DSP-Based Phono Preamp/ADC with Digital Out?

[metaleonid]

Just curious the input referred noise on the Tascam is -120dB (no weighting mentioned) and the Apogee is -129dB with no weighting per the data sheets. I assume these are at 200 Ohm or so typical mic impedances.

I didn’t pay attention to the values the specs in either device. All I can tell is that I plugged both devices to my turntable and made digital recordings of the same thing as follows: cartridge up in the air, then cartridge lowered on the record surface while the record is not spinning and then spinning the record with a certain musical fragment. I then adjusted the volume using the musical fragment as a reference. Clearly the noise on Apogee was way more.
Usually when I plug TASCAM to my TT, turn its gain all the way up and have the cartridge up in the air, the noise level is around -70 dB. Oh and by the way, I used the XLR to RCA adapter with cartridge loading of 47 Ohms. ( Ohms, NOT KOhms)

 

[scott wurcer]

Usually when I plug TASCAM to my TT, turn its gain all the way up and have the cartridge up in the air, the noise level is around -70 dB. Oh and by the way, I used the XLR to RCA adapter with cartridge loading of 47 Ohms. ( Ohms, NOT KOhms)

I use an external battery powered FET impedance converter to accommodate MC/MM/MI carts with selectable R/C loading. The mic pre-amps typically have 1k-2k input impedance and most MM carts don't like this. So I guess I'll remain curious since I won't be trying the same experiment.

 

[metaleonid]

I use an external battery powered FET impedance converter to accommodate MC/MM/MI carts with selectable R/C loading. The mic pre-amps typically have 1k-2k input impedance and most MM carts don't like this. So I guess I'll remain curious since I won't be trying the same experiment.

You do not want to plug MM and MI carts to mic preamp. For those you have instrument inputs. Mic preamp is only for MC carts. I have 2 very low output ones: Ortofon Cadenza Black and Lyra Etna SL.

I wouldn’t be bothering with this. But this is the only optimal option if you want to digitally transfer records using slow speed method.
In fact I was googling for adjustable analog RIAA converters and it took me here. I’m really trying to see if there is any way to adjust RIAA such that if I play back 33 RPM records at 22.5 RPM, it equalizes the right frequencies. So far I haven’t found it.

I’m not familiar with FET impedance converter. But isn’t it for microphones? How’s you accommodate MM and Grado?

PS. I do keep WAV files of some of my experiments if curious.

 

[scott wurcer]

You do not want to plug MM and MI carts to mic preamp. For those you have instrument inputs. Mic preamp is only for MC carts. I have 2 very low output ones: Ortofon Cadenza Black and Lyra Etna SL.

I wouldn’t be bothering with this. But this is the only optimal option if you want to digitally transfer records using slow speed method.
In fact I was googling for adjustable analog RIAA converters and it took me here. I’m really trying to see if there is any way to adjust RIAA such that if I play back 33 RPM records at 22.5 RPM, it equalizes the right frequencies. So far I haven’t found it.

Sorry I forget this is not so much a DIY site, I modify the situation if needed. What you are asking for is not difficult, just move the RIAA time constants down by 0.675 but to do that in analog requires a custom pre-amp, and you still need the frequency shift. If you are doing it digitally the frequency shift and normal RIAA works fine. BTW moving the arm cart resonance up by ~1.5 might be an issue.

 

[metaleonid]

Sorry I forget this is not so much a DIY site, I modify the situation if needed. What you are asking for is not difficult, just move the RIAA time constants down by 0.675 but to do that in analog requires a custom pre-amp, and you still need the frequency shift. If you are doing it digitally the frequency shift and normal RIAA works fine. BTW moving the arm cart resonance up by ~1.5 might be an issue.

Thank you, Scott. By custom analog pre-amp you mean DIY? If so, unfortunately I’m not that advanced.
I haven’t taken into account the arm resonance. I have completely forgotten about it. But if the end result is at the same speed as supposed to be, will the resonance still affect the sound? If so, in what way?

PS. Care to hear some samples?

 

[Frank Dernie]

Thank you, Scott. By custom analog pre-amp you mean DIY? If so, unfortunately I’m not that advanced.
I haven’t taken into account the arm resonance. I have completely forgotten about it. But if the end result is at the same speed as supposed to be, will the resonance still affect the sound? If so, in what way?

