24-Bit/192 kHz USB Audio Interface for vinyl A/D archiving?

There are several inexpensive 16bit/48kHz phono to USB 1.1 devices available to record vinyl to PC, but no one step solutions for higher res files. Would need at least a USB 2.0 interface. Wouldn't really need higher than CD quality for most LPs, but it would be nice to have options for those rare master quality records.

So far I've found a decent Schiit Phono preamp and been looking at USB home studio recording A/D USB converters that do 24-bit/192kHz. The weak link might be the A/D unit's built in preamps.
 

jaddie

Member
2011-04-01 12:06 am
You never need higher than CD quality for any LP, as all LPs are below CD quality.

However, there is one reason to capture vinyl higher than 48kHz, and that's if you plan to use a de-click process, many of which focus on the extreme high frequency produced by a record click, so another octave may be useful, or not at all depending on the de-click algorithm. Other than that, response beyond 24kHz is completely unnecessary for digitizing vinyl.

The capability of vinyl to reproduce high frequencies decreases rapidly with rising frequency and progressively inner grooves due to physical constraints of a finite stylus in a finite groove. Maximum modulation velocity is a function of frequency and groove location. If the cutting stylus moves too fast due to a combination of frequency and level for a give groove location (you can go faster at the outer edge, much slower on the inner most groove because of the inherent linear groove velocity), the rear facet of the stylus will gouge the groove wall just cut be the forward facet. So the recorded spectrum on vinyl is limited physically, and is maximum only on the outer grooves. The maximum stylus velocity for an outer groove LP is 50cm/S, which will prevent even 20khz from being ever recorded at a level as high as a 1kHz signal. The RIAA curve keeps things in check until about 7 or 7kHz for outside edge grooves, but beyond that highs must be limited in amplitude by other means or the groove cut groove will be damaged in cutting. The result of all of this is that flat response can only be maintained at a level well below maximum output level, thus limiting high level, high frequency content rather effectively. Fortunately, most music has spectral density that also falls rapidly with rising frequency, which is why this all works out well in practice.

The HF content some claim exists on vinyl is often the result of misinterpreting spectrum analysis. The stylus-groove-stylus mechanical system is loaded with distortion mechanisms which grow with rising frequency, and show up as products above 20kHz when the actual modulation is above 6kHz.

The DR of vinyl is limited fist by the vinyl itself, none of which is capable of a full spectrum noise floor even at 12 bits (re: maximum groove velocity of 50cm/S found only on outer grooves). The noise floor is a function of the vinyl surface itself, influenced by vinyl formula and variances in pressing. Then a secondary limitation occurs because of platter rumble, both in the lathe and the reproducing turntable.

As to the viability of interfaces targeted at home recording, pretty much any are fine. Whatever limitations there may be in the A/D or preamps, they are far below vinyl's requirements, especially if you use the Schiit preamp into a line input. The residual noise and distortion of any line level USB interface is orders of magnitude below anything found in the complete vinyl system.

So don't go too exotic.
 

jaddie

Member
2011-04-01 12:06 am
Correction...

Thanks, but I wouldn't quite go that far. I have done a comparison of a CD and LP both cut from identical masters with no changes (I supervised both), played on level-matched devices, with the vinyl system calibrated for flat response post RIAA, and they sounded shockingly alike, with only the surface noise being the obvious difference, but only on the first few plays. Then record wear became an additional tell. I expected a far greater difference given the capabilities.

Still no need for greater than 16/48 though.
 

MAAC0

Member
2010-05-02 10:00 pm
You never need higher than CD quality for any LP, as all LPs are WAY below CD quality.

