[dlaloum]

I did a series of tests generating square waves then passing them through various low pass and high pass filters...

I varied the filter types, so some were digital phase neutral, and others were standard phase changing types...

The square wave output clearly showed the difference between the two filters - even though they were filtering at the same frequencies and same or very similar slope...

I posted the graphs at https://sites.google.com/site/zevaudio/turt/square-waves-and-phase/minimum-vs-linear-phase-sq

Phase issues become quite visible on a square wave trace!

bye for now

David

[neobop]

Thanks David,
Fascinating stuff, but I have a question.  You say, "So as an assistance in working out what and how square waves respond to variations in frequency and phase response, I have taken  images of oscilloscope readouts of a 1kHz digitally generated square wave, adjusted with various "phase varying" filters which reflect the typical aberrations found on turntable cartridges"

How do you know that the filters are simulating cart aberrations, and not just reflecting the affects of the filters?  There's very little information (that I know of) relating to phase vs amplitude performance in phono carts. The Ortofon paper I've been referring to, is limited to amplitude vs phase at a fixed level, I believe. [Since I brought back this thread, I haven't had a chance to review it, but I think that is correct.] Are you saying that phase aberrations are easily observed when the cart is modulated at one frequency and not under amplitude plotting conditions?  If so, does this simulate actual performance?

In the old days, square wave response for carts, was sometimes published along with test reports, in Stereo Review and possibly High Fidelity.  Have you seen any of these, and do you have a test record with square waves? Forgive me if you've already posted this, memory fails me.
neo

[dlaloum]

Hi Neo,

yes I have a couple of square wave test records...
for the rest - the posted charts were artificial in that the waveform was generated digitally and then filtered digitally - I used the Ozone software for filtering as it allows me both types of filters simply and easily... as to the behaviour of the filters, I have to trust the software vendor!

What I was doing at the time was laying the groundwork, for seeking out phase anomalies in phono cartridges... ie: my collection of examples allows me (and I hope anyone interested enough to read the site) -  to identify from a square wave recording of a phono cartridge - whether there are phase anomalies or not.

Of course this would not identify the cause of the phase anomalies.... and some of those could potentially be recorded within the test tracks themselves (!).
Also the cartridge loading circuit (inductance, capacitance, resistance) also has phase impacts... which is why a potentially "ideal" cartridge should have as low an inductance and capacitance as possible.

The RIAA filter used needs to simulate the analogue circuit, as it has to correct the phase anomalies introduced during the RIAA encoding/recording - but potentially and theoretically, it might be possible to thereafter correct the frequency response of the cartridge not using cartridge loading as per standard practice, but using digital filters that do no mess with phase....

I know that this might be considered sacrilege by the analogue purists - but I believe that digital technology is at the level now where good digital equipment is on a par with some of the best analogue kit, and this method has the potential of achieving close to flat phase and amplitude response simultaneously.

My available time for vinyl experimentation however became severely constrained - and I have not taken this project any further (yet).


Based on my square wave charts, you can see from a reproduced square wave test track, how close the reproduction is to the original intent - and as you adjust the loading and/or filtering, you can also in theory identify whether the end result is in fact closer to the ideal... it seems a valid approach - albeit time consuming

bye for now

David

[neobop]

The RIAA filter used needs to simulate the analogue circuit, as it has to correct the phase anomalies introduced during the RIAA encoding/recording - but potentially and theoretically, it might be possible to thereafter correct the frequency response of the cartridge not using cartridge loading as per standard practice, but using digital filters that do no mess with phase....
I know that this might be considered sacrilege by the analogue purists - but I believe that digital technology is at the level now where good digital equipment is on a par with some of the best analogue kit, and this method has the potential of achieving close to flat phase and amplitude response simultaneously.

My available time for vinyl experimentation however became severely constrained - and I have not taken this project any further (yet).


Based on my square wave charts, you can see from a reproduced square wave test track, how close the reproduction is to the original intent - and as you adjust the loading and/or filtering, you can also in theory identify whether the end result is in fact closer to the ideal... it seems a valid approach - albeit time consuming

David,
I have to wonder about verification of simulated sq wave validity.  I assume you can't plug your cart output directly into the PC, so you would have to use a gain stage?  Maybe you can it's unclear to me if you're talking about theoretical performance or actual performance using digital filters.  I know you've corrected amplitude response with digital filters, but I thought it was after going through your phono pre?

I'm not concerned with analogue purity with this, in fact I think it might have the potential for a better phono stage. Boost the voltage to a usable level w/o RIAA and do everything digitally.  High quality ICs are faster and cleaner than discreet components and it could probably be made with capability to dial in your preferred sound, not just in general, but for each record.  Is anyone doing this commercially?