PS. Care to hear some samples?

PMFJI
The output from a seismic type transducer ( which is what a pickup cartridge is) is inaccurate up to around 2x the natural frequency of the effective mass on the suspension compliance. Simply put most of the output from the cartridge below that frequency is due to the mass of the cartridge bouncing on the suspension spring and as it approaches 2x the natural frequency the mass becomes quasi-stationary so subsequent output is due to stylus movement - ie it starts measuring the groove fairly accurately.
That is fixed by the physics, so if you run the record slowly the lowest accurately picked up frequency when you correct the pitch will go up in proportion to the difference in speed.
In short recording the output of a slow record and changing the pitch later may improve the HF of the recording but will significantly lose bass and/or bass accuracy, depending whether you properly filter the signal or not.
Not my personal idea of a good trade, given the strengths and weakness of LPs and record players.

 

[metaleonid]

PMFJI
The output from a seismic type transducer ( which is what a pickup cartridge is) is inaccurate up to around 2x the natural frequency of the effective mass on the suspension compliance. Simply put most of the output from the cartridge below that frequency is due to the mass of the cartridge bouncing on the suspension spring and as it approaches 2x the natural frequency the mass becomes quasi-stationary so subsequent output is due to stylus movement - ie it starts measuring the groove fairly accurately.
That is fixed by the physics, so if you run the record slowly the lowest accurately picked up frequency when you correct the pitch will go up in proportion to the difference in speed.
In short recording the output of a slow record and changing the pitch later may improve the HF of the recording but will significantly lose bass and/or bass accuracy, depending whether you properly filter the signal or not.
Not my personal idea of a good trade, given the strengths and weakness of LPs and record players.

Hi Scott. Thanks again. Ok, the practice shows the opposite: the slow speed method transfers are bassier but have less high frequencies. If I use speed 16 and 2/3 on 33 and 1/3 records the lack of HF is even more apparent. Now I guess I can relate this to what you wrote.
And obvious question. How can this be corrected say in a sound editor or software. I use clickrepair equalizer program that allows to basically add a custom curve for equalization. That can be added to RIAA to solve that problem.
By the way, even if the flavor of the sound is a bit off, it is still more enjoyable than having IGD.

 

[Frank Dernie]

Hi Scott. Thanks again. Ok, the practice shows the opposite: the slow speed method transfers are bassier but have less high frequencies. If I use speed 16 and 2/3 on 33 and 1/3 records the lack of HF is even more apparent. Now I guess I can relate this to what you wrote.
And obvious question. How can this be corrected say in a sound editor or software. I use clickrepair equalizer program that allows to basically add a custom curve for equalization. That can be added to RIAA to solve that problem.
By the way, even if the flavor of the sound is a bit off, it is still more enjoyable than having IGD.

I am Frank, not Scott PMFJI but record players was my job almost 50 year ago now and the physics of how cartridges work hasn't changed.

It will sound bassier because you are listening to more of the cartridge body bouncing on its suspension - not because you are more accurately transcribing the bass from the groove.

It can not be corrected afterwards because it is a physical effect which depends on the laws of physics, using this type of transducer is inherent to record players and its limitations are baked into the system.
Playing back at the intended speed and making sure the phono stage has at least the RIAA recommended high pass filter (I favour a stronger one) is the way to get the most accurate bass from a record, though some people like the "loadsa bass" effect they get from the resonance and come up with all sorts of reasons why a high pass filter sounds worse when it is actually simply removing non-music related boom!
If the arm/cartridge resonance is at 10Hz, say, and the filter is at 20Hz to remove the resonant region, that will limit the accurate transcription of bass to a lowest of 40Hz if you play back at 16 ⅔ and correct pitch in software.
This is just a fact of the physics of a record player you are stuck with I'm afraid.

 

[metaleonid]

I am Frank, not Scott PMFJI but record players was my job almost 50 year ago now and the physics of how cartridges work hasn't changed.

It will sound bassier because you are listening to more of the cartridge body bouncing on its suspension - not because you are more accurately transcribing the bass from the groove.

