Phono Cartridge Response Measurement Script

[JP]

Sure. Lemme think through how to confirm it. I think it'd also require some extra filtering as 2x bass boost can easily push LF junk up beyond the fundamental.

 

[Thomas_A]

Sure. Lemme think through how to confirm it. I think it'd also require some extra filtering as 2x bass boost can easily push LF junk up beyond the fundamental.

Yes, I usually do an HP around 400 Hz on these sweeps, to clean it up.

 

[JP]

Actually a good excuse to figure out Scott's EQ code, as I also want super accurate coefficients for 192kHz. I have them for full EQ, but not split like we need. Last time I tried I didn't get far.

 

[Thomas_A]

Actually a good excuse to figure out Scott's EQ code, as I also want super accurate coefficients for 192kHz. I have them for full EQ, but not split like we need. Last time I tried I didn't get far.

No need IMO, but ok if possible. Next is to understand why Ortofon and CA-TRS1007 can display such different crosstalk, which I do not really have a definitive answer for. VTA/SRA?

 

[JP]

Dunno. I haven't played much with either record. EDIT: while you're at it @Thomas_A see if you can find anything definitive on EQ for TRS-1005. The response looks much better without it.

 

[Thomas_A]

Dunno. I haven't played much with either record.

It’s a pity LD or Yosh is not around here. Perhaps @Balle Clorin can help with some geometry and maths in the fun thread. I don’t really know where to start.

 

[JP]

Indeed. I haven't seen LD around anywhere for several years now. I thought I saw Yosh post somewhere not too long ago, but he hasn't been that active for a long time.

 

[Thomas_A]

Dunno. I haven't played much with either record. EDIT: while you're at it @Thomas_A see if you can find anything definitive on EQ for TRS-1005. The response looks much better without it.

View attachment 491546

View attachment 491547

I do not have TRS1005 or any of the JVC records, so really no idea about it. No EQ above looks better, but which cartridge is it? I would guess an AT.

 

[JP]

Pioneer PC-1000II.

It's more of a research project than an experimental one. Most test records are cut with bass roll-off and shelf because 1) on most lathes you have to modify the electronics to disable it and 2) if there's content there you have to have it*. On TRS-1005 they don't mention anything about the recording characteristic at all. The only clue is that they say the sweep tracks are meant to be recorded with a BK 2305, but don't mention anything about using the 4416.

Even more fun, I've seen plots done by cart manufacturers where some appear to not have EQ applied, and some do.

*Per the Ortofon/Flo record and Flo's comments, it'd seem that modification is required for a half-speed cut where treble emphasis was not applied. I may need to dust off the Neumann schematics.

 

[JP]

My pal Claude and I refactored the sweep extraction function to use Hilbert Transform. The punchline is a bit more resilient on detection and around a 90% performance improvement for that process. Total processing time for STR 100 with sweep extraction on my machine is down to just a hair less than three seconds from around six seconds.

This is in SJPlot 18.5.0 or higher, which is already live for sjplot.com/online. The end of the first pilot tone must occur within the first 30 seconds of the file.

 

[Thomas_A]

Sounds good. Still need to fix the script on my new M5 MacBook. If only Santa could give me some spare time.

 

[JP]

MBP? I just picked up an M4 Air as they were on sale here. Was getting tired noise, heat, and thermal throttling on the old i7 I used in the music room. For light day-to-day workloads it's faster than my M1 Max MBP.

 

[Thomas_A]

MBP? I just picked up an M4 Air as they were on sale here. Was getting tired noise, heat, and thermal throttling on the old i7 I used in the music room. For light day-to-day workloads it's faster than my M1 Max MBP.

Yes, MBP M5. Errors with the numpy versions, no time to fix it yet.

 

[ray_parkhurst]

I have been unable to analyze the STR100 data for the last couple days. I get this error:

attempt to get argmax of an empty sequence

 

[JP]

I have been unable to analyze the STR100 data for the last couple days. I get this error:

attempt to get argmax of an empty sequence

And you get this error after the update I pushed 50 minutes ago?

 

[ray_parkhurst]

Yes, your post prompted me to try again, but no joy.

 

[JP]

In java console what's the SJPlot version being reported?

 

[ray_parkhurst]

Hmm, I forgot to refresh to the new version. Works.

 

[JP]

Yes, MBP M5. Errors with the numpy versions, no time to fix it yet.

Odd - never had an error with numpy. Only library issues I had was when the script was using librosa. Glad to have been able to dump that.

 

[JP]

@Balle Clorin when you have a moment can you let me know the duration between the start of the left sweep to the start of the right pilot tone for DIN 45543? I can't find my layback.