Maybe it would be a good time to get back in the audio game.  Well, I don't know about that but it seems like it could be killer good.
neo

[dlaloum]

I have not taken it the next step

I first had a good theoretical think about it, then did the theoretical exercise of generating square waves and filtering them (all internally within the PC) - providing a reference point as to what to expect...

My current setup had the TT connected to a JLTI preamp, which I had the maker (Joe Rassmussen) modify with an RIAA bypass switch - so I can get straight flat output from it...

Of course the gain stage itself still has an impact... total capacitive loading in this setup without additional C loading is circa 60pf (with some effort I could potentially get it down another 10 or so pf, but I am pretty close to what is possible already!)

I agree that this has the potential for being a "better mousetrap" - but it remains a very technical endeavour due to variances between differing cartridges, styli, and loadings (cable etc...)

So a "turnkey" solution would require a variable gain stage/DAC/Filter, a test track LP, and some software (or perhaps firmware) to do the necessary analysis and then configure the filters correctly.

Weaknesses as things stand today - for best results you have to be careful with the gain and signal/noise - too high and you clip (really nasty in digital!) - too low and you drop into the noise floor.
22 bit resolution seems to still be current state of the art, which provides a S/N of circa 104db - which is just fine when properly calibrated/setup - but if you have a random cartridge, with unknown output, etc...
(mind you, it should be possible to use the same test track LP to "auto set" levels....)

I have toyed with the ideal of getting back into the audio business too - but this might be just too niche.... (and the development costs would never be recovered as a result)

And yes I agree it has enormous potential...

bye for now

David

[neobop]

Something came up on Karma on the hunt for Precept PC440/550 info.  It has to do with absolute phase, but that's not we were really discussing. Specific information is on post #122 and the question posed about DC resistance relationship to impedance, is on the previous page.
http://www.audiokarma.org/forums/showthread.php?t=508578&page=9

"Actually, the DCR is most often the higher of the two figures, very roughly by a factor of two. Look at Stanton, ADC, Empire, Shure (where one of the few exceptions is the V15 II/M91E)."

DC resistance as measured by a meter at a particular frequency is a straight up measurement. With AC, inductance and capacitance (of cart) is factored in, and impedance (@1K for carts) is calculated.  I've read that impedance also determines phase angle, so this should impact absolute phase and might be part of the reason for different house sound impressions.  I've also noticed that the "best" AT carts have impedance figure close to their DC resistance and don't seem to be aggressive in the highs like some others.
neo

[neobop]

Something came up on Karma on the hunt for Precept PC440/550 info.  It has to do with absolute phase, but that's not we were really discussing. Specific information is on post #122 and the question posed about DC resistance relationship to impedance, is on the previous page.
http://www.audiokarma.org/forums/showthread.php?t=508578&page=9

"Actually, the DCR is most often the higher of the two figures, very roughly by a factor of two. Look at Stanton, ADC, Empire, Shure (where one of the few exceptions is the V15 II/M91E)."

DC resistance as measured by a meter at a particular frequency is a straight up measurement. With AC, inductance and capacitance (of cart) is factored in, and impedance (@1K for carts) is calculated.  I've read that impedance also determines phase angle, so this should impact absolute phase and might be part of the reason for different house sound impressions.  I've also noticed that the "best" AT carts have impedance figure close to their DC resistance and don't seem to be aggressive in the highs like some others.
neo

Maybe this needs some explanation, although I don't fully understand it myself.
Absolute phase refers to phase like if you have both your speakers hooked up + to - on both speakers.  Assuming your electronics aren't inverting the phase, then your speakers are in phase with each other, but absolute phase is inverted.

AT carts always AFAIK, have DC resistance lower value than impedance.  This is an explanation of DC vs impedance I looked up:
 
I think impedance is calculated, not measured.
Standard cart spec for impedance is at 1KHz and such is the case for AT carts. As I understand it, an AC circuit has reactive aspects (capacitance/inductance) and phase is taken into account along with amount of resistance. A DC circuit is purely resistive - no impedance necessary.
Impedance is defined as the frequency domain ratio of the voltage to current. That is the voltage–current ratio for a single complex exponential, at one frequency. The current gets ahead of the voltage and causes a phase shift, defined or plotted by phase angle.

That's the best I can do, I'm not an EE. I would think inductance is the obvious suspect for the difference between DCR and impedance.

The AT-440 DCR is around 770 ohms and impedance is 3.2Kohms.  Inductance is 490mH.  Some other examples of carts named above, have considerably more inductance and impedance is less than DCR?

If anyone can explain what's happening or verify the above info on Stanton, Shure, ADC and Empire, your input is welcome. 
neo

[dlaloum]

Don't have the tools at the moment to correctly measure impedance (as opposed to resistance...)

[neobop]

Can resistance be a higher value than impedance?   Seems to me, when factoring in reactance, impedance should always be the higher value.

