EQ Question: In the phono stage .vs. flat file eq'd by computer software

[allvinyl]

I'm anxious to learn opinions and experiences by users whom have used or live with a phono stage that has preset or adjustable eq curves and also heard eq applied to a flat file by computer software. A friend with the Channel D L20 phono is promoting the sonic superiority of feeding a flat signal generated by the L20 to the Pure Vinyl software for playback. I have no direct experience with the software but understand it either provides several curves or can gen curves based on user input. The same friend has listened to the Zanden 120 and prefers the software route. He is a headphone-only listener and also values the low noise floor the L20 provides. As some here may know, I am a traditional listener with the GTA3r speakers and am currently using the Channel D Lino C phono.

I recently had the Zanden 120 in my system and without applying any curve aside standard RIAA it is, to my ears, clearly a better phono than the Lino. Once you add the flexibility of being able to adjust phase and eq it becomes even less of a contest. I really like the 120 and it seems a good match with my DHT Coincident MKII linestage. I have also considered the 1200 MKIII and am interested in folks opinions on it as well.

Thanks in advance for any insights...

 

[Audiophile Bill]

I get what you are trying to do with such an approach but worry that the AD process more than offsets the “gains” with more accurate dsp based RIAA. I was planning this route a while ago (~6 years) using Acourate filters. I found that analogue going through the “transparent AD conversion” simply lost the magic. Some measurebators will tell you that the AD conversion is inaudible etc but imho it is a big no no and keeping analogue, analogue, is key to your final sound. Personally I wouldn’t countenance such an approach nowadays but I became quite a purist in my philosophy. I do get the inherent theoretical on paper advantages especially if you get fancy and add dsp room corrections and driver time alignment into the entire convolution process.

 

[allvinyl]

AB, I appreciate your reply. It seems as if my path of experiences is following where you have previously traveled. I understand where my friend is coming from but I also think his attachment to the S/N ratio of the L20 and headphone-only listening are the main factors in his conclusion. Also, there are compelling reviews on the web site.

I am not saying the Lino isn't good, it just didn't 'sound' as musical as the 120 in a direct comparison. I think the initial A/B was valid as all we did was swap the 120 for the Lino in my system. Everything else stayed the same including all the cabling. All three of us listening over those 2 days came to the same conclusion(s).

Lastly, switching the EQ curves for several of the LPs we used was pretty jaw-dropping. I especially wanted to hear this aspect in my own room as I spent 3 days listening to the two Zanden phono stages when I was helping out in the GT Audio Works/Zanden room at Axpona this past April.

Perhaps the final lesson, for me anyway, is to continue using my own ears as the arbiter in these decisions.

 

[DonH50]

As an avowed "measurebator" (had not heard that insult before) I can see definite advantages to doing the EQ in the digital domain -- assuming the front end before A/D conversion is low enough noise, can handle saturation cleanly and recover quickly, and the ADC itself has sufficient dynamic range and resolution to do the job. That likely covers many ADCs these days but I do have concerns about saturation (overload) recovery and dynamic range. The ability to tailor the EQ could be helpful though there are other ways to do it if you have a DSP in the system already.

The RIAA curve spans about 40 dB in dynamic range, a voltage ratio of 100:1. Not overly large by today's standards but significant. An ideal 16-bit ADC has about 98 dB of dynamic range so 40 dB of that is used just to handle the RIAA curve. That still leaves nearly 60 dB for the softest high-frequency signals, plenty for an LP, and hard to say how many of us can hear soft 20 kHz sounds anyway, but it is a consideration. I suspect the noise floor of the gain stages before the ADC sets the dynamic range more than the ADC itself, and of course the actual digital processing is probably at least 32 bits so plenty of margin there.

I have not heard a DSP-based phono stage recently, but in the past ticks and pops were a big issue, not that they are not for a conventional preamp as well. Bear in mind DSP, while I had a lot of experience in the past, is not my day job and I have not really touched it in ten years or so. For the analog circuits before the ADC, the problem is pretty much as it has always been: the very front end must handle that peak 20 dB HF boost and max 20 dB bass cut used in recording. A conventional preamp applies a corresponding 20 dB bass boost, which is a problem with warped records, and 20 dB treble cut. (Yes, I am looking at the frequency extremes.) The treble cut helps reduce the impact of wideband noise and ticks/pops (surface noise) to the stages following the initial RIAA filter. In a DSP-based solution, that noise hits the circuits before and the ADC itself. What happens then I do not know for modern designs because I have not looked.

