the same flac being played from a network PC running sbs or a local usb disk runing under tinysc (all via the analogue outputs).
It won't categorically prove the difference (if any) is audible, but it should provide a guide to that.
We need different facts to feed the various hypothesis.
JohnSwenson gave us one fact by declaring his perceived differences; Phil will give another one. It's an enriching process.

Themis
2009-11-06, 02:12
I don't have the ears, training, space, gear, or source material to challenge such a notion, but I find it amusing how all these products over the years are said to sound both excellent and much better than their predecessors. I especially like reading about the superiority of the newer gear's digital outputs. Deluded? Obsesssed? Entertaining, you audiophile types certainly are that. :-)
You may not have the ears, training and gear but you have a screen, a keyboard and an internet connection to the forums !
I'm glad that we can entertain each other. ;)

dave77
2009-11-06, 02:22
I don't have the ears, training, space, gear, or source material to challenge such a notion, but I find it amusing how all these products over the years are said to sound both excellent and much better than their predecessors. I especially like reading about the superiority of the newer gear's digital outputs. Deluded? Obsesssed? Entertaining, you audiophile types certainly are that. :-)

He he, I agree :) Although I probably won't be able to tell the difference myself I'll be following the topic as I plan to use my Touch as a server.

JohnSwenson
2009-11-16, 00:29
Tonight I did some careful listening with the touch, FLAC and PCM streaming and TinySC. I can't listen to Wav on TinySC since it does not support WAV at this point (I hope it does in the future).

Now these are listening tests, no objective measurements, so those of you who insist on having measurements can tune out now.

All the tests were done with the same track, a FLAC file ripped from track 25 of the King's Singers' Madrigal History Tour. This has very pure acapella singing recorded in a reverberant environment. I seem to be able to hear the differences quite well with this piece. I'm listening to the Touch with cheap $5 headphones that came with a walkman clone many years ago, plugged dirctly into the headphone jack. The Touch is stock, using the logitech supplied wallwart. This is about as "un-audiophilic" as you can get.

First streaming from a 7.5 SBS server, TinySC turned off.

I can definitely hear the difference between streaming flac and PCM. When streaming FLAC the whole presentation sounds a bit "flat" compared to PCM. With the PCM the singers sound more real, more alive. The sound has more richness. I can hear more subtlties, more neuances in expression with the PCM.

Next I tried from TinySC. This is just flac since it does not support wav and TinySC does not do transcoding, so the same FLAC procesing is going on as when streaming FLAC from the external SBS. Interestingly enough this sounded slightly better than the FLAC from SBS, but not nearly as good as the PCM from SBS. Its was similar to the FLAC from SBS but had just a little more subtlety and "sparkle".

Next I tried Streaming from external SBS, BUT TinySC was running though not playing anything. Both the FLAC and PCM stream were degraded, but the PCM was degraded the most. The PCM sounded similar tp the FLAC from TinySC and the FLAC was slightly more flat than it was with no TinySC.

I TRIED to use AudioDiffMaker to see if I could measure any of these, but I must be doing something wrong because I could not get even a good null from the same track played with the same settings,let alone different settings.

So for now this is pure listening with no objective "proof" whatsoever. When I have some time I plan on doing spectragrams of both the audio out (similar to what JA does) and of the clock going into the DAC chip which is a good indicator of clock jitter.

John S.

Phil Leigh
2009-11-16, 00:46
John - very interesting - I'm getting similar results... check my earlier post in the other forum... I think we are looking at the same problem?

rotho
2009-11-16, 01:58
Thank you John, for your careful listening tests.

If I read you correctly, it seems that your recent FLAC tests seem to favour the use of the Touch in standalone mode with TinySC :



Next I tried from TinySC. This is just flac since it does not support wav and TinySC does not do transcoding, so the same FLAC procesing is going on as when streaming FLAC from the external SBS. Interestingly enough this sounded slightly better than the FLAC from SBS, but not nearly as good as the PCM from SBS. Its was similar to the FLAC from SBS but had just a little more subtlety and "sparkle".


But in a previous test in another thread, you seemed to judge the audio quality a bit compromized with the internal server:


You CAN run a USB DAC over a USB hub, I have done it. I have also run the USB DAC AND a USB drive on the hub using the internal server. It does WORK but the sound quality suffers.

So whether you run off the internal server or over a network depends on how much of an audiophile you are and how much you value the internal server capability.

A good USB DAC used with a networked server sounds very good, its not quite as good with the internal server. (its not so much the hub, but the extra load on the processor)

John S.

So, am I missing something, or did your recent test made you change your opinion about the audio quality of the Touch with the internal server vs an external server ?