Maybe measuring parameters at different frequencies will mess up results? 
I'm just trying to make sense of the above statement, that the other carts, named above, have DCR > impedance.  Maybe he meant something else, but it still doesn't make sense to me.
neo

[dlaloum]

Need to read up on electrical theory... impedance does vary with frequency, and with coil design, but I would have thought that ultimately minimum value would be DC resistance and frequency variation would increase that by a variable amount....

[neobop]

Think I've found the answer, at least to the basic question relating to phase.

http://hyperphysics.phy-astr.gsu.edu/hbase/electric/imped.html

From the bottom of the page:
"The units for all quantities are ohms. A negative phase angle implies that the impedance is capacitive, and a positive phase angle implies net inductive behavior."

I don't see how the impedance of a phono cart could be capacitive, which leads back to the conclusion by Carlosfm, on VE.  The overwhelming determinate of amplitude response is mechanical.

http://www.vinylengine.com/turntable_forum/viewtopic.php?f=19&t=6674

neo

 

[dlaloum]

Posted this on Karma and VE - but it is relevant to this discussion too so here it is as well!

A couple of things have caught my attention over time...

1) differences between voicing of cartridges (and systems) seem primarily down to variations in frequency response - but may also involve phase effects.
(Note I am not talking about things like tracking ability, dynamics, etc... but voicing)

2) variations in F/R driven by various aspects like cantilever resonance, and cartridge loading, also have an impact on phase - these phenomena appear to be minimum phase... this means that if they are corrected using analogue filters, the phase is corrected at the same time... if they are corrected using digital linear phase filters - the F/R is adjusted, but not the phase.

3) Reading on Ozone5 software shows that the "spectrum matching" feature can be used with either analogue or digital filters... - the spectrum matching allows one to "record" a specified spectrum, compare it to a second recording and automate the required EQ curve to make the two spectrums "match" (or at least approach each other!)

It should theoretically be possible to use a pink noise test track, and the spectrum matching feature to correct a cartridges output for flat frequency and phase response using Ozone5.

One could artificially generate a "perfect" pink noise spectrum - use this as the target spectrum - then record the pink noise test track, and have the software create a custom EQ using minimum phase filters to provide a perfectly neutral end result in both Frequency and phase domains....
I am mentioning Ozone5 here because it is the only software I know of that has this type of feature....

So a theoretical digital phono stage could involve a PC with "matching" software to provide not only RIAA EQ, but also correction of frequency anomalies... (blasphemy I know!)

The end result could be that rather than seeking the perfectly "voiced" cartridge for ones system, one would choose one that matches the TT/Arm best (compliance / mass matching) - with the best possible stylus (low effective mass, high cantilever rigidity, good tracking ability) - and then rather than use crude cartridge loading methods to attempt to get a flat F/R, fine tune things with the Digital phono stage....

Would this allow a high quality budget cartridge to compete with the megabuck contenders? (eg: M97xE with SAS)

There are other aspects of performance that this approach does not cover... but it seems to me that in todays digital world this is the next logical step...

Comments?

bye for now

David

p.s. Although I have Ozone4, I have not bitten the bullet and purchased Ozone5 for experimentation....

[neobop]

David,
Interesting, to be honest I don't see the point.   You don't need a minimum phase(?) equalizer to correct the frequency response on a decent cartridge and the potential to degrade the sound is enormous.  I think your JLTI would kill the PC phono.   Would sensitivity on the PC be enough?  Even if it would, you won't make a budget cart sound like a great one.  The information isn't there to start out with.

Must be tough not having your gear at your disposal.  Only a couple more months?
neo

[dlaloum]

Hi Neo,

in a way I agree with you - my dynavector Karat has perfectly flat F/R and based on specs I believe flat phase/r as well...

But a Karat is a relatively expensive device!

What if the same result can be achieved with a good quality body, fitted with a good stylus (eg: the M97xE with SAS, AT440Mla, etc...) - the high inductance, and the lower resonant frequency of the cantilever can both be compensated for in a much more elegant manner than what I have been trying with cartridge loading....

Properly setting up a cartridge like this, and setting the loading to optimise for neutral voicing (flat f/r) takes a heap of time and effort.... With this approach, you can just run at standard 47k/250pf (or whatever capacitance your setup is at) - and let the relatively automated digital EQ take care of the rest - the tuning process for a cartridge would be under 30min rather than a weekends work.... And the result is likely to be superior...

Of course good quality front end gear is key - a good step up stage (possibly with RIAA) before the ADC - with level control so you can optimise the digital conversion process - then a good ADC, then the appropriate software, and finally a good DAC.

Given the cost of the digital gear and software this is more expensive than purchasing a Karat.... so probably a pointless intellectual exercise unless you already have the gear.

