Analog Magik

[Vienna]

Assuming this was a pivot arm, it's likely that the test tracks are in different locations on the 33 vs 45 LPs. A different relative location from the null point is likely to exhibit different crosstalk characteristics. Also the rotational speed difference technically should be compensated for with different anti-skate settings (which obviously nobody does). This will affect observed crosstalk however. Lastly, the LPs themselves could have small differences as you noted. Nevertheless, I'd contend that Analog Majik is doing its job.

Wrong assumption, the unipivot is sold
it’s the gimbal VPI 12” Tonearm. The differences are between two different 33.33 pressings of AM not between the 33.33 and 45 rpm pressings.

 

[bazelio]

Wrong assumption, the unipivot is sold
it’s the gimbal VPI 12” Tonearm. The differences are between two different 33.33 pressings of AM not between the 33.33 and 45 rpm pressings.

That doesn't matter. What I said still applies. My points were in reference to any non tangential arm. (And I said pivot, not unipivot)

 

[Vienna]

I think you are not reading carefully, the numbers you saw are from two different copies of of 33.33 rpm both made by AM where the tracks are exactly on the same position on both

 

[bazelio]

I think you are not reading carefully, the numbers you saw are from two different copies of of 33.33 rpm both made by AM where the tracks are exactly on the same position on both

Yeah, I see that now. Always multitasking while on forums!

So, the pressings are yielding differences. Do you arrive at the same setting for least crosstalk using both records? The absolute crosstalk reading is a bit less important.

 

[Solypsa]

I think you are not reading carefully, the numbers you saw are from two different copies of of 33.33 rpm both made by AM where the tracks are exactly on the same position on both

As you hint at, test records are not easy (impossible perhaps) to produce to absolute tolerances. So we use these to measure changes (improvement) rather than absolute specs. When it comes to vinyl setup I believe in having objective tools but never put your ears in 'sleep mode' because they are your ultimate guide. Not that any in this thread would

 

[Vienna]

Both are giving low values for the AM standards but the

As you hint at, test records are not easy (impossible perhaps) to produce to absolute tolerances. So we use these to measure changes (improvement) rather than absolute specs. When it comes to vinyl setup I believe in having objective tools but never put your ears in 'sleep mode' because they are your ultimate guide. Not that any in this thread would

This is what exactly I had advised earlier in this thread.

 

[leftside]

As with all measurements, all you can do is compare results from one measurement to another.

With AM, and making slight adjustments, I've not only heard better performance, but I've also seen better measurements. The two have coincided for me. I think that's all I can ask for from a piece of software.

AM has given me confidence in setting up my own turntable and cart, as I can see and not only hear the results. With AM, improved measurement = improved sound.

 

[Marcus]

Feickert's measurements are likely incorrect in absolute terms. Cartridge manufacturers will tell you that separation over 30dB is exceptionally rare to see. Yet it seems with Adjust+ to be fairly common. This isn't necessarily important though, so long as Adjust+ provides repeatable measurements that enable you to converge on the point of least crosstalk accurately. I think it does.

The oscilloscope measurement my friend did were very, very similar to Adjust+, so how could both be wrong? I think the opposite is true. Measurements with the Adjust+ were confirmed with the oscilloscope.

On what proof you base your assumption Adjust + is (“likely”) wrong? Differences in measured separation are (IME) dependent on the phono stage in use, the quality of the test record, clamp or no clamp...etc. I got considerably higher separation results with solid state phono stages than with the tube and results were quite different if I use outer and inner clamp...

 

[leftside]

Yes. I also have both and have seen the same sort of difference between the two. Richard (Mak) of AM had indicated to me in an email that AM will show lower channel separation numbers, but that AM is producing a better result if all aspects that are measured via AM are optimized. My Benz LP-S typically shows separation numbers on the order of 8 dB better when done on AP+. As to which one sounds better, I'll have to admit that once I slog through all of the tweaks needed to get through all the alignments in AM it does sound very nice indeed. Maybe a bit more refined and smoother than when done via AP+, but it's not likely a user would be dissatisfied after using either*. Both tools can help, especially if someone needs that sort of quantitative feedback. However, I've seen Harry Weisfeld set up carts in less than 2 minutes with nothing more than a VPI jig and a (amazingly accurate) guess at CW position for VTF that sound absolutely perfect. Some people have the gift ; others like me feel better with a little measured help.

