I didn't think calculating exact component values was the problem.
Interesting opportunity for an A/B experiment.
To be fair, a lot of cheap gear also gets RIAA wrong, often very wrong. A good RIAA stage is, IMO, the most difficult thing in the electronics chain to get right and most people (no matter what their market segment) don't.
FWIW, my record experience is the same as yours- I have a collection heavily weighted toward 50s-60s jazz, and all of them are (in theory) RIAA-compliant. Now, also to be fair, the sound on most of them is so colored that the >0.5dB variation of a sloppily designed stage is probably lost in the miasma. Nonetheless, I like to get my RIAA correct (and keep it VERY tight between channels) and color my sound further downstream.
Its you know your collection, you know if you need flexible. Simple as that. Its a speciality. My friend with the Japanese exotic one has it for 2 years and he could not find an LP needing the two extra curves offered (Columbia, Decca) he tells me. If he had rare old, and former Eastern Bloc stuff like Thorsten has by the boatload, only 2 extra curves would be plain misery on the other hand.
Matching both the magnitude and the phase response of the analog filter is a difficult problem. See rec.antiques.radio+phono. I can now generate these filters at essentially any sampling rate desired.
- Greg
My guess, and it is a guess, is that very much care was taken with cutting system electronics. Nor would I want an RIAA eq that was 1 dB out.
On that topic, what's the best way to measure the accuracy of the eq? I usually use an inverse network, but then I have to put my faith in the accuracy of the network components alone. I know they're very good, measurements are my specialty, but that's not the same thing as knowing the response is perfect. It seems like a point by point measurement using a 6 digit ratio transformer would be about the last word in accuracy. Or do you trust your PC card well enough? Obviously an AP system would be nice, but I ain't got one. How do you do it?
On that topic, what's the best way to measure the accuracy of the eq? I usually use an inverse network, but then I have to put my faith in the accuracy of the network components alone. I know they're very good, measurements are my specialty, but that's not the same thing as knowing the response is perfect. It seems like a point by point measurement using a 6 digit ratio transformer would be about the last word in accuracy. Or do you trust your PC card well enough? Obviously an AP system would be nice, but I ain't got one. How do you do it?
It's a reductionist math problem, the solutions exist. There is no magic or tin foil capacitors needed. For me the exotic components for 23 (Robert Anton Wilson would be proud) switched networks is a show stopper. To each his own.
SY does not appear to know or understand the difficulty of measuring the RIAA curve without a pre-correction network. It is difficult and virtually pointless to 'wrestle' with meter ranges and their relative inaccuracies, ever since Dr. Lipshitz gave us the part values to make a quality inverse RIAA.
Is it not the case that the equalisation is applied in cutting the disc? After all, in the 78 era there was no tape and the bass cut would be a characteristic of the reactance/resistance of the coil anyway.
I am probably in a tiny minority in having a high percentage of pre 1954 standard records. Both the USA Columbia's and many companies using the NAB curve really need proper eq ideally;Columbia's sound both shrill and bass heavy with RIAA as examination of the published curve would lead one to expect anyway. Also, early British HMV's, Columbia's and Decca's have no pre-emphasis and applying RIAA makes a marvellous job of dulling the treble. As has been pointed out above,the situation with mainland European records is much more complicated with many companies certainly not falling into line with the 1954 "standard" until considerably later.
John,
Good for you.
However, could it be that you know what the LPs' sound like and you expect them to sound "like that", or have you actually used some form of post EQ or adjustable EQ and found that any adjustment bring no benefit?
I used to use a digital "post RIAA" EQ before and I started researching LP EQ (past the eastern Europe CCIR) when I noticed patterns in the way I set the EQ based on labels for the LP's (especially DG's and Decca's).
First John, many Phono Stages out there, regardless of price, show much larger deviations than my design. I have read no criticism by of any of these.
So do I gather you only disapprove of the AMR PH-77's +/-0.5dB deviation from the RIAA standard but not (for arguments sake) of the -3dB/+2dB of a very expensive and ostensibly "high end" phono by Brand X (I am not mentioning the precise unit as it may be percieved as an attack on a competitors product, it was measured by Stereophile though)
Secondly, why choose +/-0.25dB specifically? It is already too much to be reliably inaudible, why not insist on +/- 0.1dB or even less? I can do a RIAA EQ in production to +/-0.05dB 20Hz-20KHz if I wanted to (I see no point, so I don't actually do it), so why not set this instead?
