Why not also trust your ears - that's the whole point of this hobby - if it doesn't sound any different then why bother going to the analogue measurements step? However if it does sound different & you can't find it on the analogue measurements - what does that tell you - your ears are wrong? And don't say it's psychology - it obviously sounds different in my & in other systems that I've heard & have been reported to me!
Hi,
By gosh. This is BAD. Sure, there is even worse, but that is such a bad result, no wonder the (near) stock Musilands often sounds better! It may have wrong impedances and levels and a noisy output, but it is not that bad.
Here, strictly for fun, a tube driven, transformer coupled SPDIF output from me, at 48KHz Btw. Before anyone asks, this is not the Lampizator design, not even remotely.
This is driven directly of the FPGA Pin on a modified Musiland PCB and it includes 2m of Gotham Cable 75 Ohm Coax and Canare RCA Sockets and Plugs (the so-called 75 Ohm variety) on both ends. The termination on the receiving socket using a pair of 0805 SMD 150 Resistors, to minimise inductance, scope probe correctly calibrated and set to 1:10...
So, if you wire your DAC's input well and correctly, this is what the receiver would see, or at least as close as we can make it...
Using a terminated RCA Plug directly on the system output looks much better, but is of course meaningless.
The Zoom level shows around 1/3rd of one square "wavelet", so this a much closer up view of the leading edge than most of the shots we see posted on-line, just in case anyone needs any references for scale.
Ciao T
By gosh. This is BAD. Sure, there is even worse, but that is such a bad result, no wonder the (near) stock Musilands often sounds better! It may have wrong impedances and levels and a noisy output, but it is not that bad.
Here, strictly for fun, a tube driven, transformer coupled SPDIF output from me, at 48KHz Btw. Before anyone asks, this is not the Lampizator design, not even remotely.
This is driven directly of the FPGA Pin on a modified Musiland PCB and it includes 2m of Gotham Cable 75 Ohm Coax and Canare RCA Sockets and Plugs (the so-called 75 Ohm variety) on both ends. The termination on the receiving socket using a pair of 0805 SMD 150 Resistors, to minimise inductance, scope probe correctly calibrated and set to 1:10...
So, if you wire your DAC's input well and correctly, this is what the receiver would see, or at least as close as we can make it...
Using a terminated RCA Plug directly on the system output looks much better, but is of course meaningless.
The Zoom level shows around 1/3rd of one square "wavelet", so this a much closer up view of the leading edge than most of the shots we see posted on-line, just in case anyone needs any references for scale.
Ciao T
Attachments
Thorsten, I've investigated both the Musiland & the Hiface & in stock form the Hiface easily beats the Musiland sonically - go figure. Obviously the above doesn't tell the full picture of the sound. I've modded the Musiland (detailed in this forum) & I would say that in modded form it comes up to about the sound of the stock Hiface.
So agreed the scope waveform does not look good but there's something else going on not shown! Of course all this can be improed & I have scope shots similar to what you show.
Remember those two ol' codgers that sat on the balcony of the Muppets show & criticised everything?
Well here's Cartman & Snotgreen (some residue left after Jocko sneezed, I presume) - have a look at their comments & have a chuckle Probably none of our business............. "What do you think of it so far?" ; "Rubbish"
So agreed the scope waveform does not look good but there's something else going on not shown! Of course all this can be improed & I have scope shots similar to what you show.
Remember those two ol' codgers that sat on the balcony of the Muppets show & criticised everything?
Well here's Cartman & Snotgreen (some residue left after Jocko sneezed, I presume) - have a look at their comments & have a chuckle Probably none of our business............. "What do you think of it so far?" ; "Rubbish"
Hi,
I think the "A beats B" depends heavily on the context. Both Hi-Face and Musiland have severely "faulty by design" SPDIF implementations. So which works better has to a degree at least to do with the rest of the system. And of course with the PC setup itself. For example, HiFace is limited to Kernel-Streaming, which may or may not have an impact (hard to predict)
I have both devices available. Once you fix the main issues in each there seems little to choose, except for Musiland offering ASIO.
With the SPDIF output as designed the HiFace actually needs an attenuator to comply with SPDIF standards.Several of the truely high erformance options for SPDIF receivers will seriously struggle with the original signals from both HiFace and Musiland.
The usual "App-Note" cirrus logic circuitry will barely lock on the Musiland's (which will generate rather high jitter) low signal levels, while the generic cirrus logic circuitry which is optimised for 5V PP AES/EBU levels will love the rather strong signal level from the HiFace.
On the other hand, sensible active SPDIF input circuitry (like I use) that deals with the low levels from a stock Musiland very well will saturate with the hi-face output and throw major wobbles, causing high jitter.
So making any such comparisons without stating the context is of very limited use.
Hence the pains I took to illustrate my test conditions. And again, please remember my measurements include the cable and the receiver side termination.
Ciao T
I think the "A beats B" depends heavily on the context. Both Hi-Face and Musiland have severely "faulty by design" SPDIF implementations. So which works better has to a degree at least to do with the rest of the system. And of course with the PC setup itself. For example, HiFace is limited to Kernel-Streaming, which may or may not have an impact (hard to predict)
I have both devices available. Once you fix the main issues in each there seems little to choose, except for Musiland offering ASIO.
With the SPDIF output as designed the HiFace actually needs an attenuator to comply with SPDIF standards.Several of the truely high erformance options for SPDIF receivers will seriously struggle with the original signals from both HiFace and Musiland.
