Riaa curve

How important is riaa accurcy in a preamp? Some state .5 db...others .25
0af4f876 eb83 4323 a292 3564f9bafea1phasecorrect
The RIAA curve for a phono preamp has a 40 dB swing from 20 Hz to 20KHz. Half of that is a 20 dB boost of the bass and the other half is a similar cut in the treble. It crosses the zero point at 1 KHz.

However, it is not a completely straight line from boost to cut.

Obviously the more accurate the curve, the closer your LPs' frequency response will be to flat when you play records.

The frequency response will also be influenced by the interaction of the phono cartridge with the preamp. One can have a preamp with a very flat response that becomes a lot less flat if mismatched to a cartridge.

It bears repeating, though, that this one parameter is not the only one that affects a phono preamp's sound quality. Distortion, noise, headroom and other factors are also very important. The ultimate test is what the user thinks of the sound when he plays a record.
to throw another variable in...

depending on the age of the vinyl, there were several variations of the RIAA curve. A few phono stages even offer switches to adjust to various labels variations on the curve.
Chosenhandle...There is only one RIAA phono equalization curve. Before this was agreed to different labels did indeed use their own propietary curves.
The choice of eq curve is stated on the record jacket, pre 1955. It does not go by year, but by maker. You can search the Internet on appropriate matchings. All the optional eq curves and RIAA are the inverse of that used when cutting the record and should be used accordingly, not for effects as some audiophiles think. I use the EMI and Columbia curve in the Boulder 2008.
the ML phono-modules in my ML326S must be really bad news, they ONLY claim +/- 1dB!

A previously owned GCPH claimed 0.25dB (+/-?)

The ML (in)boards sound better, hm.

Now take any cart MM to MC and each and every one sounds different in the same phono-modules, some actually one hell of a lot --- from to bright to balanced to dull and so forth.

I have a suspicion that even +/- 1dB is 'peanuts' compared to the much higher deviations that carts bring to the table --- YMMV.

Axelwahl...You are correct that importance of RIAA frequency accuracy is over rated. The rest of the system, notably speakers, are much less accurate.

One consideration when the the RIAA curve was adopted was that it should be easy to implement. Specs of 0.1 dB are usual and many preamps claim better. Perhaps your spec of 1 dB is a misprint.
Hi Eldartford,

>>> Perhaps your spec of 1 dB is a misprint. <<<

Maybe ML printed some figure based on reality?

I am suspicious of measured claimed performance until someone else measures and verifies it.

Looking at J.A.'s (Audiophile) RIAA measurements of some highly regarded phono-stage makes me say that.
Most, or all of them do not do any better then that, and the GCPH? They didn't bother to measure it, so how accurate it the claimed 0.25dB ? --- And I mean at least across the full 20Hz to 20kHz.

Also, looking at the SMD technique used (ML), how on earth are you going to get much better than +/- 1dB, I wouldn't know right now.
(Pre-selected components to 0.1% from a component bonding automat??) That be something new to me...
Dear Eldartford: +++++ " You are correct that importance of RIAA frequency accuracy is over rated. The rest of the system, notably speakers, are much less accurate. " +++++

yes, even some cartridges specs are less accurate and this is one way to see that RIAA accuracy.

My overall/whole point of view is different and different not only on the RIAA accuracy but in each link on the audio system chain: almost all of us ( I hope ) want to be nearest to the recording and trying to achieve this goal we need ( between other things. ) accuracy, very low noise, very low distortions, etc, etc. trying to add the less and trying to lose the less all over the audio system chain, at least this is my goal ( of course the enjoying of the music. ) maybe yours is different.

Now, what is the inverse RIAA eq. curve? well it is the main factor why phono stages are a necessity where IMHO the main target in any phono stage must be to deliver an accurate inverted RIAA curve to mimic the one used on the recording to give us a flat frequency response with the gain/low noise/low distortions that our cartridges are asking for.