It can not be corrected afterwards because it is a physical effect which depends on the laws of physics, using this type of transducer is inherent to record players and its limitations are baked into the system.
Playing back at the intended speed and making sure the phono stage has at least the RIAA recommended high pass filter (I favour a stronger one) is the way to get the most accurate bass from a record, though some people like the "loadsa bass" effect they get from the resonance and come up with all sorts of reasons why a high pass filter sounds worse when it is actually simply removing non-music related boom!
If the arm/cartridge resonance is at 10Hz, say, and the filter is at 20Hz to remove the resonant region, that will limit the accurate transcription of bass to a lowest of 40Hz if you play back at 16 ⅔ and correct pitch in software.
This is just a fact of the physics of a record player you are stuck with I'm afraid.

Hi Frank,
Sorry didn’t pay attention to the name of the subsequent responder.
That’s not quite how I conduct the process of transferring. I digitize the signal to the computer at speed 22.5 RPM. Or let’s assume 16 and 2/3 RPM at 48kHz sampling rate using microphone preamp. No RIAA is involved. Then in the sound editor I change sampling rate to 96kHz. Having done this I already have the WAV file with the correct speed to work with. Then I simply use ClickRepair Equalizer software program to apply RIAA curve. There are some settings like Low-cut from 25 Hz to 75 Hz at 18 dB and high cut from 1.8 Hz to 12 Hz at either 6 dB or 12 dB. But I don’t use any of these. I set them to off.

 

[scott wurcer]

To simplify what is being said, the cartridge/tone arm resonance is always at say 10Hz no matter what speed you record at so when you use software to correct the relative stylus/groove speed this resonance's effects are seen at a higher frequency. It does not matter how you process the recording. Since you say you are happy with the resulting bass I guess for you this is fine.

 

[Frank Dernie]

Hi Frank,
Sorry didn’t pay attention to the name of the subsequent responder.
That’s not quite how I conduct the process of transferring. I digitize the signal to the computer at speed 22.5 RPM. Or let’s assume 16 and 2/3 RPM at 48kHz sampling rate using microphone preamp. No RIAA is involved. Then in the sound editor I change sampling rate to 96kHz. Having done this I already have the WAV file with the correct speed to work with. Then I simply use ClickRepair Equalizer software program to apply RIAA curve. There are some settings like Low-cut from 25 Hz to 75 Hz at 18 dB and high cut from 1.8 Hz to 12 Hz at either 6 dB or 12 dB. But I don’t use any of these. I set them to off.

I see. In that case the bass is "wrong" or at least inaccurate, but if you like it that is fine, the faux bass is very much part of what some people enjoy about LPs anyway, it is just "faux" up to a higher frequency the way you are recording it.
Personally i would stick at doing the digitisation with the record running at its correct speed and doing RIAA in the digital realm, but using the 1976 RIAA curve which gets rid of a lot of this faux signal.
In fact that is what I actually do myself

 

[metaleonid]

The whole point of using slow speed method recording is to get rid of IGD (inner groove distortions)/mistracking which sometimes appear towards the center on the records that are pressed/recorded with louder volume. If the record is recorded with normal/quiet volume, I use the traditional method digitizing it using analog RIAA phono stage at normal speed. So for slow-speed method there's a tradeoff: slightly exaggerated bass, slightly less HF and more noise which is a bit audible during quiet parts. However, the pluses are that most of IGD are completely gone (with exceptions if it is just part of recording). Another plus is that if the record is worn out, then about 80% of pops and crackles are gone. At the end if it is possible to fix HF and bass in equalization, that would be great, if not, I can live with that. Another thing is that post processing after the recording is rather a lengthy process. So I was looking to see if it were possible to use analog RIAA with frequency shift - it would save me an extra step to apply RIAA digitally.

I looked at the pre-set curves in ClickRepair equalizer sofrware and didn't find 1976 RIAA curve. Or maybe it's under a different name? Besides RIAA there are: ffrr LP, EMI LP, NAB, CCR. And then bunch of curves for 78 RPMs which are irrelevant for my use-case.

 

[Frank Dernie]

So for slow-speed method there's a tradeoff: slightly exaggerated bass, slightly less HF

The bass isn't exaggerated, it is wrongly picked up. There is nothing to do about that since the frequency it happens at can not be compensated. The bottom octave will be mainly the flank of a resonance overlaid on the signal.