* Although I admit it does bother me that channel separation, which is a fairly simple ratio measurement between signal voltages off each channel, can yield these rather large differences between the two. I have not been disappointed in the listening, however.

Your experiences are very similar to mine. Regarding the low channel separation number reported by AM, I could noticeably hear differences when the L and R had higher measured differences (i.e. 27 and 31), so my primary focus was achieving L and R measurements that were close together (i.e. 29.45 and 29.55) and secondary focus was achieving the highest possible L and R measurements. I like the quantitative feedback provided by AM.

 

[Vienna]

The oscilloscope measurement my friend did were very, very similar to Adjust+, so how could both be wrong? I think the opposite is true. Measurements with the Adjust+ were confirmed with the oscilloscope.

On what proof you base your assumption Adjust + is (“likely”) wrong? Differences in measured separation are (IME) dependent on the phono stage in use, the quality of the test record, clamp or no clamp...etc. I got considerably higher separation results with solid state phono stages than with the tube and results were quite different if I use outer and inner clamp...

all the Lyra carts I have set, but also the Kiseki Purple Heart and AT ART 1000 have all yielded channel separation values over 30 dB and in cases over 31 dB

 

[bazelio]

The oscilloscope measurement my friend did were very, very similar to Adjust+, so how could both be wrong? I think the opposite is true. Measurements with the Adjust+ were confirmed with the oscilloscope.

On what proof you base your assumption Adjust + is (“likely”) wrong? Differences in measured separation are (IME) dependent on the phono stage in use, the quality of the test record, clamp or no clamp...etc. I got considerably higher separation results with solid state phono stages than with the tube and results were quite different if I use outer and inner clamp...

It's not necessarily the "proof" you seek. But cart manufacturers measure channel separation without the use of a phono stage (we can't assume results are independent of phono stage used) by measuring right at the cart pins with a millivolt meter. An acquaintance of mine is a cart tech at Soundsmith who not only builds carts but rebuilds carts from all manufacturers sent in to them. Over the years thousands have passed through. Probably tens of thousands. They tell me they can count on one hand the number of carts they've EVER measured over 30dB using this measurement technique. Likewise, the AnalogMajik docs make reference to 30dB being extremely rare also. I'm not saying the fault belongs to Adjust+. Maybe it does. But one way or another Adjust+ in conjunction with a phono stage, a sound card, and their software algorithms produce raw results that don't correlate with likely direct measurements.

 

[Marcus]

It's not necessarily the "proof" you seek. But cart manufacturers measure channel separation without the use of a phono stage (we can't assume results are independent of phono stage used) by measuring right at the cart pins with a millivolt meter. An acquaintance of mine is a cart tech at Soundsmith who not only builds carts but rebuilds carts from all manufacturers sent in to them. Over the years thousands have passed through. Probably tens of thousands. They tell me they can count on one hand the number of carts they've EVER measured over 30dB using this measurement technique. Likewise, the AnalogMajik docs make reference to 30dB being extremely rare also. I'm not saying the fault belongs to Adjust+. Maybe it does. But one way or another Adjust+ in conjunction with a phono stage, a sound card, and their software algorithms produce raw results that don't correlate with likely direct measurements.

You’re probably right about the highest separation measurements but that was not my point. Again, it’s the diametrically opposite L/R separation measurement difference between Analog Magik and Adjust+. How can AM show that AD+ (and oscilloscope) on four different test LP’s, is totally wrong and vice versa? And BTW, if one Benz LP-S measures 28dB’s and all others between 32 and 34dB then it’s something wrong with the cart not Adjust+.

 

[tima]

Again, it’s the diametrically opposite L/R separation measurement difference between Analog Magik and Adjust+.

