Equalization via software vs hardware

[Jon]

Since DCartMillenium offers numerous different equalization possibilities, is there some reason or advantage to doing equalization with hardware instead of the convenience of software? Does the software introduce distortion or other sampling "artifacts"?

[Rick]

Re: Equalization via software vs hardware

Jon
Once a signal has been digitized, I do not think there is any reason to use an external hardware equalizer.

On the other hand, I would apply whatever equalization you need to the analog source before it gets into you PC. Be it Vinyl, 78 or Tape; you should try to use the proper equalization on the recording before it gets into you PC.

This is usually accomplished by the phono or tape preamp. In the case of 78's a specal preamp is usually needed to get the proper EQ.

[Craig Maier]

Re: Equalization via software vs hardware

However, if you are transfering 78's and if a proper 78 pre-amp is not available, you can use a standard RIAA pre amplifier to make the transfer and reverse that curve and re-apply the correct 78 turnover characteristic using the appropriate presets found in the paragraphic equalizer.

[Jon]

Re: Equalization via software vs hardware

Hi Craig, Thanks for the response. Do you know of any advantages to doing all of the equalization with hardware as opposed to software?
Jon

[Craig Maier]

Re: Equalization via software vs hardware

I have my own personal opinion on this topic. Here is my logic. You can determine whether or not you subscribe to it. Theoretically, I believe that performing the entire process in software has advantages compared to performing any processing whatsoever in hardware. I must add the caveat that one should perform their transers with 24 bit resolution in order to have the appropriate head room for these processes. After performing the digital equalization processing using the software, then you can convert back to 16 bit and your results should be superior to analog based approaches.

Here are my reasons for believing this to be so:

1. Analog hardware is prone to frequency domain errors due to component tolerences. Even though one commonly finds 1 % resistors in good analog audio circuits, most capacitors are no better than +/- 2 % tolerence. Also, these capacitors are not electrically "perfect" in that they exhibit DA, ESR, leakage, microphonic susceptability, tempco, and electrical non-linearities The worst case time constant error will be +/- 3 percent in that circumstance. This extends to +/- 6 percent worst case between channels. Software is not prone to anywhere near these levels of errors in terms of the location of the poles and zeros required to produce an accurate and reproducable equalization curve.

2. Analog (Op Amp, transistor or tube based) equalizers all introduce distortion and noise into the signal pathway. Software does this to an insignificant level by comparison (quantization errors). If you prefer to have a little noise and distortion in the final product, you can always add it back in later using the Diamond Cut software.

The problem with this digital approach is that you need to construct a flat amplifier having the appropriate gain and termination impedance for your phono cartridge in order to interface it with your sound card. This is very easily accomplished if you have some knowledge of high performance / low noise analog circuit design. The alternative is to use a microphone preamplifier having an input termination resistance of around 50 KOhms. If you can not obtain one with that input characteristic, consider modifying an existing one having a difference input impedance. Please refer to the Diamond Cut Users Manual for some more details on this subject.

This is just my own personal opinion and does not necessarily represent the opinion of Diamond Cut Productions, Inc.

[Jon]

Re: Equalization via software vs hardware

Craig, Thank you for the fine detailed answer to my question. That is the kind of information I was looking for. Being a retired electronic engineer with a lot of analog and digital circuit design experience, I certainly appreciate and agree with your analysis of analog circuit performance. The part that I really didn't understand was what the software is doing. I know some of the shortcomings of digital filtering, FFTs etc.; but, this is a very complex and wide field and of course exactly how the software functions is most likely proprietary. So, I was unsure of how much or if any "distortion" would be introduced by the digitizing and filtering processes in software. Certainly, if digital quantitization errors are the largest source of software error, then I would subscribe totally to your opinion.

Jon

[Craig Maier]

Re: Equalization via software vs hardware

Jon,

The equalizer systems (Paragraphic, 10 and 20 band Graphic, Low Pass, High Pass and Bandpass) in the Diamond Cut software are IIR based, so you do not have to worry about FFT and IFT related distortions and / or artifacts. The 32,000 band spectral filter found in the Live / Forensics version is, however, FIR based.