I'm no expert but:

The recording/tracking ability of vinyl is easily at least 50 kHz depending upon the cartridge and stylus profile, and perhaps as high as 100 kHz. The most notable proof of this is the CD4 quadraphonic system which relied on a 45 kHz bandwidth to be accurately reproduced. That said, the high-frequency response accuracy of vinyl varies tremendously. Amplitude deviations of 5-10 dB or greater are not uncommon in the 20 kHz range for many records, but many records are cut faithfully to the source material and a high quality record player and cartridge combination should have no problems faithfully reproducing the recorded sound.
 

jaddie

Member
2011-04-01 12:06 am
I'm no expert but:

The recording/tracking ability of vinyl is easily at least 50 kHz depending upon the cartridge and stylus profile, and perhaps as high as 100 kHz. The most notable proof of this is the CD4 quadraphonic system which relied on a 45 kHz bandwidth to be accurately reproduced. That said, the high-frequency response accuracy of vinyl varies tremendously. Amplitude deviations of 5-10 dB or greater are not uncommon in the 20 kHz range for many records, but many records are cut faithfully to the source material and a high quality record player and cartridge combination should have no problems faithfully reproducing the recorded sound.
Well...that's a very miss-common conception of vinyls real capabilities. You cannot ignore the level portion of the frequency vs level profile of vinyl. Yes, you can get extremely low level 50kHz on and off (the CD 4 carrier was 30kHz with a complex modulation scheme), but human hearing becomes less sensitive above 15kHz quite remarkably, and vanishes above 20kHz (no, don't bother citing exceptions). That's why 30kHz was chosen for the CD4 carrier, and it's low level was necessary because of physical dimensions of the groove wall and stylus.

High frequency response of a recording system is not just a go/no-go, on/off parameter. To be useful, it has to provide replication of some audible event. When super-audible frequencies, which can't be heard at levels similar to those found mid-band, are produced 30dB lower, they're simply audible gone.

Vinyl cannot capture full amplitude frequencies above 3kHz on the inner grooves, by "full amplitude" I mean at an equivalent level to maximum modulation mid-band. It's capabilities decrease with rising frequency, and this is not debatable, it's documented and well known. It's a limitation that must be accounted for during mastering.

16/48 captures everything, flat, no amplitude change, up to 24kHz. That's already 4kHz above the maximum audible frequency for most of the population.
 

plasnu

Member
2011-01-27 4:39 pm
NY USA
I would recommend a professional or prosumer ADA converter. They are very reasonable price (and large used market), and many of them include mic preamp for 40dB+ gain. With a digital RIAA EQ, you can bypass phono preamp. Those builtin preamps are not bad at all in terms of cleanness. You might have to spend $1000 for better mic pre.
 
Last edited:
Last time I recorded an old LP with a Fostex FR2-LE at 96 kHz sample rate, there definitely was some signal visible above 22050 Hz when I watched a moving FFT of the recording, and it obviously wasn't an alias. So in that sense LP is definitely better than CD.

Whether that or anything else is audible is a different matter altogether, but whether high sample rates are useful was not the thread starter's question. Besides, you don't know whether Steve listens alone or in the company of a cat or a Japanese gamelan player.
 
I've spent a fair amount of time designing phono pre-amplifiers both for MM/MC and strain gauge in the past few years, and the need to do measurements of playback in particular for EQ'ing a strain gauge cartridge was a heady dose of reality.

Jaddie has really provided the straight score here.

I believe most of what you will see above 15kHz in vinyl playback except on the outermost grooves of a new record is going to be distortion.

16/48 with a good AD converter based on noise floor, dynamic range and distortion should be more than good enough.

Veils of ignorance torn away I am still very impressed with just how good vinyl can actually sound, the objective evidence doesn't really support that observation.

People reference CD-4, you should take a look at a typical CD-4 capable cartridge response, in many cases it's not going to be quite what you expect.
 

VaNarn

Member
Paid Member
2010-12-11 12:39 am
Ivanhoe
An accurate account of the high frequency limitations imposed in the cutting and replay of vinyl recordings by jaddie. I would mention that to obtain a clean carrier frequency on a CD4 recording required cutting at half speed . To reduce the wear problem on the playback of the CD4 discs a special plastic formulation was used and the Shibata profile stylus/pickup cartridge with an extended HF response being another requirement.
 

jaddie

Member
2011-04-01 12:06 am
It's always better to record at higher sampling rate because it relaxes the low pass filter at conversion.
That was absolutely true in the days of high-order analog filters. Those days are long gone. Current filtering is roughly described as digital over-sampling filtering, and the results are far superior.