Hi Marko - are the measurements so different that one would tell you, for example N degrees clockwise while the other says N degrees counterclockwise? That would be crazy enough that the two manufacturers should get involved. Too bad there's not a way to feed the same data into the two algorithms and compare results.

 

[Marcus]

Hi Marko - are the measurements so different that one would tell you, for example N degrees clockwise while the other says N degrees counterclockwise? That would be crazy enough that the two manufacturers should get involved. Too bad there's not a way to feed the same data into the two algorithms and compare results.

Something like that. Best results with the Adjust+ was (let’s say 32.6/32.8dB) but Analog Magik showed 26.4/39.2dB. When the optimal separation was achieved with the Analog Magic, Adjust+ showed the azimuth was totally wrong. Crazy. Measurements were then repeated with three or four other test records and results were always very similar as described above. As I said before, oscilloscope measurements supported Adjust+.

 

[bazelio]

You’re probably right about the highest separation measurements but that was not my point. Again, it’s the diametrically opposite L/R separation measurement difference between Analog Magik and Adjust+. How can AM show that AD+ (and oscilloscope) on four different test LP’s, is totally wrong and vice versa? And BTW, if one Benz LP-S measures 28dB’s and all others between 32 and 34dB then it’s something wrong with the cart not Adjust+.

Marcus, can you explain the diametrically opposite L/R problem in more detail? I would like to better understand what you're talking about there.

There's nothing wrong with the 28dB cart in all likelihood. In my opinion.

 

[bazelio]

Something like that. Best results with the Adjust+ was (let’s say 32.6/32.8dB) but Analog Magik showed 26.4/39.2dB. When the optimal separation was achieved with the Analog Magic, Adjust+ showed the azimuth was totally wrong. Crazy. Measurements were then repeated with three or four other test records and results were always very similar as described above. As I said before, oscilloscope measurements supported Adjust+.

Nevermind my previous response, since you already provided this info.

Which sound card was being used in this case? Was the sound card calibrated for L/R mismatch per the Adjust+ procedure? (Analog Magik specifically says NOT to do this)

 

[Marcus]

There's nothing wrong with the 28dB cart in all likelihood. In my opinion.

Probably not but I wouldn’t feel comfortable knowing all others easily achieve 33dB.

 

[Marcus]

Which sound card was being used in this case? Was the sound card calibrated for L/R mismatch per the Adjust+ procedure? (Analog Magik specifically says NOT to do this)

Don’t know which sound card was being used but it surely wasn’t calibrated. But your thought about the sound card is interesting and I will investigate this. Maybe here lies the answer...

 

[bazelio]

Don’t know which sound card was being used but it surely wasn’t calibrated. But your thought about the sound card is interesting and I will investigate this. Maybe here lies the answer...

Adjust+ instructions indicate how to calibrate the card, and they suggest the user do so. While Analog Magik instructions are the exact opposite: do not calibrate L/R channels. Analog Magik claims to work with any sound card but that the ART USB card is what they used to develop the s/w and therefore is what they recommend be used with it.

So if you hop back and forth between Adjust+ and Analog Magik without calibrating and de-calibrating the card, then you could be introducing error.

 

[bazelio]

More info regarding channel balance. Analog Magik believes that a loopback calibration (which incidentally is recommended in the Adjust+ instructions) will introduce error:

https://www.analogmagik.com/channelbalance

If there is a channel imbalance within the soundcard (Art Phono Plus), will it skew the data and cause inaccurate results?

Many have asked whether a Loopback cable should be used to measure and adjust channel imbalance. However, this method is flawed.

When using the soundcard to perform actual measurements, only the INPUT is used. But during loopback, both input and output are used, and the results will be magnified by 2x or 100%. Compensation using loopback will actually result in overcompensation, or inaccurate balancing because the Output from the same soundcard is being used.

They go on to say that they have tested many ART USB cards and found none to be out of published tolerance for L/R balance. (Another reason they site for using this particular card and not performing the loopback calibration).