Hello everyone. Recently I perused the following page written by Lampizator: CD_transport_DIY and I was interested in finding out why he recommended using balanced output instead of unbalanced spdif output between CD player and DAC. I thought about it and it came to me: why not try balanced output on my professional soundcard Prodif-88? You see. When I purchased this professional card I thought that since my DAC has spdif input then I had to convert aes/ebu signal into spdif signal and feed my DAC with this signal. I have been using this eas/ebu to spdif conversion for a few years until I stumbled across this piece of advice from lampizator. A few days ago I tried this method and I was very much surprised by how much it improved the sound of my soundcard. I just put twisted pair between aes/ebu output of the card and input of the DAC. I didn't even use any ground shield and the sound improved and was significantly better than via spdif cable. Why is that, I wonder? In theory balanced connection is useful only for long cables but it turns out that even at 1 meter length balanced digital connection is better than unbalanced one. Does it mean that spdif stream loses some bits when the signal gets transferred from the soundcard via spdif cable?
No. Digital audio either delivers perfect bits or corrupted bits. Corrupted bits will sound horrible, so if it doesn't sound horrible then the bits are perfect. You cannot improve on perfection.
However, there is also the matter of timing. The best timing (i.e. lowest jitter) comes from using the right cable, not the wrong cable. For consumer SPDIF with an unbalanced 75ohm input the right cable, the best cable, is 75ohm coax. If something else appears to sound 'better' than this then either there is no change and you are fooling yourself, or the wrong cable actually sounds worse (by increasing jitter) and you are fooling yourself.
As a general rule, lampizator's advice is best ignored as he seems to have his own private understanding of how things work.
However, there is also the matter of timing. The best timing (i.e. lowest jitter) comes from using the right cable, not the wrong cable. For consumer SPDIF with an unbalanced 75ohm input the right cable, the best cable, is 75ohm coax. If something else appears to sound 'better' than this then either there is no change and you are fooling yourself, or the wrong cable actually sounds worse (by increasing jitter) and you are fooling yourself.
As a general rule, lampizator's advice is best ignored as he seems to have his own private understanding of how things work.
You see the thing is that my set up is a little bit more complicated and I did not tell you the whole truth, all the details. The thing is that I am using a DAC with I2S input and a SPDIF to I2S converter. Following an advice by lampizator I tried to connect sound card with the SPDIF to I2S converter by a balanced twisted pair instead of coaxial spdif cable and it improved. I couldn't believe my ears. You see the converter has synchronous reclocking and the DAC has a synchronous reclocking circuit just before I2S is fed to the DAC chip. As you might understand since this is an SPDIF to I2S converter then it just extracts 3 I2s signals from the input but does not extract or use masterclock. Therefore there's no jitter to talk about. Moreover I2S signals get reclocked 2 times one time in the converter and then in the DAC. And there's one more experiment that I would like you to explain to me. Some time ago I inserted a throttle on the spdif cable between the sound card and the converter and the sound deteriorated and I couldn't listen to it for me than a few minutes. Then I tried to put a ferrite clamp on the same spdif cable and again I noticed some kind of slowness in the sound. How can that be? The converter does not extract masterclock from the signal. So jitter is no issue. Why does it happen? Any ideas? I have come to the only conclusion from all these experiments that somehow for some reason some bits get lost but this contradicts the theory, doesn't it? There's one theory that explains all these changes in sound by RF or EMI interference on the spdif cable whereas balanced connection is immune to RF interference.
Last edited:
I'm pleased that you don't believe your ears. I don't believe them either.momitko said:
No idea what you are talking about. The only throttle I have is in my car.
It can't.
It contradicts facts, and so is therefore false.
You made some changes to your system, and you thought it sounded better. How unusual is that? These apparent improvements in sound violate known engineering principles; how uncommon is that among audiophiles?
You have a choice: believe the truth, or believe your ears. Up to you. The electrons don't care as they always follow Maxwell's equations whether you know this or not.
My apologies for an error. It was a common mode choke. Someone told me that a common mode choke might reduce HF noise emanating from PC. So tried it and I heard deterioration of the sound. This is true. Judging by your replies you do not believe me but I am telling you that the sound DID change althouth it shouldn't.
There are no bit errors (they will sound really bad) however as much as everyone likes to think digital is only on or off you still have ground issues and all sorts of odd interference and modulation unless everything is "perfect" which I dont think any equipment is. More likely its a sort of balance and depending on your other stuff and what you do it will sound better to you or not.
Tweak it until you like it and by all means use the knowlegde on here but apply it with understanding and based on what you have and are looking for.
Tweak it until you like it and by all means use the knowlegde on here but apply it with understanding and based on what you have and are looking for.
A common-mode choke could help reduce common-mode noise over an unbalanced interface, however, they also exhibit parasitic reactances which will screw-up the impedance matching of the interface, especially at the high frequencies involved. For example, a common-mode choke's leakage inductance (which is due to the less than perfect magnetic coupling of the two coils) will manifest as a normal-mode impedance. As such, it will disrupt the impedance matching of the line.
It's an interesting explanation that you used. Can you explain how in hell this disruption of impedance could have caused the change of sound? And please don't forget that this was taking place on spdif cable between the sound card and the spdif to i2s converter which does not restore masterclock. So this can't be expained away by jitter.