I don't care what happen after or even before the phono stage link the subject is that in that phono stage things must be working/processed with accuracy in the right way.

If I was thinking like you that there are other audio links with a higher un-accuracies then that spec could be +,- 3-5 db and no body cares anymore about.

But things are not so easy, the RIAA is a curve not a flat line so any single deviation in that curve ( at any frequency ) affects not only that deviation on that discrete frequency but affect almost three octaves!!

Do you think that you can hear a 1db deviation on the RIAA curve at 3Khz?, certainly yes because that deviation affects a wide frequency range and puts a continuous coloration in the sound reproduction ( of course that are people that does not care about, well I care. ), Eldartford I don't know how good are your ears perception but I can tell you that you can ( if your system has the resolution. ) hear even deviations as low as 0.1-0.2db

If that 3Khz deviation was the only frequency that was affected then you maybe can't hear it or detectec it.

Everything the same IMHO the RIAA accuracy makes a wide difference in the quality performance on almost any audio system.

Regards and enjoy the music,
I'm sure there is more then one way to measure RIAA error to start with.
If so, we are arguing about the emperor beard.

For instance to get to a fat response you need to have an inverse RIAA network, yes?

Now that in itself needs to be MORE accurate than the one you try to establish in the phono-pre.

Due to component tolerances there will be no two stages the same exactly, even if you go through 100reds of components.

All of this just plain and simply just begs the question of the manufacturability of such high tolerances --- if ALL surrounding it is WAY-OF in comparison.

Seems to me, like insisting your tyre pressure to be 0.01% correct... and it will make actually no difference if it is correct to 1%.
I can claim 0.01% but how to measure/confirm that it's like this in the first place, and then the temperature might just have changed, what now?
Sorry about the typo's, (need an editor sooner or later :-)

emperor = emperor's; fat = flat; just begs = begs; WAY-OF = WAY-OFF
Dear Axel: IMHO your tyre example tell me that you don't understand anything what is on my post or you don't really care about and it is nothing wrong with that because you like a person are a unique person as everyone else.

Btw, in our unit we make the RIAA calibration with the unit " hot "/warm to avoide changes due to temperature. Of ourse that in any audio item design we have to work between the parts tolerances, in some way this part tolerances are the limits to.

Regards and enjoy the music,
Dear Raul,
>>> ... in some way this part tolerances are the limits to. <<<
Exactly, what was on my mind also, tires have tolerances so have measuring gauges.
Pushing the envelope is fine by me - but maybe not always quite 'reasonable'?

I listen to CD on my 390S and then listen to a +/- 1dB 'quoted'/spec. RIAA module, and all I hear is NO colouration other then what the different carts produce.
The least colouration CD vs. vinyl is provided by my Windfeld, known to be very neutral.

Then I listen to a S1000ZE/X that has colouration (more low end than any same CD), but it is nicer than CD.
I actually cannot easily assimilate how +/-0.0075dB error is going to make this any different.
Maybe if I put my head in vice?
The variations in moving my head will actually be bigger than that 1% RIAA error... hm.
Raul....Do you know the spec on the RIAA equalization when the LP is cut, and the frequency response of the cutter, its electronics, and the recorder that made the master, and the microphones?
Dear Axel: IMHO all your statements are out of my main/deep point of view, please re-read my post and try to understand what I really mean because maybe I can't explain in the right way.

Btw, I understand that you can't imagine how could you hear a tiny deviation in the RIAA when you never experienced on the same phono stage making RIAA deviation changes. Well, over the time in our design uit I have the opportunity to have that kind of experiences and that's why I know that we can hear those tiny RIAA deviations.

Regards and enjoy the music,
Dear Eldartford: I think that you miss my point somewhere because I posted ( between other things. ):

+++++ " I don't care what happen after or even before the phono stage link the subject is that in that phono stage things must be working/processed with accuracy in the right way. " +++++

IMHO I think that any audio device/item designer/manufacturer out there first than all take care what is " happening or not happening " with their designs trying that their designs can be of the best quality performance/accuracy at any level.