In the past, wideband noise spikes overload the ADC, which saturates (clips) and takes a little bit to recover. With a flash -type ADC, recovery is usually very quick, within 1 or 2 samples for a good design. For a delta-sigma design, such an overload could influence the output for a pretty long time due to the cascaded internal stages and long digital filters. Oversampling means it may still recover very quickly relative to the final sample rate, I do not know (hopefully somebody with more recent audio design experience can chip in). Back when I was doing them, for RF pulse and com applications, it was a big problem. I added limiters before the ADC to keep it from clipping and presumably audio ADCs would do the same but, again, I do not know. Overload can (but may not) "blur" trailing edges and introduce long settlers into the response.

It would be interesting to compare measurements and listening test results for the two schemes under a variety of less than ideal situations, like an older or more worn LP.

For now, as for many years, my preference has been a tube phono preamp as, despite its typically higher noise and reduced bias stability (causing mainly gain to wander), it handles large overloads with aplomb. A few SS designs have high overload resistance and/or recover quickly; I have used both over the years, though my main phono preamp for years was an old ARC SP3a1 (with a Denon step-up transformer for MC as needed) despite rolling through a number of newer, better-spec'd SS designs. I do not have my TT set up now and cannot comment on the latest designs.

FWIWFM - Don

 

[Audiophile Bill]

AB, I appreciate your reply. It seems as if my path of experiences is following where you have previously traveled. I understand where my friend is coming from but I also think his attachment to the S/N ratio of the L20 and headphone-only listening are the main factors in his conclusion. Also, there are compelling reviews on the web site.

I am not saying the Lino isn't good, it just didn't 'sound' as musical as the 120 in a direct comparison. I think the initial A/B was valid as all we did was swap the 120 for the Lino in my system. Everything else stayed the same including all the cabling. All three of us listening over those 2 days came to the same conclusion(s).

Lastly, switching the EQ curves for several of the LPs we used was pretty jaw-dropping. I especially wanted to hear this aspect in my own room as I spent 3 days listening to the two Zanden phono stages when I was helping out in the GT Audio Works/Zanden room at Axpona this past April.

Perhaps the final lesson, for me anyway, is to continue using my own ears as the arbiter in these decisions.

I really wanted it to be the case that current AD converters and dsp algorithms were such that one couldn’t hear any deleterious ramifications but it wasn’t so. Tbh I haven’t heard dsp based speakers either that didn’t at some point ruin the final product. A long time ago, I tried to go with a Devialet but that too also ruined analogue.

The theoretical advantages are tantalising and promise so much, but don’t deliver for me.

 

[Audiophile Bill]

As an avowed "measurebator" (had not heard that insult before) I can see definite advantages to doing the EQ in the digital domain -- assuming the front end before A/D conversion is low enough noise, can handle saturation cleanly and recover quickly, and the ADC itself has sufficient dynamic range and resolution to do the job. That likely covers many ADCs these days but I do have concerns about saturation (overload) recovery and dynamic range. The ability to tailor the EQ could be helpful though there are other ways to do it if you have a DSP in the system already.

The RIAA curve spans about 40 dB in dynamic range, a voltage ratio of 100:1. Not overly large by today's standards but significant. An ideal 16-bit ADC has about 98 dB of dynamic range so 40 dB of that is used just to handle the RIAA curve. That still leaves nearly 60 dB for the softest high-frequency signals, plenty for an LP, and hard to say how many of us can hear soft 20 kHz sounds anyway, but it is a consideration. I suspect the noise floor of the gain stages before the ADC sets the dynamic range more than the ADC itself, and of course the actual digital processing is probably at least 32 bits so plenty of margin there.