On the other hand one of the only two attempts to codify "high fidelity" into a standard, namely the West German DIN 45500 (the other was the East German counterpart TGL28660) specified for LP Playback (including cartridge) a maximum tolerance of ±2 dB 63,5Hz ... 8KHz and +/-4dB 40Hz-14KHz to qualify as High Fidelity.
I think we can safely say that under DIN 45500 the AMR PH-77 fully qualifies as "High Fidelity"; as do most if not all of the Phono stages measured by Stereophile to have more than +/- 0.1dB, those having more than +/-0.25dB, even those having more than +/- 0.5dB and even those with more than +/-2dB from 20Hz - 20KHz.
Your insistence on 0.25dB is simply a line arbitrarily drawn in the sand (drawn by you as well) that has no backing in standards or in audibility research. It is not an IHF recommendation, it is not recommended by the IEC, the DIN or VDT.
It is strictly the "John Curl" recommendation. Well, the "Thorsten Loesch" recommendation is +/- 0.5dB simply because greater accuracy is not justified by the tolerances in the actual cutting equipment and LP Pickups.
Criticism on what basis?
There is no official tolerance limit (except DIN45500).
Some manufacturers specify their RIAA EQ accuracy as +/-1dB. What do we say of them?
AMR specifies and adheres to +/- 0.5dB RIAA EQ accuracy. Where is the problem?
Ciao T
Good for you.
However, could it be that you know what the LPs' sound like and you expect them to sound "like that", or have you actually used some form of post EQ or adjustable EQ and found that any adjustment bring no benefit?
I used to use a digital "post RIAA" EQ before and I started researching LP EQ (past the eastern Europe CCIR) when I noticed patterns in the way I set the EQ based on labels for the LP's (especially DG's and Decca's).
First John, many Phono Stages out there, regardless of price, show much larger deviations than my design. I have read no criticism by of any of these.
So do I gather you only disapprove of the AMR PH-77's +/-0.5dB deviation from the RIAA standard but not (for arguments sake) of the -3dB/+2dB of a very expensive and ostensibly "high end" phono by Brand X (I am not mentioning the precise unit as it may be percieved as an attack on a competitors product, it was measured by Stereophile though)
Secondly, why choose +/-0.25dB specifically? It is already too much to be reliably inaudible, why not insist on +/- 0.1dB or even less? I can do a RIAA EQ in production to +/-0.05dB 20Hz-20KHz if I wanted to (I see no point, so I don't actually do it), so why not set this instead?
On the other hand one of the only two attempts to codify "high fidelity" into a standard, namely the West German DIN 45500 (the other was the East German counterpart TGL28660) specified for LP Playback (including cartridge) a maximum tolerance of ±2 dB 63,5Hz ... 8KHz and +/-4dB 40Hz-14KHz to qualify as High Fidelity.
I think we can safely say that under DIN 45500 the AMR PH-77 fully qualifies as "High Fidelity"; as do most if not all of the Phono stages measured by Stereophile to have more than +/- 0.1dB, those having more than +/-0.25dB, even those having more than +/- 0.5dB and even those with more than +/-2dB from 20Hz - 20KHz.
Your insistence on 0.25dB is simply a line arbitrarily drawn in the sand (drawn by you as well) that has no backing in standards or in audibility research. It is not an IHF recommendation, it is not recommended by the IEC, the DIN or VDT.
It is strictly the "John Curl" recommendation. Well, the "Thorsten Loesch" recommendation is +/- 0.5dB simply because greater accuracy is not justified by the tolerances in the actual cutting equipment and LP Pickups.
Criticism on what basis?
There is no official tolerance limit (except DIN45500).
Some manufacturers specify their RIAA EQ accuracy as +/-1dB. What do we say of them?
AMR specifies and adheres to +/- 0.5dB RIAA EQ accuracy. Where is the problem?
Ciao T
Hi,
I did look if I could find any official specs.
I do seem to remember the Neumann lathes electronics to have a spec much worse than +/-0.25dB, maybe +/-1dB 40Hz-14KHz, worse above and below, but my last exposure to these is now well over two decades past.
I heard the Westrex Lathes common in the US had much worse accuracy.
Ciao T
I did look if I could find any official specs.
I do seem to remember the Neumann lathes electronics to have a spec much worse than +/-0.25dB, maybe +/-1dB 40Hz-14KHz, worse above and below, but my last exposure to these is now well over two decades past.
I heard the Westrex Lathes common in the US had much worse accuracy.
Ciao T
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.