An amplifier designer can't design thinking that because the accuracy on the speaker/room is of only 10 db ( deviations. ) then the amplifier accuracy can be at 9-10 db on frequency deviatins, at least I don't thnk that is the way to design/build high-end audio items.

Regards and enjoy the music,
Dear Raul:
I actually think you explained your position VERY WELL, and I also believe what you say when referring to your testing and listing --- AND I also think that the better tolerance is a better thing to have, --- if it makes commercially viable sense. This is here, where we are at X purposes.

If 0.01% RIAA error HAD TO BE, you could 'trash' just about every phono-pre under sun, except the 3160 --- and this is where I beg to differ.
Simply because a GOOD non-RIAA replay unit like a CD player sounds 1:1 with a 1% error spec. phono-stage.

I think, I would first have to look very seriously at my speakers to be able to appreciate ANY difference.
Call it a question of the right priority?
Yet, I also know that the Burmester 961's have a VERY linear frequency response also, so then it'd be back to my room?
But now I also know that I have very good room lock / good diffraction / no standing waves...
OK, now we need a much bigger listening room first and I would not disagree but my pockets do :-)
And all this so I can truly appreciate a better RIAA spec?

Again, there is NO, ZERO notable difference in tonality between CD and Vinyl recordings OTHER than provided by different cartridges.
The more neutral the cart the closer in tonality to the CD, with the difference in vinyl adding some more LIFE, simply by the analogue's higher resolution of micro details, e.g. hall information, etc.

We all can agree, that the RIAA error does impact tonality but NOT resolution, (given that L/R are ~ <0.75dB out)
Hi Eldartford
I'd love to learn from the pro's perspective about that "RIAA equalization when the LP is cut", please if you would share it.
I think it should add some perspective to the subject on hand i.e. "Riaa curve" and not RIAA curve error tolerance...

Eldartford, my cutting lathe has a set of 'pre-emphasis' modules that are in fact the inverse RIAA curves- if you put their output into a phono preamp, the result should be flat.

It sounds to me that what you are getting at is that not all cutting systems are created equally, not all mics are neutral as well, that sort of thing, and you are absolutely correct.

However, you might be surprised but the cutter manufacturers were actually quite concerned about getting the pre-emphasis correct, to the point of hand-picking components to make it so. The same is true of any high end phono preamp manufacturer; after the hand-picking process we wind up with a substantial stack of unusable EQ components!

Raul is absolutely correct in saying that EQ errors of only 0.1db can be heard and he also gave the correct reason why- its a spectrum, not a single frequency. That makes a huge difference to the human ear.

In the end, this makes it possible to evaluate components used in the recording process, such as mics and recorders. Although 0.1 db seems like its not very much, it **can** be measured quite easily, and the math for generating the EQ networks is well established.

I have seen at least one manufacturer claim that due to other errors, RIAA EQ is not important. IMO such a claim is rubbish and if a manufacturer says that to you, turn around and run as fast as you can!
Hi Atmasphere
good points you make, Raul will have a field day :-)

This also means, there is no way, but NO way, to do this RIAA circuit other then by a most tedious hand selection and soldering in/out process. In the case of your cutting lathe I've no problem, after all it the one that sets the 'standard'.

But as to phono-pres it means 'bespoke' only! --- and well that just about killed 98% of all phono-pres by then.

Food for thought...
I don't know how good the cutting equalization is...that's why I asked. Atmasphere might know.

If you can hear a 0.1 (or 0.01) dB boost over the range above or below the RIAA break frequency (1000 Hz I think) who is to say that the boost is bad? It might even compensate for error of the speaker (although 0.1 dB would not be enough to do that).

Raul...Your efforts to achieve perfection are an excellent example of overdesign. Overdesign is a common fault of engineers...I know: I am one.

Remember the saying...