I have not heard a DSP-based phono stage recently, but in the past ticks and pops were a big issue, not that they are not for a conventional preamp as well. Bear in mind DSP, while I had a lot of experience in the past, is not my day job and I have not really touched it in ten years or so. For the analog circuits before the ADC, the problem is pretty much as it has always been: the very front end must handle that peak 20 dB HF boost and max 20 dB bass cut used in recording. A conventional preamp applies a corresponding 20 dB bass boost, which is a problem with warped records, and 20 dB treble cut. (Yes, I am looking at the frequency extremes.) The treble cut helps reduce the impact of wideband noise and ticks/pops (surface noise) to the stages following the initial RIAA filter. In a DSP-based solution, that noise hits the circuits before and the ADC itself. What happens then I do not know for modern designs because I have not looked.

In the past, wideband noise spikes overload the ADC, which saturates (clips) and takes a little bit to recover. With a flash -type ADC, recovery is usually very quick, within 1 or 2 samples for a good design. For a delta-sigma design, such an overload could influence the output for a pretty long time due to the cascaded internal stages and long digital filters. Oversampling means it may still recover very quickly relative to the final sample rate, I do not know (hopefully somebody with more recent audio design experience can chip in). Back when I was doing them, for RF pulse and com applications, it was a big problem. I added limiters before the ADC to keep it from clipping and presumably audio ADCs would do the same but, again, I do not know. Overload can (but may not) "blur" trailing edges and introduce long settlers into the response.

It would be interesting to compare measurements and listening test results for the two schemes under a variety of less than ideal situations, like an older or more worn LP.

For now, as for many years, my preference has been a tube phono preamp as, despite its typically higher noise and reduced bias stability (causing mainly gain to wander), it handles large overloads with aplomb. A few SS designs have high overload resistance and/or recover quickly; I have used both over the years, though my main phono preamp for years was an old ARC SP3a1 (with a Denon step-up transformer for MC as needed) despite rolling through a number of newer, better-spec'd SS designs. I do not have my TT set up now and cannot comment on the latest designs.

FWIWFM - Don

Very interesting - thanks for your insights.

Have you compared your valve phonostage with and without dsp after the preamp just to see whether the ADA is good enough for you? I have heard quite a few implementations - the most horrific was a Munich demonstration of the Techdas AF1 into Wilson Benesch speakers with Trinnov. A total disaster.

 

[DonH50]

Very interesting - thanks for your insights.

Have you compared your valve phonostage with and without dsp after the preamp just to see whether the ADA is good enough for you? I have heard quite a few implementations - the most horrific was a Munich demonstration of the Techdas AF1 into Wilson Benesch speakers with Trinnov. A total disaster.

Thank you. And no, I have not had my phono rig running in 10 years or so. Last time I did any comparing was probably around the late 1990's/early 2000's and at that time preferred my valve (tube over here) preamp. Noise was higher and pops and ticks would absolutely trash the sound. ADCs and DACs have come a long way since then thus all my waffling about what might or might not sound better today. I flat out do not know.

My current system includes DSP for processing and room correction; I listen mainly to CDs and SACDs though work hours are such that most of the time it is just on the boob tube (a much nicer boob tube than I used to have, and actually no tube involved, but programmatically much the same though movies are much better -- when I have time to watch!) The times I have compared analog to digital or "direct" I have preferred using digital processing, but much of that is because my room has issues that really require DSP to solve (though now I have multiple subs to handle room modes I could go back, but I've gotten into surround sound a bit and don't want to go back).

 

[Audiophile Bill]

Thank you. And no, I have not had my phono rig running in 10 years or so. Last time I did any comparing was probably around the late 1990's/early 2000's and at that time preferred my valve (tube over here) preamp. Noise was higher and pops and ticks would absolutely trash the sound. ADCs and DACs have come a long way since then thus all my waffling about what might or might not sound better today. I flat out do not know.

My current system includes DSP for processing and room correction; I listen mainly to CDs and SACDs though work hours are such that most of the time it is just on the boob tube (a much nicer boob tube than I used to have, and actually no tube involved, but programmatically much the same though movies are much better -- when I have time to watch!) The times I have compared analog to digital or "direct" I have preferred using digital processing, but much of that is because my room has issues that really require DSP to solve (though now I have multiple subs to handle room modes I could go back, but I've gotten into surround sound a bit and don't want to go back).

Don - what room correction software (or hardware) are you using now? Did you do much testing of the various options? Have you tried room correction only for the bass and leaving the mids/treble to passive filters?