Intellectual exercises are what I am sticking to at the moment until I have the gear here!

All I have at the moment is a TV and a boom box.... (actually squeezebox boom connected to my digital music library on a NAS, with a Gallo sub attached - so it is not as bad as it sounds...)

bye for now

David

[neobop]

Hi David,
Of course that opens up the possibility of a digital phono stage having the potential to correct other aspects or imperfections in record playback, including the record itself?? 
If PCphono  could correct phase nonlinearities instead of just minimum phase filters for frequency response, it might be flat and have 3D imaging as well.
Imperfect records or pressings could be compensated for by altering the amplitude and phase response - separately.  One control for brightness and another labeled hall.  Requirements for the cart would be good detail and transient response.  Of course impulse noise reduction (click and pop) could be incorporated. 

I have no idea how you could accomplish this or if it could even be done, but it seems to me there is the possibility for digital domain correction that goes way beyond what is available for the end user now, whether the source material originates in digital or analog domain.  Sometimes I hear phase anomalies in recordings, both analog and digital.  Have you ever heard a recording where one mike is out of phase?  Correcting something like that would be a little trickier, well, maybe a lot trickier, but you couldn't do much worse than some remastering commercially available now.

I used to have an old Charlie Parker record ('40s) that was recorded off a portable tape machine, monitoring a little hallway speaker.  That one and a couple of others were reserved for playback with tone controls.  I guess there's only so much you can do if the source material is really poor quality, but we're talking home remastering here, at least to some extent.  Right now we're limited to EQ.  That, and playing with different cables is about it.  You could even save your remasters for future use, but if you change your speakers you might have to do it over.   

If the home remastering lab were available, it might be fun to record you stuff and play with the sound, but I think I'd rather just play a record.  Something is lost when an analog recording is digitized and that something is a big part of why records and record playing equipment are still around.  Besides, we need a hobby and something to drool over and spend our money on, sort of like classic cars only slightly more affordable.  There's no substitute for owning cool tables and a multi-thousand pound record collection. 

What kind of electricity do you guys use down under, 120/60?  You need a killer DD table like an old Goldmund with a JVC motor.  That would be a nice toy to celebrate setting up your equipment again.  It would make you forget about this digital phono, at least for awhile.  It might be interesting, but I suspect  your resources and time could be better allocated.  With your new family member you probably won't be able to listen to music while you're on duty, and while you're off duty time will probably be limited.   I'll see if I can find you a nice used Goldmund and send you the bill.   
neo

[dlaloum]

I remember auditioning a Goldmund reference in the late 80's at Kostas Metaxas Melbourne audio salon....

Played through his own electrostatic speakers, and amps.... truly wonderful sound.
And the table was (and still is) a technological and engineering marvel 

They are however a little bit rare - and even today still fetch a reasonable bit of change!

[neobop]

Yea, reasonable bit of change, you're not kidding, last one I saw for sale was $35K.   But the Reference is servo controlled belt drive.  The 35lb platter was a bit much for direct drive, and they didn't have Harry Weisfeld"s submarine motor.

I was thinking of a Studio or Studetto DD.  The ones with JVC motors are the ones to get.  Depending on the arm you might have to mod the suspension, could be a little tricky, but doable and much less expensive.
neo

[dlaloum]

I like the original linear tracker - wasn't it a modified rabco design? (or was Pierre Lurne involved.... or both?)

[neobop]

I think it is similar in some ways to Rabco, but a different design.  One nice thing is the arm tube is longer than most linear arms.  This will minimize record warp VTA anomalies and make VTA/SRA less of a problem in general.  That's the T3F arm on the Reference and Studio.  The Studietto had an optional T5 linear arm.  Mine was w/o the arm and Goldmund supplied two piece aluminum armboards for SME or Linn type mounts.  The last table from that era was the ST4 which came with a dedicated T4 linear tracker. 

Pierre Lurne designed some of the arms and I think he headed the design teams for the DD tables.  I'm not sure of the extent of his involvement, but he later sold a similar linear tracker through his own company.  Lurne later abandoned linear trackers for unipivots he sold with his tables at Audio Mecca.
There are a number of interesting interviews in Stereophile and TNT with Lurne. Do a google search and they'll come up.  That should make for interesting reading while you mark time.
neo

[dlaloum]

I think if I was to setup a TT/Arm (ie: manual rather than the automatic integrated setups I now have...) - I would look at something like a classic JVC or Denon DD with the Dynavector arm - that arm fascinates me, huge mass horizontally and adjustable mass vertically...

Reviews show it working well with everything from high compliance  V15V's and Grado's to mid compliance MC's (not sure I've seen a review of a true low compliance cartridge in the dynavector arm).... magnetically damped, versatile, and steam-punk sexy....