Equalization via software vs hardware

[Jon]

Re: Equalization ??

Doug, I looked at the links in your post and probed around that web site some. I couldn't find the references to constant signal velocity type equalization. I wonder where you found that?
I am not an expert on audio recording so this is just my opinion but here is what I get out of the discussion on that site. The early simple cutting heads were driven by an open loop amplifier and the velocity (in the axis of signal motion) was proportional to the amplitude of the signal. This means that for a low frequency signal (say 200Hz) the actual distance of the cutting head travel would be greater than it would be for a higher frequency signal (say 500Hz) even though the two signals had the same amplitude. This means that as the frequency of a recorded tone continues to get higher, the physical excursion of the cutting head gets smaller and eventually gets so small that it approaches the size of imperfections, granularity of the media and also of dust and impurities that may be picked up after the cutting. To get around this limitation of high frequency cuttings being down in the noise level, an electronic boost was given to the signal prior to applying to the cutting amplifier. Also to limit the low frequency excursion from cutting into the adjacent groove, an electronic attenuation was applied to the bass signal. The end result of all of this, if you look at it from the stand point of the size of the signal excursions on the groove, the best you could hope for would be constant amplitude excursions of the cutting head at any audio frequency for constant level input signals. This is referred to on some articles as "constant amplitude recording". To achieve this you have to attenuate the bass and boost the highs prior to applying the signal to the cutting head amp. If you play such a recording with a perfectly flat playing system, you won't hear much bass (because it was attenuated) and the highs would be overly loud because they have been boosted by the equalization curve to keep them from becoming so small that they would be overwelmed by the noise created by the miroscopic bumps or dirt in the groove of the recording. To account for this on playback, you need to boost the bass and roll off the highs above a certain frequency at 6DB per octave. After you do this you should have achieved a "flat" or true reproduction of the audio signal that was originally picked up by the recording microphone. I think that this really what was meant by the terms constant amplitude and flat. Maybe this helps? Maybe not.

John

[Craig Maier]

Re: Equalization ??

Jon

You have to look at the electrical schematic of the feedback network to discover the constant velocity circuit. But, overall, your dissertation of the matter is a good summary. The lower corner frequency is referred to as the "turnover frequency" (the frequency used to keep the bass from over-displacing the cutting stylus) and the upper corner frequency is called the "rolloff frequency" to attenuate the overly amplified top end of the spectrum for S/N reasons. This "rolloff frequency" is sometimes referred to as the "pre-emphasis frequency" from the cutting head amplifiers perspective. Of course, recording vs. playback are compliments of each other meaning the curves are equal but opposite. A negative going slope during the recording process is corrected with a positive going slope during playback and visa versa. All of the attenuation values vs. frequency that were in common usage for turnover and rolloff can be found in the Diamond Cut Users Manual. There are also a number of presets that can be found under the Paragraphic Equalizer that may be of interest.

[Doug]

Re: Equalization

Yes, Thanks John & Craig...I think I was reading too much into it. I understand that the Equalisations are used to try and obtain the original signals, I just found the last two EQ's that they titled constant amplitude and constant velocity confusing (both had me stumped). - still a little confused on the constant velocity but I do have a better understanding of what it may be (I think)... At least I've got a grasp of the constant amplitude - Thanks John

Just another question on the topic of EQ's, did any of the manufacturers use slopes other than 6dB/8ve ?? AND, if so, how important is it that you use the correct slope (say 6dB/8ve compared to 9dB/8ve - if possible?)

Thanks again for being so informative [img]/ubbthreads/images/graemlins/wink.gif[/img]

[Craig Maier]

Re: Equalization

I have never seen any slopes used in phonographic or tape eq circuits that exhibit greater than 6dB / Octave anywhere on the audio spectrum. I have seen some circuits that exhibit gentler slopes than that through the use of pole zero pairs located closely together in the frequency domain, however. I believe that if you look at the RIAA curve and take the average slope over the entire audio spectrum, it would calculate out to be about 3 dB / Octave. You can see such a curve in the Diamond Cut Paragraphic EQ by bringing up the appropriate preset.