 

[allvinyl]

As an avowed "measurebator" (had not heard that insult before) I can see definite advantages to doing the EQ in the digital domain -- assuming the front end before A/D conversion is low enough noise, can handle saturation cleanly and recover quickly, and the ADC itself has sufficient dynamic range and resolution to do the job. ...

...
I have not heard a DSP-based phono stage recently, but in the past ticks and pops were a big issue, not that they are not for a conventional preamp as well. ...

...
In the past, wideband noise spikes overload the ADC, which saturates (clips) and takes a little bit to recover. With a flash -type ADC, recovery is usually very quick, within 1 or 2 samples for a good design. ...

FWIWFM - Don

I know the newest version of the Pure Vinyl software has a tick-pop remover that my friend says works wonderfully. I've only heard the results once through his headphones so I really can't speak to its overall effect(s).

Also, I believe the L20 circuits address the overload issue, though I can't speak definitively about the hardware. As a side note, the Lino C is a reduced to a single processor version of the 16 processor L20. There are also differences in how the battery is implemented but I can tell you it is very quiet.

From the Channel D site:
"In short, ultra-ultra wide frequency bandwidth delivers the ultimate in sound staging and definition, especially when paired with the best moving-coil cartridges. Additionally, the outstanding signal to noise ratio of the L20 provides the blackest of black backgrounds for "hearing" and enjoying the silence between the musical notes.

This feat was accomplished by connecting 32 (16 per channel) of the proven, ultra wide bandwidth front end preamplifier modules from our $4998 Seta L in parallel (a detail not revealed in our news release for the L20)."

John

 

[DonH50]

Don - what room correction software (or hardware) are you using now? Did you do much testing of the various options? Have you tried room correction only for the bass and leaving the mids/treble to passive filters?

Dirac Live in my main system. I have used Audyssey and Audyssey XT (not XT32), MCACC, and YPAO in the past.

The most recent "main" systems used MCACC and then (and now) Dirac Live. In all cases I run the program, then use a measurement mic and analysis SW (was an Earthworks M30 and pro SW, now a CSL UMIK-1 and REW) to tweak settings for best response. Mainly because dialing in multiple subs with the mains in my room is a PITA.

I have tried various schemes for only the bass but my room has a multitude of absorbers to help solve bass modes. I could rip them open and add plastic membranes to liven the room, but so far (for like 10+ years now) I've just let the room correction program tweak up the treble to compensate. I do not mind a "dead" room as much as others. I am not sure what you mean by "passive filters"; other than the processor and crossover I don't have passive filters in my system now. With my old ARC I did use passive filters to implement a bi-amped system for my old Maggies but that was years ago.

I have also run without any sort of room correction (heck, I started piddling with this stuff in the early 1970's, long before room correction SW was a thing). By and large, in my current room, I prefer correction. I have been backing off the correction and treatments (pulled some off the walls) lately (in the past year or two) whilst playing with my new speaker setup.

I much appreciate the ability to adjust Dirac Live's target curve; none of the other options I have had allowed that though MCACC did allow you to go in and tweak PEQ settings after the run. I actually got the flattest response and best step response with MCACC but Dirac Live does very well, is much more powerful, and much more flexible. I probably have not spent as much time tweaking it. My work hours have been crazy the past few years so I'm more into getting it good enough and enjoying it when I can than spending hours measuring and tweaking by ear. I tend to spend some time (hours, days, week or three) measuring and tweaking when I first get a new toy, then once it is measuring and sounding good I tend to turn off the engineer and engage the musician when listening. And I am pretty tolerant of sound when watching movies; I focus mainly on the video action. The exception is when I watch a music DVD/BD; bad sound kills it for me.

@allvinyl: Not ignoring you, John, just really nothing to say. I have no experience with that system, and almost none with phono stages generally for a number of years now, so am not competent to comment. I can talk a bit to past experiences and the technical issues I have seen but (as folk tend to remind me) that may not apply at all today.

 

[Audiophile Bill]

Dirac Live in my main system. I have used Audyssey and Audyssey XT (not XT32), MCACC, and YPAO in the past.

The most recent "main" systems used MCACC and then (and now) Dirac Live. In all cases I run the program, then use a measurement mic and analysis SW (was an Earthworks M30 and pro SW, now a CSL UMIK-1 and REW) to tweak settings for best response. Mainly because dialing in multiple subs with the mains in my room is a PITA.