In 1976 the RIAA curve was modified to reduce the amount of the signal produced by the cartridge mass bouncing on the suspension (rather than from the groove) getting to the amp and speakers.
It is not always implemented because there is a (ill conceived scientifically) belief amongst some LP fans because of its phase shift, rather forgetting that there is a 180 degree phase shift due to resonance in this frequency band anyway despite it having been well understood for many decades.
So the RIAA in your software could either be the old 1950s one which exaggerates the defect or the 1974 one which reduces it.

 

[metaleonid]

Do you mind if I share samples of the same fragment done at normal speed using traditional analog phono stage and slow speed method?

 

[Frank Dernie]

Do you mind if I share samples of the same fragment done at normal speed using traditional analog phono stage and slow speed method?

Are you asking me?

There will be differences I know the reason for and others I don't, such as the accuracy of the RIAA in your analogue phono stage and any difference in loading effects of the cartridge.
I am not sure what type of sample would cast more light on the situation than we have now.

In the end the whole LP system is, and was, a (good IMO) set of compromises to produce relatively inexpensive copies of music recordings for the mass audience.
Its shortcomings allow significant latitude for experimentation and tuning to taste, but the balance of compromises is just that, a generally well thought through balance. Reducing one shortcoming may well spotlight one of the others, which I think is what is happening here.

In the end it is whether you enjoy the results you are getting that counts, the whole LP thing is limited in ways that have been known for decades (they were all documented when I was working in the engineering of it almost 50 years ago). I can point out engineering facts but not what you enjoy.

Personally I enjoy my LPs as they are and no longer attempt to rip them to digital. The copies were fine IMO but I decided the work involved wasn't worth it for me so I stopped.

 

[metaleonid]

Either you or whoever else is interested. I'd be curious to get the second opinion from the real samples how far the sound has deviated from what it is supposed to be. Then I was also thinking if it were a good idea to get a low compliance cartridge with compliance between 6 and 7 Hz so that when done at lower speed, the target will be between 9 and 10.5 Hz. Any thoughts?

 

[Frank Dernie]

Either you or whoever else is interested. I'd be curious to get the second opinion from the real samples how far the sound has deviated from what it is supposed to be. Then I was also thinking if it were a good idea to get a low compliance cartridge with compliance between 6 and 7 Hz so that when done at lower speed, the target will be between 9 and 10.5 Hz. Any thoughts?

The problem is knowing "what it is supposed to be". We can be sure that the bandwidth limit of bass is raised doing what you are doing, but nothing else.
You would need a higher compliance to lower the resonant frequency but yes, since a seismic transducer becomes accurate around 2x the natural frequency, lowering the resonant frequency also lowers that point. It may/will be more sensitive to warps for normal use though.

 

[metaleonid]

Assuming the reference is a normal speed transfer with analog phono stage, then how far the sound from the slow speed method deviated from the reference.

Actually, AFAIK my current resonance is somewhere at around 9.5 Hz. Since I am only using 22.5 RPM (the ratio is then a bit less than 1.5x), going to around 14 Hz is not that critical. If 14 Hz were at normal speed, it would result in mistracking. But in my case, mistracking will never occur because the speed is slower.

 

[Frank Dernie]

Assuming the reference is a normal speed transfer with analog phono stage, then how far the sound from the slow speed method deviated from the reference.

Actually, AFAIK my current resonance is somewhere at around 9.5 Hz. Since I am only using 22.5 RPM (the ratio is then a bit less than 1.5x), going to around 14 Hz is not that critical. If 14 Hz were at normal speed, it would result in mistracking. But in my case, mistracking will never occur because the speed is slower.

If the resonance is 9.5Hz the output from the cartridge begins to be a reasonable transduction of the groove from around 19Hz (depending on damping) so the lower limit of accurate bass in your system using 1.5x speed is around 30Hz.
The rest of the bandwidth may sound different because of the difference between your analogue phono stage's RIAA circuit and the (probably more accurate) digital implementation.

I only play LPs of music I haven't got on CD myself, and never for SQ only to enjoy the music.

 

[metaleonid]

The software RIAA done at normal speed doesn't differ much from analog RIAA. If there's a difference, it's very subtle and barely audible. I also have samples of software RIAA using 16, 22 and 33.