[Jon]

Re: Equalization ??

Craig, My problem is that I don't know the definition of a constant velocity circuit. I don't have any recollection of such a beast and can't find it in any of my references. I did find some references to "velocity constant" having to do with control system theory; but, I don't believe that would be intended in this context.
I'll offer a guess as to what is meant by constant velocity: This is what you get if you don't apply frequency dependent correction to the record head drive amp. If no frequency dependant boost or cut is applied to the drive head amplifier then it will move at the same velocity per volt of drive signal regardless of the frequency. (Within limitations of the power of the drive amplifier of course). In other words the velocity per volt is constant across the entire band. With the typical equalization (frequency dependent, gain alterations) that occurred later as the recording technology improved we were achieving more or less a constant amplitude per volt across the band instead of constant velocity per volt. This is the only sense that I can make of this. Does anyone have any better information?

Jon

[Jon]

Re: Equalization

Craig, Just did a little charting on the RIAA curves. It is interesting to note that from 20Hz to 20Khz which is 3 decades, the decible values are 19.2dB and minus 19.5 (scaling from my chart) which is a 38.7dB change over 3 decades or approximately 10dB/decade which is 3dB per octave as you stated for the average over the whole audio spectrum. However, the portion of the curve above about 3KHz is closer to 6dB/octave.
Jon

[Craig Maier]

Re: Equalization

Jon,

Constant Amplitude or Constant Velocity (or potentially even Constant Acceleration), all do relate to control loop theory. That is where this terminology comes from. In a feedback system, constant amplitude simply means that the loop closes on the amplitude of the signal. The "error amplifier" compares the reference signal amplitude to the actual output amplitude (or displacement in terms of a closed loop cutting head) and drives the power amplifier until the error amplifier is satisfied. Amplitude (or displacement) is held constant independent of outside variables such as friction, load, or reactance. In a constant Velocity system, that paramater (dv/dt or ds/dt) is held constant independent of all of the outside (influencing) variables. I have never seen a constant acceleration system employed in audio, but have seen it used a lot in industrial control systems and robotics.

For example, lets look at the RIAA curve. Below 500 Hz (referred to as the "Turnover Frequency") the system operates in Constant Amplitude mode. That reduces the potential for overdisplacement of the cutting stylus on loud bass notes. Between 500 Hz and 2120 Hz the system operates in Constant Velocity mode. Above 2120 Hz (referred to as the "Rolloff Frequency"), the system reverts back to Constant Amplitude mode in order to allow for improvements in the S/N at high frequencies.

[Craig Maier]

Re: Equalization

Here is an RIAA Decoding Table for reference purposes:


Frequency / / Relative Amplitude (dB)
20 Hz / / + 19.3
30 Hz / / + 18.6
40 Hz / / + 17.8
50 Hz / / + 17.0
60 Hz / / + 16.1
80 Hz / / + 14.5
100 Hz / / + 13.1
150 Hz / / + 10.3
200 Hz / / + 8.2
300 Hz / / + 5.5
400 Hz / / + 3.8
500 Hz / / + 2.6
800 Hz / / +0.7
1,000 Hz / / 0.0
1,500 Hz / / - 1.4
2,000 Hz / / - 2.6
3,000 Hz / / - 4.8
4,000 Hz / / - 6.6
5,000 Hz / / - 8.2
6,000 Hz / / - 9.6
8,000 Hz / / - 11.9
10,000 Hz / / - 13.7
15,000 Hz / / - 17.2
20,000 Hz / / - 19.6

[Jon]

Re: Equalization

Craig,
Thanks again for all of the information. Based on our discussions, I think I would like to go with a flat pre-amp in my next system upgrade. However, looking around, I don't see any on the market. I may have to build one. It has been a while since I have done any circuit design as I have been retired about 7 years now. It will be fun to see what components ares still available 7 years later; with the rate at which technology changes I may not be able to obtain the same op amps or other ICs that were available then. At any rate I am not ready to embark on any construction project right now as there are too many other irons in the fire.
Thanks again.
Jon

[Craig Maier]

Re: Equalization

I would recommend modifying the input terminating circuit of a stereo mike pre-amplifier. It has the right gain and is "flat." That would be the easiest way to go. BTW - all the olde circuits still work, even if the op amps are not available. Pinouts are still the same on the newer types. They just have lower noise floors, higher slew rates, better gain bandwidth products and lower power consumption. Just be sure to use "internally compensated" types unless you are familiar with gain crossover loop compensation techniques for second order op amps.