Phil Leigh
2009-11-16, 02:16
I'm not John but I believe he is saying that in ascending order of replay quality from lowest to highest (from the analogue outputs) the results are:

FLAC from SBS
FLAC from TinySC
WAV from SBS

rotho
2009-11-16, 02:51
I'm not John but I believe he is saying that in ascending order of replay quality from lowest to highest (from the analogue outputs) the results are:

FLAC from SBS
FLAC from TinySC
WAV from SBS

That's also what I understood from his last post, but seems to contradict his earlier findings concerning SBS vs Tiny SC.
Anyway, let's wait for his answer on that subject!

I admit that I am a bit shocked by these tests, because I am currently in the process of ripping my entire collection (2000 CDs) in FLAC format. Should I stop right now and use PCM instead? As I am using a Mac platform, I suppose I would get AIFF files. Would that be different from WAV files, regarding audio quality? And what are the tagging possibilities with such file format?

Phil Leigh
2009-11-16, 03:08
That's also what I understood from his last post, but seems to contradict his earlier findings concerning SBS vs Tiny SC.
Anyway, let's wait for his answer on that subject!

I admit that I am a bit shocked by these tests, because I am currently in the process of ripping my entire collection (2000 CDs) in FLAC format. Should I stop right now and use PCM instead? As I am using a Mac platform, I suppose I would get AIFF files. Would that be different from WAV files, regarding audio quality? And what are the tagging possibilities with such file format?

No you can carry on - FLAC is fine as a storage medium - the best choice.

Much depends on the following:

1) are you planning on using TinySC or a seperate server capable of transcoding FLAC to WAV? if so, there is no issue

2) are you going to use an external DAC? if so, there is no issue

technobear
2009-11-16, 04:24
...I can't listen to Wav on TinySC since it does not support WAV at this point (I hope it does in the future)...

What!

I don't see this on the list of P1 and P2 bugs.

Does this mean Logitech thinks it can launch the 'Touch without WAV capability?

For me that's a show stopper.

Could the developers please chime in and tell us whether WAV will be supported by TinySC at launch.

My finger is hovering over the 'pre-order cancel' button.

rotho
2009-11-16, 05:24
No you can carry on - FLAC is fine as a storage medium - the best choice.

Much depends on the following:

1) are you planning on using TinySC or a seperate server capable of transcoding FLAC to WAV? if so, there is no issue

2) are you going to use an external DAC? if so, there is no issue

I intend to use the Touch in standalone mode (i.e. with TinySC) with an external HDD, and connect it via S/PDIF to a high quality DAC (Lindemann 820S CD/SACD player, which has 4 digital inputs), and I have big expectations regarding the audio quality! I was first planning to buy a Transporter, but the possibility given by the Touch to do without a PC made me think twice.

Now, regarding the FLAC vs PCM issue, initially, I thought that John has used FLAC and PCM files, but in fact, he did use the same FLAC file for each test.
So, can you precisely explain what John means by "streaming FLAC" and "streaming PCM" ? How does the FLAC to PCM conversion occur in each case ?
And, in particular, how does this conversion occur when using TinySC ?

Bytec
2009-11-16, 05:45
"streaming FLAC" - FLAC is decoded by player
"streaming PCM" - input file is decoded by server and raw PCM is streamed.

rotho
2009-11-16, 06:42
"streaming FLAC" - FLAC is decoded by player
"streaming PCM" - input file is decoded by server and raw PCM is streamed.

OK, now what happens when using TinySC ? Are there several FLAC to PCM decoding options ? And which sounds better ?

Bytec
2009-11-16, 06:57
FLAC and PCM should sound the same because both are losless formats.

technobear
2009-11-16, 07:52
FLAC and PCM should sound the same because both are losless formats.

If only it were that simple!

There is no disputing that the 1's and 0's are correct. This, however, does not tell the whole story.

The differences being heard here will be due to increased jitter or noise - or possibly both - when the processor is working hard to decode a FLAC file.

pfarrell
2009-11-16, 07:57
technobear wrote:
> The differences being heard here will be due to increased jitter or
> noise - or possibly both - when the processor is working hard to decode
> a FLAC file.

What?!? Pure speculation.

Flac was designed specifically to be easy to decode by low powered
devices. A Touch is far more powerful than flac requires.

Go dissemble wild speculation elsewhere.

--
Pat Farrell
http://www.pfarrell.com/

andyg
2009-11-16, 08:04
On Nov 16, 2009, at 9:57 AM, Pat Farrell wrote:

> technobear wrote:
>> The differences being heard here will be due to increased jitter or
>> noise - or possibly both - when the processor is working hard to decode
>> a FLAC file.
>
> What?!? Pure speculation.
>
> Flac was designed specifically to be easy to decode by low powered
> devices. A Touch is far more powerful than flac requires.
>
> Go dissemble wild speculation elsewhere.