The usual "App-Note" cirrus logic circuitry will barely lock on the Musiland's (which will generate rather high jitter) low signal levels, while the generic cirrus logic circuitry which is optimised for 5V PP AES/EBU levels will love the rather strong signal level from the HiFace.
On the other hand, sensible active SPDIF input circuitry (like I use) that deals with the low levels from a stock Musiland very well will saturate with the hi-face output and throw major wobbles, causing high jitter.
So making any such comparisons without stating the context is of very limited use.
Hence the pains I took to illustrate my test conditions. And again, please remember my measurements include the cable and the receiver side termination.
Ciao T
If I hear a difference under blind conditions (i.e., trusting my ears, not my eyes) and there's no difference in the analog signal, there's probably a Nobel Prize for me.
If I don't hear a difference, that doesn't mean no-one else can, so the analog measurements are of interest. Since you haven't done any controlled testing for audibility (or if you have, you've kept it a secret), then I'm mystified at how you can blithely dismiss even the possibility that the differences are psychological.
Good points, Thorsten & indeed you are correct - comparisons without context are of limited use. I forget the set-up I used, I would have to look back at any notes I might have taken.
Your plot is impressive! You've used a tube OPS, different output transformer as well as some PS changes & other tricks? I see you've used those Canare RCA sockets & plugs but how does one of these RCA sockets perform if being plugged into a non-Canare RCA socket - impedance issues?
Your plot is impressive! You've used a tube OPS, different output transformer as well as some PS changes & other tricks? I see you've used those Canare RCA sockets & plugs but how does one of these RCA sockets perform if being plugged into a non-Canare RCA socket - impedance issues?
If I hear a difference under blind conditions (i.e., trusting my ears, not my eyes) and there's no difference in the analog signal, there's probably a Nobel Prize for me.
Since you haven't done any controlled testing for audibility (or if you have, you've kept it a secret), then I'm mystified at how you can blithely dismiss even the possibility that the differences are psychological.
Because you only have to resort to psychological explanation when the differences are of a slight & undefined nature i.e. does this cable sound better than that one type comparisons? I know when the threshold of doubt has been passed & am confident in what I hear! This applies more so to the stock Vs modified Hiface differences & to a lesser extent, the attenuators!
Last edited:
Hi Sy,
And I am mystified how you can blithely dismiss even the possibility that the absence of observed differences are psychological (expectation bias).
I got caught out by that myself a few times (not the way you think - my DBT showed "perfect random" in cases where I had expectations, while others in the same tests showed statistical significant indications of audible differences).
The bottom line is that a Blind test is not "Blind" if you are aware of the subject and have expectations with regards to the outcome. Any expectation bias increases the "noisefloor" dramatically. Certainly individual test cannot overcome it, an n= 50...100 seems a sensible minimum sample size in case expectation bias is present.
Ciao T
PS, fifteen people all individually scoring "random" in fifteen trials may still produce a statistically significant "positive" if the results are combined.
And I am mystified how you can blithely dismiss even the possibility that the absence of observed differences are psychological (expectation bias).
I got caught out by that myself a few times (not the way you think - my DBT showed "perfect random" in cases where I had expectations, while others in the same tests showed statistical significant indications of audible differences).
The bottom line is that a Blind test is not "Blind" if you are aware of the subject and have expectations with regards to the outcome. Any expectation bias increases the "noisefloor" dramatically. Certainly individual test cannot overcome it, an n= 50...100 seems a sensible minimum sample size in case expectation bias is present.
Ciao T
PS, fifteen people all individually scoring "random" in fifteen trials may still produce a statistically significant "positive" if the results are combined.
Hi,
Yes, quite a few mods. Powersupplies, clocks, etc.
To be honest, I have only compared Canare RCA to Canare BNC (Canare BNC does better, but it takes close comparison) and some generic locking RCA Plugs on Canare sockets with non 75Ohm cable (way bad).
But the Canare RCA stuff is not expensive and it is possible to upgrade everything from Source to DAC. Of course, Canare RCA's are not true 75R, but they seem non too bad in practice.
BTW, the way the FPGA Pin's are wired to the Output stage also has a major impact on waveform and thusly to a degree on jitter.
Ciao T
Yes, quite a few mods. Powersupplies, clocks, etc.
To be honest, I have only compared Canare RCA to Canare BNC (Canare BNC does better, but it takes close comparison) and some generic locking RCA Plugs on Canare sockets with non 75Ohm cable (way bad).
But the Canare RCA stuff is not expensive and it is possible to upgrade everything from Source to DAC. Of course, Canare RCA's are not true 75R, but they seem non too bad in practice.
BTW, the way the FPGA Pin's are wired to the Output stage also has a major impact on waveform and thusly to a degree on jitter.
Ciao T
Why you made this comment is a mystery to me.
I haven't been able to discover anywhere that SY has done so, in this thread at least. Your contribution has not been unhelpful, up to this point, but remarks such as this merely tend to reduce any confidence other readers have in the rest of your assertions.
w
We all find it amazing that you don't. After all there's a lot more evidence supporting our point of view than there is yours.
w
Let's not get into us & them - I expressed an opinion that SY might want to answer! If you want to answer for yourself then go ahead & say at what point you no longer have to resort to measurements to be sure of differences in sound - can you tell the difference between an acoustic guitar & an electric guitar for instance?
Last edited:
Hearing has nothing to do with it. My brain, like all human brains, is quite subject to error- that's why I enjoy watching skilled conjurers. Unlike many others, I freely admit it.
- Status
- Not open for further replies.