TNT: That may have been what this thread was originally about, but it was pulled wildly off topic. I didn't even know about attenuators until John's unit arrived here. The tests I offered to do in another, now-closed, thread (and will still do) are all analog domain tests. I'll throw the attenuator into the mix as well, but still, the measurements are entirely analog measurements. That's all that really counts.
I suggest everybody should send their equipment to SY before laying down hard cash - just with a few measurements he can tell you that device A sounds exactly like device B - the rest is Booga-Booga
1-Attenuator reduces signal level so reduce jitter also, that's logical where ilogical if only reduces only one thing?
2-Or if signal level is too high the attenuator reduces signal & the DAC works better?
So my point of view it's all depends your system, because if you don't have a high level signal & you will use the attenuator you decrease jitter but decrease also signal, what do you prefer? always you get something but you loose something too....
I only believe that attenuator is right for high level signal, it's the only application that's make sense, other things is "snake oil".
2-Or if signal level is too high the attenuator reduces signal & the DAC works better?
So my point of view it's all depends your system, because if you don't have a high level signal & you will use the attenuator you decrease jitter but decrease also signal, what do you prefer? always you get something but you loose something too....
I only believe that attenuator is right for high level signal, it's the only application that's make sense, other things is "snake oil".
I started this thread introducing these attenuators with my premise about how they worked (btw, Merlin, the reflections are doubly attenuated compared to the signal). I never said I had the answers, it was meant as a cheap experiment for people to try & maybe find some answers. 45 pages later & one person has tried them! I think this says a lot about how this thread was wildly taken off topic early on & something about the DIY part of this forum. We are now at the stage of "testing" these.
Last edited:
But wasn't that question already answered by some of the links to e.g. Stereophile articles. I subscribe to that publication (oki oki don't shoot me) and they test very consistently a verity of products in different price classes and I see differences between all products and so does the listening test fall out as well.
What will you really contribute to wrt. the state of current general knowledge?
What will you really contribute to wrt. the state of current general knowledge?
Hmmm, that's worth pondering!! What exactly is SY's test about? Maybe Stereophile should just forget about all the expensive equipment they use & get down to the nitty gritty - just do analogue output waveform testing - that's where it matters. Jeez, I'm going to tell them straight away!
That's the least of Stereophile's problems. But their business is selling magazines and advertising space, so entertainment is far more important than technical relevance.
Don't worry, you are not alone 😀
Attached is Db <=> Ratio attenuation table. Just in case if you will need this in further discussion 😉
Attachments
Last edited:
Thanks storm, I have this table but from my own company, Minicircuits 😀 but I see you bought from the competition 
. BTW, check out the female ends of those Suhner ones - do they not look like 50ohm configuration? Joseph K already agreed with this observation! Shh, don't tell Waki or I'll get accused of bringing the reputation of all electronic engineers into dis-repute - imagine that!
They look like this right?
. BTW, check out the female ends of those Suhner ones - do they not look like 50ohm configuration? Joseph K already agreed with this observation! Shh, don't tell Waki or I'll get accused of bringing the reputation of all electronic engineers into dis-repute - imagine that!They look like this right?
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Complete nonsense. Jitter can be seen in an FFT of the original digital signal. See below. If those same side bands can be seen in the analog signal, then your equipment is sensitive enough to detect the jitter.
Once this is established then jitter reducers can be tested because we know that the test equipment can see at least some level of jitter. In the plot below, fairly small jitter. (Stereophile jitter test track)
That will NOT tell you that that "Device A sounds exactly like device B" it will just tell you if the jitter reducer is really doing something. If you can hear the difference, it will show up in a test like this.
Attachments
Hehe, "looks like" and "they are" are two different things. Don't mix them , your eyes may trick you, like Waki posted here 😀
Yes, I'm guilty for owning those Suhner attenuators before name "Minicircuits" become known and before you found datasheet posted HERE.
At this time, you were still waiting for your Minicircuit attenuators to arrive, bigmouthing HERE and trying to rip off several unaware and less knowledgeable people. After that they banned you.
As posted several times, where is your problem, why can't you give someone credit for something you learned from him?
PS: I bet when you put together (or buy) TDR, you will start to teach Jocko, Joseph K and others 😀
, your eyes may trick you, like Waki posted here 😀Yes, I'm guilty for owning those Suhner attenuators before name "Minicircuits" become known and before you found datasheet posted HERE.
At this time, you were still waiting for your Minicircuit attenuators to arrive, bigmouthing HERE and trying to rip off several unaware and less knowledgeable people. After that they banned you.
As posted several times, where is your problem, why can't you give someone credit for something you learned from him?
PS: I bet when you put together (or buy) TDR, you will start to teach Jocko, Joseph K and others 😀
Last edited:
Just to be clear, do you specifically mean the jitter on the spdif line or the jitter of the DAC's timing?
I don't know, really.
If the clock is derived from the spdif line, then would they be the same? And what about jitter in the ADC?
If the clock is derived from the spdif line, then would they be the same? And what about jitter in the ADC?
I purposely said looks like - have you tested them? Do you know if that connector is 75ohm or not? Why not show us your tests?Hehe, "looks like" and "they are" are two different things. Don't mix them
Can you not read yet? Ah, shame! Let me help you - go down two posts on the head-fi thread that you linked to & you will find this post from me
I'll do it out phonetically for you if you like?
No, I'll leave that up to the ones who can't read - I don't do machines just text!
I think you are mixing up my equipment with SY's equipment - this is a test that SY needs to produce to show that jitter can indeed be detected by him on the analogue out - something he so far has refused to do! Maybe you should ask him?
Exactly so after SY has done above he can then reliably do this step with the degree of confidence that the above step will reveal! Excellent logic, Pano.
