active AD/DA xover as RIAA-amp?

miksin

Member
2005-04-02 6:58 pm
Has someone ever tried to use active x-over as RIAA equalizer?
I use one such,( running active two-way system) namely Alto Maxidrive , and of course there is out plenty of them, all Behringers and DBX's and so on, but the question is: most such devices has easily enough EQ power to make RIAA curve. One could possibly need line pre or step-up tran between a xover and a cartridge, but is there some window to aim, to make A/D converter to work in 16-bit or perhaps even 24bit and not yet saturating ADchips with RIAA.s uncorrected tr3eble emphasis? complicated, ehe? Yes , I think so but I also think that could be something to at least to try, namely, if you already has one such device (Active xover, that is if you are usin active multiway system with such device) so why not to try to skip separate RIAA preamp with it, if possible., or replace it with just line pre. Help with maths please? How much one could boost carts signal before ad-chip begins to clip, and does it seems possibe to achieve reasonable line level takinf into account chips sample rate and dynamics? ...


Mika
 

EC8010

diyAudio Moderator Emeritus
2003-01-18 7:57 am
Near London. UK
I'm sorry to rain on your parade, but RIAA is best done analogue. If you do it digitally, you immediately use up 40dB of digital dynamic range. Given that the A/D chips have only about 106db of dynamic range (and that's a weighted figure), that only leaves 65dB. Further, by applying an unequalised signal to the A/D, you are stressing it just where it's weakest - its HF performance. And what about leaving some overload margin etc?
 

miksin

Member
2005-04-02 6:58 pm
So it does not matter whether the xover is fed with -60 or +6 dbU, the converter anyway uses its entire dynamic range to input signal, right?
And all eq-ing-lets say for example -12 db automatically reduces dynamic range with same number?
If you feel like I dont really know what I am talking about you propably just are right! Sorry if askin stupid questions, that is only because I liked to learn some more :) And thanks for "raining on my parade", that was exactly what I was askin for!
 

EC8010

diyAudio Moderator Emeritus
2003-01-18 7:57 am
Near London. UK
miksin said:
So it does not matter whether the xover is fed with -60 or +6 dbU, the converter anyway uses its entire dynamic range to input signal, right?

It does matter. I'm afraid I was assuming very best case. An A/D needs to be fed with a signal large enough to almost overload it, but no more. That's why many people who use digital crossovers use stepped attenuators on the output of the crossover to feed their valve amplifiers etc. That way, the crossover always sees a large signal.
 

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
If you must do it the hard way:

1) Put about 10:1 of gain in front. That will put the hot 5KHz-20KHz tones around the A/D's nominal input level.

or

2) Use an integrator in front. The main slope in the RIAA curve is really due to the differentiator action of the dynamic (velocity) pick-up. If you used a displacement pickup, the RIAA curve would be nearly flat, a 6dB shelf from 500Hz to 2KHz. Integrator with gain of about 100 at 1KHz would be about right.

Note that the standard integrator has low input impedance and excess voltage noise. You need to complicate it. Pretty soon it is more complex than a standard analog RIAA amp.

Since all the slopes in a phono recording and pickup are simple analog effects, and well standardized, I'm not sure what you'd gain with exotic digital EQ. While a phono needle could have a few degrees excess phase at 20KHz, they do not have the huge internal resonances of cone speakers, and don't need nearly as much EQ complexity.

I could picture using digi-EQ in 78 transcription where every company used a different EQ. Still I suspect it would be best to start with a semi-universal analog preamp, and then use the digi-EQ to give the variations between different recording curves. But good 78 transcription is such an Art, that each person finds his/her own preferred techniques.
 

miksin

Member
2005-04-02 6:58 pm
I could picture using digi-EQ in 78 transcription where every company used a different EQ.

that's exactly one of the things I had in my mind too.. I am spinning an TD124 :)

Since all the slopes in a phono recording and pickup are simple analog effects, and well standardized, I'm not sure what you'd gain with exotic digital EQ.

Well, if I already have one such device in my setupp with with one possibly could make a riaa-correction, why couldnt I at least to think or try it? you see, it took me a while to accept use of such device -digi-xo- in my place. I used to listen for four or five years only fullrange drivers and I felt almost as a sinner, you know:) So I could see some beauty in a idea to avoid one amplification\equalization device with use of another one, namely that XO, I am sure you understand that idea. Unfortunately it seems to be not a good idea.
Still, I have some RCA-6,3mm-plug adapters and because that XO accepts both xlr- and 6,3-mm inputs I will try it as soon as I next time swap carts, now I am running a DL103 and I think that is perhaps too long shot but I also occasionaly use a Benz MC Silver and high output mc could perhaps be more realistic :)
Perhaps I will report conclusions, that is if you are interested.

