Bill, I went back to your first post on the subject to see if I (& other scallywags) were taking this thread off topic.
My interest is to help people use simpler electronics (I think Guru Wurcer's stuff is too complicated but I forgive him 😀) and DSP to get flat vinyl replay response.
I think we are very near being able to do this ... certainly with a 2K EVIL FIR incorporating Guru Wurcer's RIAA as well ... a bit further away with the present zillion bi-quads MiniDSPs in you and George's possession.
I have little interest in stuff above 20kHz and ringing in square waves ... from nearly 2 decades of DBLTs .. though my techniques can be used to correct these too.
Hence I think trying to disassociate electrical resonances and mechanical resonances are only of interest to da pedants (touching my forelock to da true pedants like Lucky 🙂 )
Should I continue on this thread or move my contributions to Digitising Vinyl?
George, i think we are very close to 'choosing' a Test Record. When we decide, we can get a 2K FFT correcting your cartridge and incorporating Guru Wurcer's RIAA too almost immediately.
My interest is to help people use simpler electronics (I think Guru Wurcer's stuff is too complicated but I forgive him 😀) and DSP to get flat vinyl replay response.
I think we are very near being able to do this ... certainly with a 2K EVIL FIR incorporating Guru Wurcer's RIAA as well ... a bit further away with the present zillion bi-quads MiniDSPs in you and George's possession.
I have little interest in stuff above 20kHz and ringing in square waves ... from nearly 2 decades of DBLTs .. though my techniques can be used to correct these too.
Hence I think trying to disassociate electrical resonances and mechanical resonances are only of interest to da pedants (touching my forelock to da true pedants like Lucky 🙂 )
Should I continue on this thread or move my contributions to Digitising Vinyl?
George, i think we are very close to 'choosing' a Test Record. When we decide, we can get a 2K FFT correcting your cartridge and incorporating Guru Wurcer's RIAA too almost immediately.
In Linear Audio Vol 6, I compared A-Weighting against IEC468 for many different types of Riaa preamps (Bipolar and Fet).
And although the basic principles of IEC468 are looking good, the outcome revealed that the IEC468 was all the time 6dB +/-0.3dB above the A-Weighted figure.
Apart from the 6 dB scale factor, +/- 0.3dB is IMHO an insignificant difference.
Probably for non riaa Preamps like microphones IEC468 may give a better view, but apparently not for these annimals, attenuating with 6dB/oct over the whole audiospectrum.
The noise plot as shown in #675, is of course in relative bandwidth, in this case 1/100 of an octave.
Hans
Then it's necessary to understand what is happening with cartridge mechanics, generator magnetics, and interface electronics. Otherwise it's guesswork, and unlikely to be repeatable between cartridges and setups.
There's far more to playback performance than 'flat response'. In any event, it's very probable that many vinyl playback systems have a response that varies significantly with level. Digital correction offers fresh opportunity to correct such non-linearities, if they are understood. Which, AFAIK they aren't......
So I figure it's sub-optimal to take a punt on digital playback eq based on poor understanding of the overall system, or historic beliefs which are, frankly, often baseless and don't stand scrutiny, IME. Assuming accurate, repeatable vinyl playback across systems is the aim, ie achieving its potential in more than one system........?
Well, I disagree. You'll just fall into the same pitfalls as decades before have, in terms of repeatability and linearity. All one is then doing is replacing analog eq with digital, which is fine and can have advantages, but is rather like rearranging deckchairs on the titanic in terms of improving the fundamentals of signal recovery from the record, which is where the big hole is 😉
Nor me. Unless it's indicative of some mechanism that can affect the audioband in other ways, as seems possible here. It's a mistake to think that f response is all there is to it: it's is a non-linear system. In general I can't see the point in reproducing programme material above about 15kHz for me, though I know I can't hear much above 12kHz these days unless it is pretty loud.
Well, it should be interesting and useful to anyone who's into extracting fine performance from vinyl playback, methinks. Or making informed decisions about cartridge choice, for example.
IMO it's far more pragmatically useful discussion, and far less pedantic, than, say, exploring extreme limits of opamp noise performance which are swamped by intrinsic vinyl noise, for example? Or discussing digitising algorithms which introduce distortions already swamped by intrinsic non-linear distortions in vinyl playback........ apparently? And which, you might have noticed, I don't participate in 🙄
LD
Last edited:
The noise plot in #675 looks like it has A weighting applied. The most useful graphical display is the raw noise without weighting.