Early analog anti-aliasing and reconstruction filters suffered from high frequency nonlinearities that resulted in some measurable (but arguably inaudible)in-band phase shift, but the real problem was high frequency intermodulation distortion that caused "images" of high frequency signals to be folded down into the mid and upper-midband.

Again, those days are long gone.
 

jaddie

Member
2011-04-01 12:06 am
People reference CD-4, you should take a look at a typical CD-4 capable cartridge response, in many cases it's not going to be quite what you expect.
I'm a relic of the CD-4 days. Well, ok, I'm just a relic. But I did have hands-on experience with CD-4. The simple answer is, it only sorta worked, the sound quality was not that impressive, and record wear was 10X the problem of normal vinyl.

The short story is, CD-4 never made any significant market penetration, and didn't even survive as an archivable medium. Some of that could be said of all "quad" formats on vinyl, but at least there were more releases, they were more or less stereo compatible, while they did require specific decoding hardware, they didn't require special stylii, cartridges, tone arm wiring...etc., etc....

CD-4 is not to be taken as proof of vinyl's high frequency capability. It's a special case, and one that ultimately failed.
 

jaddie

Member
2011-04-01 12:06 am
Last time I recorded an old LP with a Fostex FR2-LE at 96 kHz sample rate, there definitely was some signal visible above 22050 Hz when I watched a moving FFT of the recording, and it obviously wasn't an alias. So in that sense LP is definitely better than CD.
Classic case of not knowing what you're looking at.

Just a test here: what's the 3rd harmonic of 8kHz? 24kHz. At what level would harmonics be for 3% THD? -30dB. 3% is a "good" figure for a stylus tracking a high frequency groove.

So, now, what were you seeing above 22.5K? And how would you know if it's real audio or distortion products?

The visible presence of signal above 20kHz is not proof of the existence of real or useful audio. It's a measure of the inability of the system to reproduce a clean signal at high frequencies.

Whether that or anything else is audible is a different matter altogether, but whether high sample rates are useful was not the thread starter's question. Besides, you don't know whether Steve listens alone or in the company of a cat or a Japanese gamelan player.
My cat doesn't seem to care if the music I play him contains extreme HF. I've played him music on a low-bandwidth computer speaker system, and I've played him the same music on my 'good' system. His disinterest in both is equally intense.

But let's address something else here. How are you going to get, say 25kHz to your ears? Assuming your tweets will even do that at all, and assuming you're close enough that air absorption isn't a factor, their directional pattern is progressively narrowing as frequency rises, to the point where at ultrasonics (25kHz+) they radiate something like a pencil beam. While mid and normal HF reflects around the room, every surface absorbs ultrasonic energy to a much greater degree, many nearly completely. The 24kHz "sweet spot" is so tiny so position critical that you probably can't be in it with more than one ear at a time.

My cat's idea of the "sweet spot" includes more of a different kind of energy: heat.

I'm not aware of the hearing acuity of Gamelan players as being significantly different than average for their age.
 

jaddie

Member
2011-04-01 12:06 am
I would recommend a professional or prosumer ADA converter. They are very reasonable price (and large used market), and many of them include mic preamp for 40dB+ gain. With a digital RIAA EQ, you can bypass phono preamp. Those builtin preamps are not bad at all in terms of cleanness. You might have to spend $1000 for better mic pre.

A mic preamp input has a very inappropriate load characteristic for a MM cartridge. Carts expect 47K, the typical mic pre is about 2K.

Spending up on a mic pre doesn't change this (or anything else). The real art in mic preamps relates low noise, maximum gain, and how they load different kinds of microphones, which does affect the end result audibly. A recently introduced Drawmer channel strip includes adjustable mic loading, something I think long over due.

But none of that, and I do mean none of that, has any value to reproducing the output of a phono cartridge. Not only do carts expect 47K, they also expect a capacitive load range. The total load is a function of the cable, the deliberate load, and the virtual load of the front end of the phono preamp. These factors have been well known since the late 1970s, but still are ignored in practice today.

I'm not an advocate of using a mic input for a phono cart, then applying RIAA as a post process, because the RIAA curve will not be tracked well at the input stage making the precision digital application inappropriate. I am fully aware that there are many proponents of the technique.

Get a good phono pre, then use the line input on the ADC. It's not expensive.