Yeah this thread worries me a bit, not because I think TinySC is doing something bad to the sound but because of the FUD being spread in this thread, it will scare people away from using TinySC for no reason.

rotho
2009-11-16, 09:12
Yeah this thread worries me a bit, not because I think TinySC is doing something bad to the sound but because of the FUD being spread in this thread, it will scare people away from using TinySC for no reason.

I started this thread because the Touch does indeed offer a new procedure to listen to music vs the other Squeezeboxes (i.e., by using TinySC) and perhaps the fact that it has less processing power than a dedicated PC, leads to a certain loss in audio quality. That what at least what John seemed to have experienced.

But I had absolutely no intention to denigrate the Touch and "scare" people. I am in fact very impatient for the Touch to be released, because it will certainly meet my needs to have a music server with no PC. As an audiophile, I am just interested in getting the best of all the FLAC files I am currently filling my HDD with...

So perhaps we should just wait for the final product (hardware + software) to be released, and then judge by ourselves. (And also, maybe after the Touch, there will be a "Touchporter", that is an audiophile version of the Touch ?)

technobear
2009-11-16, 09:20
technobear wrote:
> The differences being heard here will be due to increased jitter or
> noise - or possibly both - when the processor is working hard to decode
> a FLAC file.

What?!? Pure speculation.

Flac was designed specifically to be easy to decode by low powered
devices. A Touch is far more powerful than flac requires.

Go dissemble wild speculation elsewhere.

--
Pat Farrell
http://www.pfarrell.com/

Let's hear your alternative theory then.

I can hear a clear difference using a Transporter between FLAC decoded on a PC and FLAC decoded on the Transporter.

Of course you need a high-end system and well-recorded material to hear the difference. Most people simply will not hear any difference because their systems or their ears are not up to the job.

I don't really care either way as I only use WAV.

Themis
2009-11-16, 09:38
C'mon guys, Andy is right.
If it's to discuss these matters, it's better to bring this discussion to the Audiophile forum.

The SB Touch forum is a general forum, and most people don't understand the nuances of your argumentation.

JohnSwenson
2009-11-16, 09:44
Let me reitterate, In these tests the same FLAC file was used for all tests. TinySC as commonly configured does not transcode, so when playing a FLAC file it can only send a FLAC stream to the player. Currently it does not support WAV files so I can only hear FLAC decoding from TinySC right now.

When I say SBS I'm refering to an external SBS connected by network to the Touch (wired in this case). The SBS can either send the data as is (FLAC format) or it can transcode the FLAC file into PCM and stream that to the Touch.

So for these tests when using SBS the Touch can either decode PCM or FLAC, but on the Touch only FLAC.

Here are rankings, the on on the TOP is the best sounding:

source stream

SBS PCM Best
TinySC FLAC
SBS FLAC Worst

The last two are very similar but the first is clearly superior. So this seems to indicate that its NOT TinySC itself that is the issue but that PCM sounds quite a bit better than FLAC, and because TinySC doesn't do PCM at the moment its not sounding as good.

As to the format to use I agree with the others, go with FLAC, As long as the streaming can be done with PCM I have yet to hear a difference.

Things get a little more complicated here with TinySC. As it stands now you have to do Flac decoding in the player. Since TinySC and the Touch player are running on the same hardware, I'm not sure it would make any difference whether you decoded in the server or in the player. The transcoding capability of TinySC can supposedly be turned on by using a command line argument, so if I can find out how to specify the file types without using a webui I'll try transcoding FLAC to PCM in TinySC and see what difference that makes.

I gotta go now, We'll see what happens with more tests.

John S.

dean
2009-11-16, 09:47
On Nov 16, 2009, at 8:20 AM, technobear wrote:
> I can hear a clear difference using a Transporter between FLAC decoded
> on a PC and FLAC decoded on the Transporter.

If this is reproducible (i.e. measurable) then it's a serious bug in Transporter and would need to be fixed. Same with Touch.

IIRC, as tested in the lab, the audio performance of Touch was not measurably different based on the CPU load.

If you can't measure it, then you can't fix it.

andyg
2009-11-16, 09:51
On Nov 16, 2009, at 11:44 AM, JohnSwenson wrote:

>
> Let me reitterate, In these tests the same FLAC file was used for all
> tests. TinySC as commonly configured does not transcode, so when playing
> a FLAC file it can only send a FLAC stream to the player. Currently it
> does not support WAV files so I can only hear FLAC decoding from TinySC
> right now.