[tomd]

Re: Equalization

This thread is kind of confusing. Quoting Craig's post #3292:

"Below 500 Hz (referred to as the "Turnover Frequency") the system operates in Constant Amplitude mode. That reduces the potential for overdisplacement of the cutting stylus on loud bass notes. Between 500 Hz and 2120 Hz the system operates in Constant Amplitude mode. Above 2120 Hz (referred to as the "Rolloff Frequency"), the system reverts back to Constant Amplitude mode in order to allow for improvements in the S/N at high frequencies."

Except for a typo--he meant Constant Velocity between 500 Hz and 2120 Hz--this agrees with what I read elsewhere. "Constant Amplitude" or "Constant Velocity" refers to the recording characteristic of the disk cutter. With an audio input to the cutting equipment whose amplitude is constant at all frequencies, the cutting stylus motion will be either constant in amplitude (excursion) at all frequencies, or will move with constant velocity at all frequencies (which means that its excursion or amplitude has to be made less at higher frequencies because the motion is faster).

The electrical output of a magnetic cartridge when playing a record is proportional to the velocity of stylus motion, as shown in Craig's post #3279. During the constant-amplitude portions of the RIAA cutting characteristic at the low- and high-frequency ends, the playback cartridge's output rises with frequency compared to the original audio signal, and the RIAA playback EQ falls with frequency to make the overall record-playback response flat. During the constant-velocity portion at the mid-frequency part of the RIAA characteristic, the RIAA playback "EQ" is flat (no EQ really, no slope). The velocity-proportional response of the playback cartridge itself makes the 500-2120 Hz constant-velocity range come out flat. (The cartridge output is proportional to velocity and the velocity is constant with changing frequency, so the output is constant with changing frequency, i.e., record-playback response is flat.)

This isn't consistent with the idea in post #3279 that the cartridge termination in the preamp compensates for the cartridge's velocity-proportional response. It is the EQ that does this. The RIAA EQ converts the velocity-proportional signal to one with a flat frequency response at the ends of the audio band, and uses the velocity response as-is in the midband. I believe that the cartridge termination has to do with cartridge resonances, not with the basic voltage-generating mechanism.

In the R. F. Wilmut switchable-EQ preamp circuit linked from Doug's message #3269, the "constant amplitude" position 11, whose frequency response is flat in the audio range, will give flat frequency response for a constant-velocity recording played with a magnetic cartridge. The "constant velocity" position 12, which rolls off at 6dB/oct as frequency rises, will give flat frequency response for a constant-amplitude recording played with a magnetic cartridge. (The -3dB point is 72 Hz, maybe for some rumble filtering). The switch positions are misnamed. The C-A position 11 simulates electrically the overall audio-to-groove amplitude response of a C-A *cutting* process, while the C-V position 12 simulates electrically the overall audio-to-groove amplitude response of a C-V cutting process. When applied to *playback* with a magnetic cartridge, the names should be reversed: position 11 to play constant-velocity records and 12 to play constant-amplitude ones.

Perhaps there aren't any constant-amplitude phono records. I would have guessed that acoustic 78's were, theoretically, though in practice they weren't constantly anything, but at this link:

http://www.kabusa.com/eqs12rvw.htm

Gary Galo says that acoustic 78's are approximately constant-velocity. This would explain why he and others say that acoustic 78's should be played with flat "EQ"--which is no EQ--for a magnetic (velocity-responding) cartridge.