The ART USB Phono Plus provides the right input stage, 47K 100pf cart load, RIAA eq, 16/48 ADC, selectable low-cut filter, 90dB s/n re: 0dBFS, front panel gain trim, clip light, and monitor in one box for $100. I'm not convinced vinyl transfer demands much more than that.
 
Thanks, but I wouldn't quite go that far. I have done a comparison of a CD and LP both cut from identical masters with no changes (I supervised both), played on level-matched devices, with the vinyl system calibrated for flat response post RIAA, and they sounded shockingly alike, with only the surface noise being the obvious difference, but only on the first few plays. Then record wear became an additional tell. I expected a far greater difference given the capabilities.

Still no need for greater than 16/48 though.


Well the similarity is due to the ear not being able to distinguish the higher quality of CD with a complex signal - give a variety of single and two tone tests you should easily be able to pick out harmonic and intermodulation distortion from loud tones on vinyl.


Pre-echos are a great way to distinquish vinyl/CD on some recordings, the clocks in "Time", on DSotM for example.
 
My cat doesn't seem to care if the music I play him contains extreme HF. I've played him music on a low-bandwidth computer speaker system, and I've played him the same music on my 'good' system. His disinterest in both is equally intense.

It depends very much on the cat. Both systems probably sound poor to him, with his 85 kHz hearing range (if he is a young cat with no hearing damage). Then again, the same holds for my cat and he responds to bird sounds and cat sounds even when they are played through the television. He also tends to get nervous when I play dancehall reggae through my normal audio system.

But let's address something else here. How are you going to get, say 25kHz to your ears? Assuming your tweets will even do that at all, and assuming you're close enough that air absorption isn't a factor, their directional pattern is progressively narrowing as frequency rises, to the point where at ultrasonics (25kHz+) they radiate something like a pencil beam. While mid and normal HF reflects around the room, every surface absorbs ultrasonic energy to a much greater degree, many nearly completely. The 24kHz "sweet spot" is so tiny so position critical that you probably can't be in it with more than one ear at a time.

Those are all things that cause relatively smooth roll-off. That's why cats and rodents are very well capable of hearing frequencies in that range indoors.

My cat's idea of the "sweet spot" includes more of a different kind of energy: heat.

Yes, that is very typical of cats (-;

I'm not aware of the hearing acuity of Gamelan players as being significantly different than average for their age.

In the 1990's, Japanese researchers published the results of a very peculiar experiment involving Japanese gamelan players, supertweeters and electroencephalograms. I still don't know what to make of it.
 
Last edited:

jaddie

Member
2011-04-01 12:06 am
OK, but then what causes the supposed brick-wall filter behaviour of a vinyl record-playback chain?
Huh? What "brick wall filter"?
I would expect everything in the chain to roll-off smoothly, quite unlike the anti-alias filters used for low sample rate digital recordings, and also quite unlike the 15 kHz low-pass filters used for FM stereo.
It's not a "smooth" roll-off, it's not a roll off at all, it's a physical geometric limit to the maximum stylus velocity. Stylus velocity is determined by amplitude and frequency, limited geometrically by groove dimensions and the linear velocity of the groove determined by the groove diameter. But what it means is the system has no capability to record or reproduce high level high frequencies. And low level high frequencies cannot be heard because of masking by dominant signals mid-band, and by human hearing insensitivity. Components above 20kHz at low (recordable) levels do not contribute audibly, and in FFT spectrum analysis are swamped by distortion products of similar or higher level.
 

jaddie

Member
2011-04-01 12:06 am
Well the similarity is due to the ear not being able to distinguish the higher quality of CD with a complex signal - give a variety of single and two tone tests you should easily be able to pick out harmonic and intermodulation distortion from loud tones on vinyl.
Partially agreed...but vinyl is also not the worst HF IMD offender in the system. Take a look at tweeters!
Pre-echos are a great way to distinquish vinyl/CD on some recordings, the clocks in "Time", on DSotM for example.
Yup, totally true. Pre-echos are the result of the need for playing time vs the degree of tolerable pre-echo. More land between the grooves, lower (or no) pre-echo, and lower play time. Another one of the many vinyl quality trade-offs.