Phono Cartridge Response Measurement Script

[JP]

3.5.2

It only affects plot 2 for me.

View attachment 272385View attachment 272386

Around line 463, replace

#fig.legend(loc=3, bbox_to_anchor=(0.899, 0.11))
fig.legend(loc=3, bbox_to_anchor=(0.125, 0.11))

with

lines1, labels1 = ax1.get_legend_handles_labels()
lines2, labels2 = ax2.get_legend_handles_labels()
ax1.legend(lines1 + lines2, labels1 + labels2, loc=3)

The latter lets matplotlib auto-locate the legend at the bottom left. I've updated the code in the original post.

 

[Thomas_A]

Test records are all over the place. My copy of the CA-TRS1007 suffers not only by poor x-talk, but has some weird noise in the left channel, which shows as wavy pattern for the L signal channel including distortion and an even more wavy pattern for the left silent channel (x-talk).

 

[JP]

File?

 

[Thomas_A]

File?

Only 44.1 kHz, need to record again, but here it is.

TRS1007 Clearaudio sweeps.wav

Shared with Dropbox

www.dropbox.com

 

[JP]

Only 44.1 kHz, need to record again, but here it is.

TRS1007 Clearaudio sweeps.wav

Shared with Dropbox

www.dropbox.com

Thanks will check tomorrow. Did you check record centering?

 

[Thomas_A]

Thanks will check tomorrow. Did you check record centering?

Yes, I believe so.

 

[USER]

Around line 463, replace

#fig.legend(loc=3, bbox_to_anchor=(0.899, 0.11))
fig.legend(loc=3, bbox_to_anchor=(0.125, 0.11))

with

lines1, labels1 = ax1.get_legend_handles_labels()
lines2, labels2 = ax2.get_legend_handles_labels()
ax1.legend(lines1 + lines2, labels1 + labels2, loc=3)

The latter lets matplotlib auto-locate the legend at the bottom left. I've updated the code in the original post.

Perfect. Thank you!

These were quick measures so crosstalk can be improved, but the former would never be helpful because of the test record. Each has its problems. I'll use style 3 for my CBS measurements for now.

I'm hoping to have a first go at a CBS to Clearaudio (hopefully JVC) EQ curve this weekend. They seem comparable to what you get difference wise. It'll be a start, anyways. I'll use it to compare future measurements and see how it does.

 

[Thomas_A]

With the sweeps I get the following approximate crosstalk figures at 1 kHz:

CA-TRS1007 (new): -22/-25 dB
Denon XG7001 (mint): -28/-41 dB
QR2010 (not so good condition): -26/-34 dB

Same directions, perhaps I need to check with the 1 kHz Ortofon reference signal and adjust.

 

[JP]

7001 on my setup favors the left channel. Haven’t dug in to it yet.

 

[Thomas_A]

7001 on my setup favors the left channel. Haven’t dug in to it yet.

My results for the Ortofon test record and 1 kHz tracks, things look good.

Raw RMS:
L->R 1: -33.4 dB
R->L 1: -32.4 dB
L->R 2: -32.6 dB
R->L 2: -32.9 dB


Filtered HP 100 Hz RMS:
L->R 1: -40.8 dB
R->L 1: -35.1 dB
L->R 2: -35.4 dB
R->L 2: -35.9 dB

 

[JP]

Thanks will check tomorrow. Did you check record centering?

Left channel of tracks 1 and 4 are modulated by a signal that cycles every 1.8s - double modulation, if you will. Is the record warped?

 

[Thomas_A]

Left channel of tracks 1 and 4 are modulated by a signal that cycles every 1.8s - double modulation, if you will. Is the record warped?

No, the record is quite flat. I bought it new and it was like that from start. The wavy pattern is the same in left channel whether it contains the signal or not. Crosstalk of the 1 kHz Ortofon test record is way lower.

 

[JP]

Highlighted section is 1.8s. I see it on @USER CA sweeps too but not nearly as pronounced. I supposedly have a few copies of this record on the way but don't currently have one here.

 

[JP]

Interesting - I can't filter it.

sos = signal.iirfilter(17,[10, 25000], rs=60, btype='band',
                            analog=False, ftype='cheby2', fs=Fs,
                            output='sos')
audio = signal.sosfiltfilt(sos,audio)

 

[JP]

It doesn't show up in track 3, but track 3 does show a ~1.25dB channel imbalance...

 

[JP]

Ya it's not modulated, it's an amplitude variation that repeats once per revolution. If you can duplicate it then I'd do a layback at 45 to see if the issue is in the grooves or not.

 

[Thomas_A]

It doesn't show up in track 3, but track 3 does show a ~1.25dB channel imbalance...

View attachment 272539

Yes it may be this stylus which has a .7-.8 dB imbalance measured by other means. I usually add .5-1 dB when playing. My previous stylus had around .3. That said I have an issue with my Muffsy DIY phono pre that I need to fix - the left ch drops out sometimes and there some poor contact that seems to be chassie-related. Don’t know ir that might be the problem, it is more an on-off effect.

It is however only the CA-TRS1007 record that shows the modulation in L channel. No other record.

 

[JP]

I'd do a layback at 33 and 45. If the pattern changes from 1.8s to 1.33s, then the source of the issue is the record, for whatever reason.

Silver lining - I've added an HP pre-filter to the script.

EDIT - similar cyclic patterns are prevalent in the distortion axis approx 400-900Hz and +2kHz typically. Have been curious about that for some time.

 

[Thomas_A]

I’ll do some testing later on.

 

[JP]

Post number three will house the latest test code along with release notes and to-do/known issues list.