Phono Cartridge Response Measurement Script

[JP]

It always has been.

 

[JP]

Thanks but I need start of left sweep to start of right pilot, not start of pilot to start of pilot.

@Balle Clorin ? Or the file.

 

[Balle Clorin]

Thanks but I need start of left sweep to start of right pilot, not start of pilot to start of pilot.

The seconds are shown in the plot.

 

[JP]

That's fine - I'll just remove support for the record until/unless I'm able to get thorough timing info for it.

 

[JP]

Some cool progress. Unified script that does "standard" plots and "compare" plots.

Nothing new about "standard" except the config file syntax.

{
  "plot_info": "Standard",
  "equip_info": "Arm -> Phonostage -> ADC",
  "author": "Plot Author",
 
  "files": [
    {"path": "VM745xML-STR100_30k-105pF.wav", "label": ""}
  ],
 
  "mode": "standard",
  "style": 4,
  "dimensions": ["fundamental", "crosstalk", "h2", "h3"], 
  "extract_sweeps": true,
  "extract_channel": "both",
  "test_record": "STR100",
 
  "normalize": 1000,
  "normalize_mode": "relative",
 
  "start_f": null,
  "end_f": 20000,
  "smoothing": 0.5,
  "override_ylim": null,
 
  "riaa_mode": 0,
  "riaa_inverse": false,
  "str100": true,
  "xg7001": false,
 
  "plot_data_out": false,
  "log_level": "info"
}

"Compare" is using full sweep recordings, extracting the left channel sweeps from each, plotting only the fundamental, and showing 1-20kHz using a more aggressive data smoothing filter setting (the one I usually use for loading comparisons).

{
  "plot_info": "Compare",
  "equip_info": "Arm -> Phonostage -> ADC",
  "author": "Plot Author",
 
  "files": [
    {"path": "VM745xML-STR100_30k-105pF.wav", "label": "30k"},
    {"path": "VM745xML-STR100_33k-105pF.wav", "label": "33k"},
    {"path": "VM745xML-STR100_38k-105pF.wav", "label": "38k"},
    {"path": "VM745xML-STR100_40k-105pF.wav", "label": "40k"}
   ],
 
  "mode": "compare",
  "style": 5,
  "dimensions": ["fundamental"],
 
  "extract_sweeps": true,
  "extract_channel": "left",
  "test_record": "STR100",
 
  "normalize": 1000,
  "normalize_mode": "absolute",
 
  "start_f": 1000,
  "end_f": 20000,
  "smoothing": 0.02,
  "override_ylim": null,
 
  "riaa_mode": 0,
  "riaa_inverse": false,
  "str100": true,
  "xg7001": false,
 
  "plot_data_out": false,
  "log_level": "info"
}

Nevermind the legend placement. Will make it pretty later.

 

[JP]

GitHub - FidelisAnalog/SJPlot at v19

Phono Cartridge Frequency Response Plotting Script - GitHub - FidelisAnalog/SJPlot at v19

github.com

Rough alpha code. Not for production plots.

 

[JP]

@ElectricWarrior thanks for the referral!

 

[JP]

In other news, pending separate verification, but looks like some good progress on higher sample rate RIAA biquads:

 

[USER]

Oh cool. Reminds me of when I tried to test for the accuracy of the filters. Mine are certainly not accurate but these do show the scale and how it would "affect" the results.

 

[JP]

If you're free tomorrow I have 24 filter coefficients that need to be verified

 

[ElectricWarrior]

@ElectricWarrior thanks for the referral!

Thank you for providing this. It's very useful. And useful stuff gets referred to.

 

[ray_parkhurst]

I'm testing my P-Mount cartridges and am seeing some of them, mostly the MCs, where the script seems to reference the plots to the L channel crosstalk rather than main signal, as seen below for the AT312HEP. What am I doing wrong?

 

[mackat]

I'm testing my P-Mount cartridges and am seeing some of them, mostly the MCs, where the script seems to reference the plots to the L channel crosstalk rather than main signal, as seen below for the AT312HEP. What am I doing wrong?

View attachment 499401

I have had this happen many times, but with older versions of the script that required manually trimming the audio.

I seem to recall that it was caused by too much silence after the right channel sweep, but I can’t say with certainty.

Not sure why that would happen with the current version, though.

 

[ray_parkhurst]

I have had this happen many times, but with older versions of the script that required manually trimming the audio.

I seem to recall that it was caused by too much silence after the right channel sweep, but I can’t say with certainty.

Not sure why that would happen with the current version, though.

I've been manually trimming excess at beginning and end.

Here is a link to the file:

Dropbox

www.dropbox.com

 

[JP]

I've been manually trimming excess at beginning and end.

Why?

Here is a link to the file:

You've a pathological amount of HF noise in your recording. If I run a steep LPF @30kHz on the file it plots fine, or if I remove the clipping at the sweep start (your levels are too hot - think about what happens when inverse RIAA is applied) it also works fine.

 

[ray_parkhurst]

You've a pathological amount of HF noise in your recording. If I run a steep LPF @30kHz on the file it plots fine, or if I remove the clipping at the sweep start (your levels are too hot - think about what happens when inverse RIAA is applied) it also works fine.

I tried lowering the levels by 12dB and still have same issue. Tried all different modes...MCPrepre-MM, MCPrepre-Line, MM direct, and nothing worked. Seems it might be associated with the Soundblaster box.

 

[narud]

I'm testing my P-Mount cartridges and am seeing some of them, mostly the MCs, where the script seems to reference the plots to the L channel crosstalk rather than main signal, as seen below for the AT312HEP. What am I doing wrong?

View attachment 499401

ive gotten results like this when i try to run the sweep as a single file sometimes it works, sometimes i get stuff like this. i stopped trying to figure out what i was doing or what was going wrong and have continued to trim the files and run them independently. that's worked 100% of the time.

 

[ray_parkhurst]

I assume you mean L and R independently? I tried that but got errors.

 

[Balle Clorin]

No just trim the ends, and run script with both files left and right , no auto extraction. That usually works for me. If my recording level is low or crosstalk too good and end not clear, The script seem to become confused

 

[ray_parkhurst]

Hmm, what do you mean by "both files"? In the "advanced" section I see an option for a second file but it's not clear what will be done with it. I tried using trimmed L and R there but it gave a garbage output. I don't see any option to turn off auto extraction.