I've started messing around with considering cross-talk and what it actually means in vinyl replay. As part of my homework on this trying to work out how cross talk varies with frequency and how variable this is.
Does anyone have any recordings of a single channel sweep that I can extract cross-talk from (i.e both channels recorded for a sweep on only one)?
Does anyone have any recordings of a single channel sweep that I can extract cross-talk from (i.e both channels recorded for a sweep on only one)?
From my personal (based mostly on listening to cartridges and amplifiers with different crosstalk values both in their in their specs and measurement results) experience for me it means - foremost - presence or lack of finest detail in sound.
How are you intending to measure it - after RIAA phonostage?
Pre-or post RIAA doesn't matter too much as I can correct for that. Initially just trying to work out what the frequency dependence is and if the phase stays the same (which it should). Cartridge cross talk is out of phase, so in theory a cross feed of an in phase signal would cancel this out.
Now here we hit conjecture, but based on some anedotal evidence doing this (basically reducing stereo width) might actually improve the centre image and 'snap things into focus'. You will lose some of the out of phase 'vinyl ambience' so its not clear which would be preferable. I'm just interested in the actual scale of the problem and how much correction could actually be done.
Now here we hit conjecture, but based on some anedotal evidence doing this (basically reducing stereo width) might actually improve the centre image and 'snap things into focus'. You will lose some of the out of phase 'vinyl ambience' so its not clear which would be preferable. I'm just interested in the actual scale of the problem and how much correction could actually be done.
How exactly are you going to "correct" the crosstalk (both originating from the cartridge and the RIAA-stage)?
That depends to a degree on what the frequency dependency is. It might be as simple as splitting the L-R signal into M-S (mid side) and filtering the side with a tiny bit of attenuation. Stereo width controls have gone out of fashion, but they may have a use.
As I said entirely conjecture at the moment
As I said entirely conjecture at the moment
I don't think that there is any RIAA anything like as poor as even the best cartridge
I can record the Ortofon test LP tonight and upload. With respect to phase and cross-talk you will get in and out of phase of the crosstalk signal depending on how you turn it. Optimal is when you have the crossing.
An externally hosted image should be here but it was not working when we last tested it.
And as I wrote on VE about the azimuth measurements and interpretations on the korf blog. I think he is wrong since generator tilt (and not stylus tilt) will cause the effects he is measuring:
Azimuth Adjustment Myths and Reality, Part III
Azimuth Adjustment Myths and Reality, Part III
If you have a perfect symmetry of goove walls and generator ie +\- 45 degrees you would theoretically get zero crosstalk. Turning the generator 2 degrees will cause +43/-47 degrees of the generator and for one channel the will be a partial vertikal movement and for the other channel partial horisontal movement. So crosstalk signals will be in phase for one channel and out-of-phase for the other. Turning the cartridge the opposite diection will just switch the in-and out of phase between the channels. I’ve done these measurements many times and the Feickert software shows the same thing. They had many examples for this on the web before but have not been able to find them lately. The further effects of this is that for a mono signal both channels, the out-of- phase crosstalk adds to the signal and and cause a slight decreased output. The opposite happens in the other channel, the in phase signal will add to the signal and increasw output. In addition, the crosstalk leads to different distortion in the two channels.
That is why I think the Korf blog is wrong. The generator tilt may well cause the measured effects and have nothing to do with the stylus azimuth. Stylus contact patch is also within reasonable limits rather insensitive for azimuth for conical and classical elliptic stylii. For fine-line stylii you may get the contact patch to ride higher and lower for right/left channel if azimuth is off. This may lead to different noise for left and right channels and perhaps less tracking ability. One can also look at Yosh excellent pages about stylii and contact patches and test his Excel simulations.
When measuring crosstalk of a "typical" preamp then you see something like on the attached graph.
Most probably a similar phenomenon takes place in the cartridge body where mechanical movement of the cantilever is converted to output voltage. And judging from the quite different values given for different cartridges the crosstalk incurring during that conversion may be of significantly different level (or character).
I wonder how different results (also frequency-domain wise) the voltage generators of different cartridges can produce ...
Attachments
ok well with a cartidge good is -25dB at 1kHz and excellent is -30dB. Jolida claims they can get to -40dB with their magic box. At 10kHz it is always lower. I haven't seen meaurements sub 1kHz.
Stylus azimuth is one cause of this
Generator azimuth is another
the general flailing of the cantilever in the groove is a 3rd
There may be others, right now I assume they are down in the noise
By 'cartridge body' do you mean the generator?
Stylus azimuth is one cause of this
Generator azimuth is another
the general flailing of the cantilever in the groove is a 3rd
There may be others, right now I assume they are down in the noise
By 'cartridge body' do you mean the generator?
As said the crosstalk varies quite some between cartridges, and this is dependent on the accuracy of the parts and assembly of the cart. The Shure V15Vx/SAS I have is between 35-40 dB in separation at 1 kHz.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.