You see the thing is that I have a friend who believes that computer is not in any position to compete with CD transport. He does not think that synchronous reclocking and USB to I2S or SPDIF to I2S converter can neutralise horrific amounts of jitter coming from computer. A few years ago I got myself a DIY DAC with I2S input and spdif to I2s converter and I tried to implement this method but I had some problems with grounding and shielding i2s cable and due to this his CD players outperformed my system. This only hardened his position. I tell him: Listen, synchonous reclocking completely reclocks all 3 i2s signals. So any jitter is killed. But he won't believe me and we argue via Skype all the time about all these issues. There is one more thing that I came to in my experiments. When I was tinkering with my DAC I noticed that the sound improved when I shielded i2s cable despite that all 4 i2s signals are transmitted in balanced mode. And the best sound was achieved only when I made separate shields for all 4 i2s signals. Can you believe that? When I put all 4 signals inside one shield the sound was better than without a shield but I heard some kind of distortion and only after placing every single i2s line inside a separate shield the sound improved significantly. How can you explain that? When I was content that I have rectified all the problems inside my DAC he came to my place with his CD player and again his CD player outperformed my system. This was a shocker to me. I didn't know what to do and then out of the blue I stumbled across the webpage by lampizator where he recommends using balanced mode instead of regular spdif and I try it and hear a significant change in the sound of my system. Now, I think, my system can compete with all his CD players.
Last edited:
For your information the last time my friend came up to my place and we held a comparison test I a created a topic about it on this forum here: www.diyaudio.com/forums/digital-source/291196-jitter-bitclock.html It was half a year ago and this topic is a continuation of that topic.
If you connect a balanced digital output to an unbalanced digital input (with a different impedance too, IIRC) then there will be some common-mode signal and also the possibility of common-mode to differential-mode mixing and interference. Adding a common-mode choke could modify this. It should not change the sound unless bits were being corrupted. However, the solution to corrupted bits is to use a proper connection - not patch up a wrong connection.
People often use the wrong cable in their audio system (whether through ignorance, DIY or 'high-end snake oil') and find that it changes the sound. This change, if genuine, will always be for the worse - physics guarantees this. However, having spent money or time on making the change it is very common for the change to be perceived as an improvement.
People often use the wrong cable in their audio system (whether through ignorance, DIY or 'high-end snake oil') and find that it changes the sound. This change, if genuine, will always be for the worse - physics guarantees this. However, having spent money or time on making the change it is very common for the change to be perceived as an improvement.
In this case the change has definitely been for the better, the sound improved, I started hearing more information at HF spectrum. Besides that the sound card has a native balanced output. Why would a designing engineer design a balanced output?
There are essentially three mechanisms for how a CMC choke could change the sound, two are jitter related. One mechanism is by causing an impedance mismatch (or bandwidth limiting) of an otherwise impedance matched transmission line of sufficient bandwidth. I would need to know the details of your soundcard->I2S converter->DAC interface in order to eliminate this as a possibility.
A second mechanism is by suppressing common-mode noise (assuming the DAC's input isn't already transformer coupled) via the unbalanced S/PDIF link. So, rather than hearing the introduction of jitter due to impedance mismatch it's possible you are hearing the altering or suppression of pre-existing jitter due to the suppression of common-mode noise.
The third mechanism is by common-mode noise coupling to the analog stages of the DAC, or to the analog stages of a following preamp or amp through the DAC.
Last edited:
I wonder if anyone could give me a link to any resource with detailed explanation on synchronous recklocking, I mean an article with detailed information on how this process works. Some people say that recklocking can not eliminate all jitter. I suppose this is valid only for asynchronous recklocking but not synchronous recklocking. Do you know any white papers on this subject?
There's a scientific term for believing something's better just on expectations. It's a placebo effect. If you give an emotionally drug addicted person a placebo solution they might feel relieved too.
Yeah. I have heard this term quite a lot. But in this particular case this term does not apply. You can try the same experiment as I did: just change spdif connection to a balanced one or insert a common mode choke in spdif line. And how will you explain the reason for change of sound when I put a separate shield on each i2s line instead of 1 shield for all 4 i2s signals?
If there is audible difference in sound it usually is because the system is poor in jitter rejection.
Placebo effect can still happen even when you know about it. i2s signals are not meant to be sent along cables, but short PCB traces only.
You did not read my post thoroughly. I sent I2S signals in balanced mode via RS-485. Some people use LVDS for I2S and some use RS-485. In my case there was no placebo. You can easily try the same mod as I did and I am sure you will hear the same change of sound.
Last edited:
For sure it is placebo.
If it not, than is only the common mode noise that is not correctly filtered by very audiophile equipment's (to be read incomplete and poorly made)
I like very much how every day appear new people that know the absolute truth and the thousands of engineers who have studied the phenomena of thousands of hours with equipment that ordinary people do not even think that exist, failed to discover it.
If it not, than is only the common mode noise that is not correctly filtered by very audiophile equipment's (to be read incomplete and poorly made)
I like very much how every day appear new people that know the absolute truth and the thousands of engineers who have studied the phenomena of thousands of hours with equipment that ordinary people do not even think that exist, failed to discover it.
- Status
- Not open for further replies.