This is what I show in #662 .. which is plot off a B&K 1/3 8ve filter set.
I can assure you that the audible difference in noise is more than the straight numbers show. 'Whitish' noise as with most MM/MI systems is always more irritating as it is obviously 'artificial'.
___________________
Lucky, you've pooh poohed the efficiency advantage of MC cartridges for S/N. But why then are your spruiking Aurak type virtual earth inputs?
It seems to me with virtual earth you discard
- manufacturer's recommendations on capacitance on the usual 47k input
- for a dubious method of guessing the somewhat wonky L & R of the cartridge requiring fudge factors for flat response
You do make the wiring 'immune' to total capacitance but ..
It's far easier for the punter to measure his total capacitance from cartridge pins to preamp than do to Julian's curve fitting for the wonky L & R on his cartridge. THEN he has to compensate for the EQ that the manufacturer's recommended C + wonky L & R + 47k provides (which of course compensates for the evil mechanical resonances discussed in this thread). Then he has to adjust Aurak type EQ to suit.
George's measurements show the recommended procedure to do the above (if any) don't quite get the response to what the manufacturer intended.
BTW, I like the idea of virtual earth input. I've been trying for years to make one that is quieter than my enormously complicated 1981 battery circuit. I think the advantage is on the noise front though MM/MI still can't beat the MC with optimised preamp.
But today, I would use DSP to do the twiddly bits. 😀
I'm not sure there is anything OT in this thread as it meanders around. It would be good to keep DSP techniques in the other thread, but anything around what goes on in cartridges is good for here. But that is just my view. I am happy as long as there is robust discussion on all the things we have forgotten over the last 40 years and a few accepted facts shot down 🙂
Edit as Richard posted as I was completing. Measuring capacitance of a cable is beyond the average Joe. At least there are some carts for which relatively accurate impedance measurements are available and as you are only changing R to correct it is possible to adjust in circuit. For me this makes a slightly more applicable tweaky edge case on the understanding that few will embrace these methods.
Noise wise Hans as shown that transimpedance gives noise performance below best vinyl noise floor so I reckon we are into good enough territory and we can focus on other things. There is much yet to explore
Edit as Richard posted as I was completing. Measuring capacitance of a cable is beyond the average Joe. At least there are some carts for which relatively accurate impedance measurements are available and as you are only changing R to correct it is possible to adjust in circuit. For me this makes a slightly more applicable tweaky edge case on the understanding that few will embrace these methods.
Noise wise Hans as shown that transimpedance gives noise performance below best vinyl noise floor so I reckon we are into good enough territory and we can focus on other things. There is much yet to explore
Last edited:
Shall we just get the response flat to start with? We are very close to doing this simply.
When you decide what non-linear stuff you want done, there are enough DSP gurus (and some pseudo ones 😀 ) on this thread who might be able to help you.
My previous-life expertise was using electronics, linear & non-linear and DSP, linear & non-linear, for better speakers. If you think vinyl playback isn't linear time-invariant, you should look at speakers. It's just often convenient for us to pretend that they are.
And if you want to know what better for speakers is in my book .. its stuff that will improve its performance in DBLTs ... which is my other previous-life expertise. And BTW, 'flat response' usually trumps THD for transducers in DBLTs ... though this is one of dem sweeping generalisations 🙂
Last edited:
He uses one of these
Pocket Digital Multimeter DMM Frequency Capacitance Tester Auto Range Tester GH | eBay
probably cheaper than his OFC cable. There's a deluxe version at $19.11 🙂
Waddya mean accepted facts shot down?? All dem Vinyl Resonance Deniers and supa LoNoise Deniers. They should stand for POTUS 😀
yeah, and then what about the input capacitance of the phono stage? Very few manufacturers actually publish that, just the additional C they have added as they can't be bothered. It's not easy.