[Doug]

Re: Equalization ??

My thoughts about recording "Flat" follow Craig - I've had no trouble to date. However, while on the topic of EQ's, could someone shed some light on why a person would play a 78 record using a preamplifier with a constant amplitude ("flat") or constant velocity ("-6dB per octave") EQ?
[img]/ubbthreads/images/graemlins/confused.gif[/img]

The Preamplifier and feedback circuits under question can be found at the following locations:
Preamp Circuit
Feedback Circuit

[Craig Maier]

Re: Equalization ??

Doug,

I do not completely understand your question. But, I can say that sometimes you want to play a 78 with no EQ if that recording was acoustically mastered. Those recordings were not "EQed" during the mastering proces and do not need to be reverse EQed meaning that they were meant to be played flat. Some people play these records with a pre-amplifier having a slope because that is all that they own. That needs to be reversed so that the transfer sounds natural. Some people record all material "flat" and then use the software to add the correct turnover and rolloff later with Diamond Cut. Does this help?

[Rick Barber]

Re: Equalization ??

I sort of remember that there was a list somewhere of what EQ's to apply to a flat recording of various manufacturers of disks. IE Victor, Columbia, blue bird and so on. Did I see that list here on the forum? [img]/ubbthreads/images/graemlins/confused.gif[/img]

[Craig Maier]

Re: Equalization ??

Rick,

I think that you probably saw that chart or listing in your Diamond Cut Users manual. I am certain that it exists in the DC5 version in the Appendix section, but I believe that it may also be in the printable Millennium version.

[Rick Barber]

Re: Equalization ??

Thanks Craig. I'll chec,k it out.

[Doug]

Re: Equalization ??

Sorry Craig, a little hard to explain,

The last two Eq's for the preamp the person designed (a)altered the amplitude 6dB for every octave and (b) to keep the sound wave velocity constant..

"Flat" for their preamp meant that even though signal power increased with frequency, the preamp rolled the signal amplitude off 6dB per octave to keep the amplitude of the signal at each frequency constant. - I thought of it as a type of signal compression?

The other EQ for constant signal velocity - well, I have no idea why

[Craig Maier]

Re: Equalization ??

Ahhh - maybe a preamp with the rolloff as you describe is simply an integrator. Integrators have that rolloff (-6dB/Octave). A Magnetic Phono cartridge outputs an "open circuit voltage" having the following relationship (to express this mathematically correctly, I would need a Greek Character set, so I will spell it out):

V = N (A) dphi / dt

wherein N is the number of turns on the cartridge coil, A is the loop area circumscribed by one turn, and dphi / dt is the time derivitave of Magnetic Flux Density.

A pre-amp Integrator could merely be to compensate for the natural rising output voltage of the phono cartridge with rising frequency (dv / dt being proportional to dphi / dt). But, proper phono cartridge termination impedance (Rl and Cl) should also accomplish the same result. There is really no need to get that fancy about it if the cartridge is properly terminated.

However, upon quick study of the circuit that you pointed me to, it looks like a pretty standard feedback equalization pre-amplifier. I do not see a "dominent pole" representing the -6 dB / Octave slope across the entire audio band for all eq settings other than its input termination network. And, the schematic shows the proper termination impedance of around 47 K (it has 2 x 100K in // which is close enough). The 5 pF miller capacitance in conjuction with the 15 K load resistance on one of the voltage gain stages is too short of a time constant to perform the integration function in the audio portion of the spectrum. That cap is probably there to assure stability of the feedback circuit at very high frequencies (above 3 MHz). I noticed a 0.082 uF cap which would act as a partial integrator as one of the selectable EQ's (actually it is a pole zero pair since it is in parallel with 27 KOhms if I remember the overall circuit correctly), but that would be redundant upon the input termination network functionality. So, I am not sure why that preamp circuit has that so called "constant velocity" feature. Were some records recorded at "constant velocity" across the entire audio band? I was not aware of that if that is the case.

[Jon]

Re: Equalization via software vs hardware

Craig, Thanks for the clarification. That puts a different "light" on things. I think I will go for a flat preamp in my improvement effort and do all of the EQ or unEQ in software.
Jon