I am just thinkin, 65 db of dynamics is still quite respectable amount. There are quite a few recordings with larger dynamics and perhaps even less domestic systems capable producing 50-60db (lowest levels in a recording in normal listening eniviroments background noise, I think? add there 65db, you'll end around 115-120?
I think I will at least try, to see how much microdynamics and details suffers. Perhaps it sucks, but at least, I tried it :)
That 10:1 stepup tran to cart could be nice idea:) perhaps with treble equalization:)
 

miksin

Member
2005-04-02 6:58 pm
-Just got to write this more:
I tried, with DL103 putting out whoooppin 0,3mV, directly in that Alto Maxidrive 3,4.
Needless to say, I listened very quiet, -70 db compared what my RIAA (Musical Fidelity x-lps v3, that is the newest edition) gives. Also, time here is 00.40 am, but I often listen in the middle of the night at very low levels and that was exactly right level for my night listenings.
I made _very_ rude correction with input eq, putting -15db in 12khz and and +15db in 70 hz, peak width set wit ears and a handful of familiar recordings.
Unfortunately difference in to favor of that Alto was not subtle. Except that output level there was not a single thing that my RIAA makes better, and again, difference was not subtle.
Perhaps I really should set up one pre-pre with perhaps Lundahl LL9206 step up for Denon and transformer coupled e55l or somethin, perhaps intentionally miscalculating output iron to achieve some treble attenuating:)
Or perhaps it is just that Musical fidelity which sucks.
Really, has some one ever tried something similar? how do you know how does it sounds if you not try?
Perhaps propellerhead is my middle name :)
 

EC8010

diyAudio Moderator Emeritus
2003-01-18 7:57 am
Near London. UK
Bravo to you for trying the experiment.

Why would I not have bothered (even though I have a DCX2496 digital crossover)? Well, you see, when it comes to digits, some things are cut and dried. It's inescapable that applying RIAA in the digital domain uses up dynamic range. PRR's point that a magnetic cartridge is a velocity transducer and that what we call RIAA is really a combination of RIAA and the 6dB/octave response of the cartridge is very pertinent. It would make sense to use a digital equaliser to apply the myriad of 78 and mono LP settings - until you've seen how long it takes to load the settings (I use my crossover to apply NICAM de-emphasis when watching television). Perhaps your kit is different.

By the way, can you think of any good reason why this thread should not be in the "Digits" forum?
 

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
> It would make sense to use a digital equaliser to apply the myriad of 78 and mono LP settings - until you've seen how long it takes to load the settings

True. I would assume that a true fanatic would use some storage scheme. I have not worked much with digital EQ and X-over, but I know they have storage, some internally and more when used with a PC. I was cussing a System Controller (fancy crossover/EQ/limiter) and I can teach it 100 different rooms and recall them with (many-many) finger-presses, or save settings to/from PC (the PC does not have to stay attached). So after a wild week of entering factors, you look at the disk and recall the "RCA, 1923, Studio 4, plus 4KHz cut" setting, and bingo.

> has some one ever tried something similar?

Yes, though with mike-amps and universal analog equalizers instead. Pickup loading, dynamic range, and EQ flexibility are real problems. If the crowd is drunk/stoned enough, they don't care; for "good" audio, it is either not-quite-right or an awful lot of work.

> how do you know how does it sounds if you not try?

After trying a few thousand wild ideas, I've gotten some sense of what will work, what might work, and what will "work" in quotes. No doubt you can play an LP using a digital EQ as your preamp; but I don't expect it will be "better" than just using a standard phono preamp.

I was trying to suggest that there are easier ways to get good clean sound. For really good LPs, I use only a phono preamp and a power amp. If an LP needs EQ, of course "virtue" goes out the window and I will slap as much EQ on it as I feel it needs.

OH: I have, more recently, played 78s through RIAA (33 LP) preamp and used PC-based digi-EQ to balance the tone. The story is long, but the client was happy.

> Perhaps propellerhead is my middle name

I've seen propeller beanies go in and out of "style" several times. I'm sure all my non-geek friends assume I have one.
 
PRR said:
For really good LPs, I use only a phono preamp and a power amp. If an LP needs EQ....
Too often the LP has really poor sound so it doesn't matter to be too exact. A normal 10- or 31 band EQ will probably be enough and such EQ can be either in hardware or in software plus some decent soundcard.

One good example of a really bad LP is Vinegar Joe (remember with Elkie Brooks and Robert Palmer) and the album "Rock and roll gypsies". A complete failure from the recording and mixing engineer.

If you are into variable RIAA EQ you may check ELBERG MD12 MK2

http://www.vadlyd.dk/English/RIAA_and_78_RPM_preamp.html
 

miksin

Member
2005-04-02 6:58 pm
"By the way, can you think of any good reason why this thread should not be in the "Digits" forum?"

No, I can not. neither can I think of any good reason why this thread should not be in the "analog" forum, or in "everything else"?

I really was thinkin for a while where to post this, and decided to do it here because I was afraid analogue geeks would course me and my herecy and ban me immediately:) and thought also that digit geeks (Hi to every one :) )would think that this is not an interesting problem.

Perhaps subject could be " Could there be any theoretical advantages using active x-over also in RIAA-correction and thus replacing one active eq-device with simple gain stage if it is needed to achieve needed gain/dynamics level in a system which anyway is built around XO using it always?"

Also, I liked to note that I am not using that XO as a eq unit, except perhaps with this issue, but only as a preamp and a crossover.
Now this is not meant to be a commercial but: if interrested of such devices please give that Alto a try, if you find a dealer. I think it is of extremly good value for its price.
 

hatyri

Member
2004-08-20 9:50 am
Salo
Maybe PCM4202 and Fully-Differential OP (example OPA1632) give enough resolution..

PCM4202
* Dynamic Range (VIN = -60dBFS, fIN = 1kHz, A-Weighted): 118dB
* THD+N (VIN = -0.5dBFS, fIN = 1kHz): -105dB

I think that solution is something what we get also out passive riaa.
Passive have as well that correction what eat input dynamic 40dB.

I don't believe there is more that -100dB snr before riaa so that 24bit a/d enought. If we rewiring better wiring and use differential solution we maybe get even better snr what original. It is easy do good quality digital correction. Same quality analog correction is very hard to do and hard to find good components.