I will try and get to the bottom of why WAV/PCM does not work today, so you can test that too.

JohnSwenson
2009-11-16, 13:30
If you can't measure it, then you can't fix it.

There is a test which can measure the audio effects of very small amounts of jitter, its actually quite simple: play a 15KHz sine wave, run it through a 24 bit ADC and run a very high resolution FFT. Someone did the math which shows that you should be able to see jitter differences in the 5ps range with this test.

There is one drawback, the jitter of the ADC has to be significantly lower than the jitter of the tested system. The best ADC I have right now has about 50ps jitter which makes it unusable for this test. I have one which would be fairly easy to add some very low jitter clocks to but that will take some time and money, neither of which I have right now so its going to be a while before I could do such a test.

John S.

JohnSwenson
2009-11-16, 13:33
I will try and get to the bottom of why WAV/PCM does not work today, so you can test that too.

Thanks Andy. My guess is that its just a file type thing and WAV has been turned off in the file types list. I'm not sure what file to look in for this since I always use the webUI to change it.

John S.

andyg
2009-11-16, 13:58
Yep, WAV was broken because of --notranscoding, but it was easily fixed.

technobear
2009-11-16, 18:39
Yep, WAV was broken because of --notranscoding, but it was easily fixed.

That's excellent news. Top stuff Andy http://www.technobear.btinternet.co.uk/emoticons/thumbsup.gif

Phil Leigh
2009-11-17, 01:14
C'mon guys, Andy is right.
If it's to discuss these matters, it's better to bring this discussion to the Audiophile forum.

The SB Touch forum is a general forum, and most people don't understand the nuances of your argumentation.

Agreed. Let's take this discussion elsewhere.

pippin
2009-11-17, 05:04
On Nov 16, 2009, at 8:20 AM, technobear wrote:
> I can hear a clear difference using a Transporter between FLAC decoded
> on a PC and FLAC decoded on the Transporter.

If this is reproducible (i.e. measurable) then it's a serious bug in Transporter and would need to be fixed. Same with Touch.


Or it's a serious bug in whatever does the decoding on the PC.

Phil Leigh
2009-11-17, 05:22
Or it's a serious bug in whatever does the decoding on the PC.

That's a very valid point.

technobear
2009-11-17, 08:03
Guys, the tiny differences we are talking about here are NOT ABOUT THE 1's AND 0's.

How many more times!

The distortions that we are hearing here are not caused by missing or wrong bits.

They are far more subtle than that.

They fall into two broad categories - amplitude domain distortion and time domain distortion - commonly known as noise and jitter as I said in an earlier post.

Where these distortions arise and why is a matter for investigation but it has nothing to do with the software or the decoding - that was perfected long ago.

Phil Leigh
2009-11-17, 08:18
Guys, the tiny differences we are talking about here are NOT ABOUT THE 1's AND 0's.

How many more times!

The distortions that we are hearing here are not caused by missing or wrong bits.

They are far more subtle than that.

They fall into two broad categories - amplitude domain distortion and time domain distortion - commonly known as noise and jitter as I said in an earlier post.

Where these distortions arise and why is a matter for investigation but it has nothing to do with the software or the decoding - that was perfected long ago.



You are making an assumption that a "bug" is always in the software... I never said that. It is indeed nothing to do with the zeroes and ones.
Whatever does the decoding includes the hardware...

pfarrell
2009-11-17, 08:24
technobear wrote:
> How many more times!

Why are you ranting? The device is not released yet.

> They fall into two broad categories - amplitude domain distortion and
> time domain distortion - commonly known as noise and jitter as I said in
> an earlier post.

You just described all possible errors in any audio devices. What is
your point?

> Where these distortions arise and why is a matter for investigation but
> it has nothing to do with the software or the decoding - that was
> perfected long ago.

On what basis do you make this claim?

In most modern audio devices, the distinction between hardware and
software is much more fuzzy than you are making it out to be. Nearly
everything has a DSP.

My troll detector is starting to move the meters.

--
Pat Farrell
http://www.pfarrell.com/

dean
2009-11-17, 09:01
Ah, "sample rate drift". Very interesting. Sounds likes samples are being dropped.

Is it possible to do this with the digital output?