Well, you need to take up this point with SY because his claim is that if it can be heard, it will show up on the analogue output waveform in his measurements i.e if A measures the same as B then they are identical sounding - everything else is Booga-boo
You really are trying to do a number on me - an emissary of Jocko's no doubt - here to cause trouble - I never sold the attenuators so how can I be ripping off the unwary? I got banned because I kept calling to task people who made unfounded, stupid & technically unsupported claims about the stock Hiface:
- being 3dB louder than any other transport;
- having tipped up High Frequency,
- Marco was a liar , etc
BTW, all these problems magically disappeared in the modified Hiface according to the posters on Head-fi so it was not in my commercial interest to be defending the stock Hiface - so I guess I was ripping off the unwary in that case too!
you really should read same more before utterances like these or are you just a puppet for Jocko?
Last edited:
Roll up all you children, have a whack at the Piñata, come on, don't be shy - I'm used to it by now!
Last edited:
Thorsten,
It's funny, those AP plots made by J. A. in the linked Stereophile article are almost exactly the same test setup like my EMU plots here, so are directly comparable..
I mean it's the same test signal, same frequency, same freq. window.
Mine plots are having more dynamic range and resolution- due to his MF Dac.
One combination with higher resolution capability would be an CS8414+ESS dac, with it's internal ASRC bypassed..
The Bel canto unit + MF Dac measures as bad as the ESIu2A plot which I had shown here. That means several nanosecs jitter! In fact he cites 2.91nsec pp jitter.
In the same time, his direct digital testing of the SPDIF line is giving less insight then what I had shown here with the Tek jitter analysis package. They should invest..
By the way, his numbers are a good chance for me to scale a bit my results with the two Hiface units.
I did not put up the numbers, but:
The graphs shown in this post:
http://www.diyaudio.com/forums/digi...tenuators-jitter-reducers-30.html#post2352987
are equivalent of ~54 psec jitter 1sigma (or rms jitter) - if one takes the histogram like a normal distribution, which it is really not, because of the visible separation of the random (Rj) components by some amount of deterministic peaking (Dj)
The results in this plot:
http://www.diyaudio.com/forums/digi...tenuators-jitter-reducers-30.html#post2352977
instead are more close to a real random, normal distribution, and such can be taken for good like a real 37 psec 1sigma (or rms jitter) value.
Now I think the values cited by J.A for the Lindemann & Stello converters are peak. Which is a bit vague in case of normal ditributions, but let's consider 4sigma for the Hiface units: we are about 150psec and 240psec, respectively for my two units.
His numbers are around 400 psecs for both converters measured.
I would not want to say what is better here: the numbers here will be depending on the definition of peak jitter.
For me these seem to be in the same league. Would be nice to see the TIE trend track..
Finally, would like to add that I did the same jitter analysis for the EMU1616m digital output with the scope as well - but no, I don't have the data any more.
Anyway, it was ~300psec 1sigma (or rms jitter) !! Almost as bad as the Bel canto unit.. Which is simply ugly for a unit with local crystal oscillators, "asynchronous" type of operation..
PS.: Another difference: If I understood well, he used the JTest signal in the measurements.
For those plots of the Hiface, I used real music playing. That means something, because of the deterministic components it worsens the situation. The test signal is a good thing, but music is even more "brutal" in this meaning..
It's funny, those AP plots made by J. A. in the linked Stereophile article are almost exactly the same test setup like my EMU plots here, so are directly comparable..
I mean it's the same test signal, same frequency, same freq. window.
Mine plots are having more dynamic range and resolution- due to his MF Dac.
One combination with higher resolution capability would be an CS8414+ESS dac, with it's internal ASRC bypassed..
The Bel canto unit + MF Dac measures as bad as the ESIu2A plot which I had shown here. That means several nanosecs jitter! In fact he cites 2.91nsec pp jitter.
In the same time, his direct digital testing of the SPDIF line is giving less insight then what I had shown here with the Tek jitter analysis package. They should invest..
By the way, his numbers are a good chance for me to scale a bit my results with the two Hiface units.
I did not put up the numbers, but:
The graphs shown in this post:
http://www.diyaudio.com/forums/digi...tenuators-jitter-reducers-30.html#post2352987
are equivalent of ~54 psec jitter 1sigma (or rms jitter) - if one takes the histogram like a normal distribution, which it is really not, because of the visible separation of the random (Rj) components by some amount of deterministic peaking (Dj)
The results in this plot:
http://www.diyaudio.com/forums/digi...tenuators-jitter-reducers-30.html#post2352977
instead are more close to a real random, normal distribution, and such can be taken for good like a real 37 psec 1sigma (or rms jitter) value.
Now I think the values cited by J.A for the Lindemann & Stello converters are peak. Which is a bit vague in case of normal ditributions, but let's consider 4sigma for the Hiface units: we are about 150psec and 240psec, respectively for my two units.
His numbers are around 400 psecs for both converters measured.
I would not want to say what is better here: the numbers here will be depending on the definition of peak jitter.
For me these seem to be in the same league. Would be nice to see the TIE trend track..
Finally, would like to add that I did the same jitter analysis for the EMU1616m digital output with the scope as well - but no, I don't have the data any more.
Anyway, it was ~300psec 1sigma (or rms jitter) !! Almost as bad as the Bel canto unit.. Which is simply ugly for a unit with local crystal oscillators, "asynchronous" type of operation..
PS.: Another difference: If I understood well, he used the JTest signal in the measurements.
For those plots of the Hiface, I used real music playing. That means something, because of the deterministic components it worsens the situation. The test signal is a good thing, but music is even more "brutal" in this meaning..
Last edited:
- Status
- Not open for further replies.