[Craig Maier]

Re: Equalization

Be real careful when you cite Gary Galo as a reference in the field of audio. Gary Galo has a degree in teaching and teaches English in High School from what he has told me. He pretends to be the world renowned expert in Audio Engineering and insists that such things as Tube Preamplifiers are the only acceptable way to do that job. Hmmmm. Doesn't something smell here. Tubes produce considerable levels of distortion. Anyone who is technically inclined is aware of that. If you like distortion, fine. The reasonable approach is to record without distortion. If you want distortion, you can add it later with the VVA. But do not say, Gary, that tubes are the way to produce an amplification system given todays op amp technology. I will take 0.001% THD over 1 % THD any day of the week. Most people who understand Electrical Engineering appreciate that concept. I have spoken with Gary on the phone about op amp amplification circuits that I have designed. He denounces them as completely out of hand and has promised to provide bad reviews if they are ever put into production. I have come to realize that he knows absolutely nothing about Engineering and is exceptionally good at owning an opinion, but not being able to defend it. Sorry, Gary, but that is my take on your self professed expertise. If you want to become a real expert in audio engineering, consider going out and obtaining a real Electrical Engineering Degree as Rick and I have done. As for Magnetic Phone Cartridges, yes I did make a typo. And it was correctly identified and I have corrected it. Thanks for the heads up. However, Magnetic Phono Cartridges produce a linear frequency response output iff (if and only if) it is properly terminated. Please refer to any magnetic cartridge manufacturers output voltage curve vs. Frequency. Unsimilarly, Ceramic phono Cartridges (or other forms of Piezoelectric types) do not produce a linear output. Their output is not unsimilar to the RIAA curve only inverse. That is why one does not see an RIAA eq network on the input side of a pre-amplifier designed for any type of piezoelectric cartridge.

[Doug]

Re: Equalization

Not meaning to start the argument but Craig, I totally agree with your concep about valve preamps/amplifiers. If one is trying to best reproduce the recorded materials, then you would want to do it via a method introducing as little distortion as possible. Ie, not via a "out dated" method - the only method that was available at the time. Although I am not expert in audio engineering, nor do I have an Electrical Engineering Degree, I have discussed the theory with my father (an experienced audio engineer) many times with the same conclusion. Besides, I believe you can mimic a valve preamp/amp using transistors/op-amps anyway (just include some design faults and incorrect EQs into the circuit! - just kidding [img]/ubbthreads/images/graemlins/tongue.gif[/img])

[Craig Maier]

Re: Equalization

Doug,

Just use a single JFET as your Voltage amplifier stage and use no local feedback around it (omit the source resistor). You will get "tube sound" with about 2 percent Harmonic Distortion using this technique with a square law transfer characteristic. Remember, you can also simulate tube distortion using the Diamond Cut VVA with a great degree of versatility, which contains the actual transfer characteristics of a number of commonly used audio Triodes and Pentodes. If you want to construct a compound tube based pre amplifier and power amp, configure a number of VVA's in cascade using the multifilter function.

[Doug]

Re: Equalization

You know what's funny Craig? - my father said something similar to what you said:
[ QUOTE ]
I will take 0.001% THD over 1 % THD any day of the week. Most people who understand Electrical Engineering appreciate that concept.

[/ QUOTE ]

However, he took the approach of "I wouldn't bother designing a valve amplifier, especially with the low noise/high output transistors and Op-amps around today." Then he tells me how all the Op-amps of today are composed from the best of the transistor amplifier circuits people come up with (both back then and now). Many of the Op-amp designs (Ne5534 for example) were originally larger transistor circuits that were of such quality or flexibility; they packaged them into an Op-amp configuration. I understand that this allows greater design flexibility and sets a worst case or standard S/N ratio that can be obtained. Nothing wrong with that...

Also, at least when something goes wrong, you don't send 600V-1kV into your prized audio equipment - nothing can withstand that! (Including the person who goes to turn it off) [img]/ubbthreads/images/graemlins/crazy.gif[/img]

But while on the topics of EQ and Preamplifiers:
If Op-amps can operate over a wide voltage range of +/- 3v up to +/- 18V (in some cases), why is the most commonly used rail voltage +/- 12 or 15. Why not use +/- 5 ???

I've been told; most Op-amps will operate correctly on a lower supply rail without compromising S/N.
Also, what happens in the computer, do the buffer stages use +/-5 or +/- 12-15???