On Nov 17, 2009, at 12:08 AM, Phil Leigh wrote:

>
> dean;486074 Wrote:
>> On Nov 16, 2009, at 8:20 AM, technobear wrote:
>>> I can hear a clear difference using a Transporter between FLAC
>> decoded
>>> on a PC and FLAC decoded on the Transporter.
>>
>> If this is reproducible (i.e. measurable) then it's a serious bug in
>> Transporter and would need to be fixed. Same with Touch.
>>
>> IIRC, as tested in the lab, the audio performance of Touch was not
>> measurably different based on the CPU load.
>>
>> If you can't measure it, then you can't fix it.
>
> Dean/JS:
> http://forums.slimdevices.com/showpost.php?p=486288&postcount=2
>
>
> --
> Phil Leigh
>
> You want to see the signal path BEFORE it gets onto a CD/vinyl...it
> ain't what you'd call minimal...
> SB Touch Beta (wired) - TACT 2.2X (Linear PSU) + Good Vibrations S/W -
> MF Triplethreat(Audiocom full mods) - Linn 5103 - Aktiv 5.1 system (6x
> LK140's, ESPEK/TRIKAN/KATAN/SEIZMIK 10.5), Townsend Supertweeters, Blue
> Jeans Digital,Kimber Speaker & Chord Interconnect cables
> Kitchen Boom, Outdoors: SB Radio
> ------------------------------------------------------------------------
> Phil Leigh's Profile: http://forums.slimdevices.com/member.php?userid=85
> View this thread: http://forums.slimdevices.com/showthread.php?t=70979
>
>

technobear
2009-11-17, 09:46
Is it possible to do this with the digital output?


Good question! I wasn't using an external DAC when I made the comparison. I only listened to the Transporter's analogue outputs.

If I can find the FLAC files, I'll give it a go using an external DAC (a Beresford Caiman).

Phil Leigh
2009-11-17, 10:00
Ah, "sample rate drift". Very interesting. Sounds likes samples are being dropped.

Is it possible to do this with the digital output?



Dean, I'm not sure - I'm at the boundary of my knowledge. As far as I understand things, no samples are being "dropped" or "repeated", but the clock frequency in the DAC is wandering (very slightly) so all of the samples are playing, but some of them are not at exactly the right time. If this was exaggerated, we'd perceive it as something like the "wow" or "flutter" of a turntable, but it's obviously much, much more subtle than that.

I have no way of measuring the s/pdif output except by playing it through an external DAC. Clearly no samples are dropped etc otw DTS wouldn't work.

I'm off to research the audible effects (if any) of sample rate drift.

Merlinwerks
2009-11-17, 10:38
Tonight I did some careful listening with the touch, FLAC and PCM streaming and TinySC. I can't listen to Wav on TinySC since it does not support WAV at this point (I hope it does in the future).

John S.

Hi John,

When you conducted these tests, how much delay, if any, was there when switching between streams?

Thanks

radish
2009-11-17, 10:48
I have no way of measuring the s/pdif output except by playing it through an external DAC. Clearly no samples are dropped etc otw DTS wouldn't work.


I have a decent sound card with SPDIF in which I could hook up to the Touch and run audiodiffmaker against if that would be at all useful.

Phil Leigh
2009-11-17, 11:02
I have a decent sound card with SPDIF in which I could hook up to the Touch and run audiodiffmaker against if that would be at all useful.

Radish - that would be great - can you download my test file
http://rapidshare.com/files/307731080/WAV_Limelight.wav.html
and make a flac clone? - then play the wav and the flac from the s/pdif out of the SB and record using your soundcard, then load the files into ADM?

Please let us know what you find.
Regards
Phil

PS Don't forget to set File types to stream wav as wav!!!!

radish
2009-11-17, 11:24
Will do. Can you let me know the settings you used in AD? I seem to remember being a little bewildered last time I played with it and it makes sense to me to be as consistent as possible.

Phil Leigh
2009-11-17, 11:36
Will do. Can you let me know the settings you used in AD? I seem to remember being a little bewildered last time I played with it and it makes sense to me to be as consistent as possible.

For the s/pdif test you want your sound card to get it's clock from the incoming stream - not sure how you do that as it is card-dependant.
If the card rate and s/pdif rate are different you will get clicks/pops which will rather distort things...

The only setting you might want to change is to check the "compensate for sample rate drift".
Try it with the defaults first and see what happens.
You might also want to make 2 seperate wav and flac recordings and compare them against each other as a control. I found the 2 wavs and 2 flacs compared pretty well to each other - but there was a big difference going cross-format...

JohnSwenson
2009-11-17, 14:24
I'll do a S/PDIF to file test tonight, if the bits are the same that SHOULD give a complete null.

I agree, this should probably be over in the audiophile forum, there is already a thread there.

John S.

JohnSwenson
2009-11-19, 21:52
Just to let everyone know that WAV now works fine in TinySC.

Thanks Andy!

John S.