As for accepted facts, too bunged up with flu to spar with you today. Maybe when I feel better 🙂
As for accepted facts, too bunged up with flu to spar with you today. Maybe when I feel better 🙂
As already said, it's not about intrinsic cartridge noise. Though Aurak topology preamp noise performance is very creditable, there are many other ways of achieving perfectly adequate noise performance for all cart types. Hard to imagine, but preamp noise really should be a non issue, IME, though MC can be hard work 😉
There are potential environment sourced noise rejection advantages to Aurak, due to low impedance presented to cartridge and wiring. That is system specific, and hard to quantify though. And cable parameters, including length, become a non-issue, of course.
You're not getting it, Richard.
Manufacturers' recommendation for C/47k is based on conventional LCR resonance, and post resonance roll-off. Aurak eliminates this, and its artefacts, so of course if the manufacturer was relying on this resonance, the native audioband f response of the cartridge may not be flat.
So, with Aurak, one should select cartridges which have mechanical resonance above the audioband, and a generator with a flat native audioband response. George didn't do this: the cart he used to test was chosen for other random criteria by Bill, perhaps with the intent to see the mechanical resonance.
It won't, and doesn't, work well with all MM/MI carts IME, for reasons beyond f response too.
The matter of a 'correct' electrical model for cartridge output impedance is a work in progress, as per a few pages ago. I have preliminary thoughts and simulations which I'll share, nearly there, for George's cart. That's the nature of exploration of such things, who knows where it might lead?
But, with respect, you can't know properly what to twiddle, Richard.......🙄
LD
Last edited:
Lowest range on that is 4nF. Given it will be at best +/-5% thats 200pF of error band, or more than the capacitance you are trying to measure.
Yup, and that is very probably where the real opportunity for digital correction might lie IMO.
LD
Off-topic:
I've been probing my picoFarads using this handy meter which is available in kit form: LC METER PROJECT AND KIT - L/C Meter IIB
The L metering is really only good for Hams though.
I have projected in green the raw unweighted noise for a 3nV/rtHz / 0.3pA/rtHz amp with a 600 mH / 800 Ohm cart into your B&K plot.
Vertical and horizontal scale are exactly adjusted, just as the fact that 50Hz hum on the B&K plot was displayed at 40Hz.
And I have added for the same configuration the A-weighted plot, this time also on a logarithmic Y axis.
Which of the two tells me more ? Of course the A-weighted version, without a doubt.
Hans
Is that done with SY's spreadsheet?
It doesn't correspond with real life. Your green line is what I'd expect of a MC cartridge with a SOTA noise preamp. What preamp is assumed?
I'll have to get someone to translate his XLSX spreadsheet into something I can read. 🙂
In the accompanying text I have mentioned 600mH and 800 Ohm Cart, just as in your B&K recording.
And the green line came from LTSpice, calculated with 1/100 steps of an octave, just as in the A-weighted plot below.
Nothing SOTA about the Amp, just an ordinary NE5534.
Or was your B&K noise recording from an LP instead of only showing the noise from Cart + Preamp ?
In that case we are comparing apples to oranges and your B&K plot was not fitting in the discussion we had.
Hans
What termination network? This all becomes moot anyway when you plot the quietest blank groove you can find.
In #694 I plotted linear spectra instead of 1/3 octave bands.
So the B&K plot and my plot are incomparable.
I will redo a new version, now with 1/3 octave bands.
Hans
So the B&K plot and my plot are incomparable.
I will redo a new version, now with 1/3 octave bands.
Hans
I don't get the message. My plot had nothing to do with blank grooves.
It was the noise spectrum of preamp with cartridge connected.
That was what the discussion was about in #675.
And then the B&K plot popped up, where it is unclear to me whether this was recorded while playing a blank groove of a record, or just displayed the noise from preamp with cart connected.
Hans
Here are the noise spectra for a 5534 riaa preamp, but this time not linear but with 1/3 octave filters.
Both the A-weighted and the non weighted spectra are displayed.
This is just the noise from preamp + MM Cart connected, terminated with 47k and 125pF. Cart is 600mH / 800 Ohm.
The A-weighted plot should conform to the noise spectrum we actually hear.
Hans
Both the A-weighted and the non weighted spectra are displayed.
This is just the noise from preamp + MM Cart connected, terminated with 47k and 125pF. Cart is 600mH / 800 Ohm.
The A-weighted plot should conform to the noise spectrum we actually hear.
Hans