I have tried various schemes for only the bass but my room has a multitude of absorbers to help solve bass modes. I could rip them open and add plastic membranes to liven the room, but so far (for like 10+ years now) I've just let the room correction program tweak up the treble to compensate. I do not mind a "dead" room as much as others. I am not sure what you mean by "passive filters"; other than the processor and crossover I don't have passive filters in my system now. With my old ARC I did use passive filters to implement a bi-amped system for my old Maggies but that was years ago.

I have also run without any sort of room correction (heck, I started piddling with this stuff in the early 1970's, long before room correction SW was a thing). By and large, in my current room, I prefer correction. I have been backing off the correction and treatments (pulled some off the walls) lately (in the past year or two) whilst playing with my new speaker setup.

I much appreciate the ability to adjust Dirac Live's target curve; none of the other options I have had allowed that though MCACC did allow you to go in and tweak PEQ settings after the run. I actually got the flattest response and best step response with MCACC but Dirac Live does very well, is much more powerful, and much more flexible. I probably have not spent as much time tweaking it. My work hours have been crazy the past few years so I'm more into getting it good enough and enjoying it when I can than spending hours measuring and tweaking by ear. I tend to spend some time (hours, days, week or three) measuring and tweaking when I first get a new toy, then once it is measuring and sounding good I tend to turn off the engineer and engage the musician when listening. And I am pretty tolerant of sound when watching movies; I focus mainly on the video action. The exception is when I watch a music DVD/BD; bad sound kills it for me.

@allvinyl: Not ignoring you, John, just really nothing to say. I have no experience with that system, and almost none with phono stages generally for a number of years now, so am not competent to comment. I can talk a bit to past experiences and the technical issues I have seen but (as folk tend to remind me) that may not apply at all today.

Yes sorry for my sloppy wording - I just meant in an active system using dsp on the bass with passive crossover for mids/tweeter.

Have you tried Dirac via mini DSP?

 

[DonH50]

Yes sorry for my sloppy wording - I just meant in an active system using dsp on the bass with passive crossover for mids/tweeter.

Have you tried Dirac via mini DSP?

NP. DSP on bass only: tried, yes, but for room reasons not doing it now. Dirac Live allows you to limit the correction range so it is relatively easy to try. It is on my list to try again.

I have an Emotiva XMC-1 with the "full" version of Dirac Live so don't need a miniDSP (i.e. "no"). The base miniDSP model is not all that great and the ones that do Dirac Live are more expensive than I want to pay just to try it on a different platform. Yes, I know this is WBF, but "Best" for me is an order of magnitude or three below what y'all can afford. I blew all my money on speakers.

 

[allvinyl]

...
Not ignoring you, John, just really nothing to say. I have no experience with that system, and almost none with phono stages generally for a number of years now, so am not competent to comment. I can talk a bit to past experiences and the technical issues I have seen but (as folk tend to remind me) that may not apply at all today.

Don - I understand because we are in the same position. I've only ever read about systems setup as yours is configured and my limited actual exposure has been at shows. I am trying to use the crude tools I have to guide me and so far have used app software, room treatments and speaker/listener positioning to tune response at the listening chair(s). Increasingly, I feel as if I will likely end up where you are as I anticipate downsizing knowing that a dedicated listening space will probably collapse to become part of the living space of any place we move to.

That's how this whole thing started for me as I anticipated ripping all of my vinyl to digital media and playing it back that way selling off the record collection as I go or as finances dictate needing to raise money. That, and the fact that even though my daughter/son-in-law like vinyl they aren't off into the variety of minutia it takes to really maintain state of the art vinyl playback. In the end, they'd be happy with the music and would love the convenience of the digital playback. I'm trying to optimize the digital result of vinyl ripping and like the theory of the flat file optimized at playback by software. So the theory kind of fell apart on me when we put the Zanden 120 into my system and I had access to the EQ curves. As Audiophile Bill wrote, the magic was lost. Well, in my case the magic came back with the 120 in the system. So, before I go any farther building out the ripping environment around the Lino, I felt I should slow down and reassess. Ask questions, think about it some more.