"In every program there comes a time to shoot the engineers and get on with the job".
Eldartford, when dealing with calibration of a cutting lathe, the speakers are not part of the equation. To answer your question to Raul though: we are, 'we' meaning the people who make this stuff to standards like this. 'We' say its bad when there is a boost or a roll-off that does not conform to the curve. Its not a matter of over-design either. You *can* find parts that a close enough in tolerance that they will do the job.
Atmasphere...What is cut in the vinyl is heard through speakers...the "equation" for what is heard most certainly includes speakers, and a lot of other stuff, and the cutter.

You say that you can find parts that are "CLOSE ENOUGH" to "DO THE JOB". Right on!! That philosophy indicates you don't overdesign.

I am still waiting to hear the frequency response of a typical cutter. and its "pre-emphasis" module.
Eldartford, the unit that I use has bandwidth out to just shy of 30K. The electronics have a bit more than that, but not much (they were designed in the late 60s). I don't think you could say that there is a 'typical' cutter, as they are all designed (like anything else in this world) to meet certain parameters that the designer felt was important.

I chose the unit I did based on hearing the LPs that it was know to have made, and also for the fact that I can get our amplifiers to drive the cutter as well- resulting in the first vacuum-tube transformerless cutter system.
Atmasphere...Ok. I guess I just said "frequency response" when I meant fidelity to RIAA equalization across the whole band. I think that was what we were talking about.

When the record is mastered equalization is usually applied so that the result sounds right to the producer, and when the record is played back equalization may be applied by the user so that it sounds right to him (through his speakers an in his room). With all this going on extreme accuracy of the RIAA equalization seems unnecessary to me.
you mentioned the cutting lathe and applied (inverse) RIAA.

I would like to know about % deviation found with well regarded 'commercial' phono-pres (not boutique audio).

Looking at the few test graph showings in audiophile’s measurements --- "With all this going on, extreme accuracy of the RIAA equalization seems unnecessary to me." --- it seems some more manufacturer sharing this opinion.

We all know that accuracy will be nice to have in deed, but is it THE prime design parameter I seem to hear it is? Aren't there other much more important ones?

In other words: will it render another product much inferior just because it does not go to the ultimate in this RIAA matter?

Again, (yes, my system sux) listening to my 390S CD and then to the same on LP, I can hear NO colouration with a +/- 1% spec. We also know there are units out there with more than 1% deviation (and not too cheap either).
Eldartford, you are right about that process, but the producer of an LP would never have the EQ of the RIAA messed with! Usually the issue is a private studio and producer have made a recording. They want an LP, so they send the tape to someone who can do that. Usually they also send a tape of test tones made on their tape machine, so the LP manufacturer can set his tape playback to match the calibration of the original tape machine. Then he makes the LP.

The producer then gets a test of the LP and listens to it. He may ask for changes in the tape playback calibration (this is something *very* different from EQ, BTW) in order to get the LP to sound more like it did when the recording was made.

If the LP is being pressed in a different country, this process is often broken. That is one reason why it is so important to get an LP that is pressed in the country where the recording was originally made!

It is a tricky process, no doubt, and full of errors, but one thing that no-one ever seems to mess with is the EQ curves. For the most part, there is a very genuine effort to get the LPs to sound like the playback of the master tape.
Atmasphere...Agreed that no one would deliberately screw with the RIAA equalization. My original point was that 0.1 dB is readily obtainable and more than adequite. 0.01 dB is a waste of effort. IMHO.
Eldartford, it might be, but in the world of high end, the difference between HIFI and the ability to sound like real music exists in the nuances. That is why there are Teflon caps, high precision resistors, care in grounding layouts, spec-ing out the gear octaves beyond human hearing, use of high purity wires, balanced operation (in our case anyway), non-ferrous chassis, *tubes* and whatnot.

We are trying to make it sound real. All designers probably have a blind spot, usually a focus on what they think is important :) so you will see a wide variation in designs as a result. It is the mark of a good designer to know what specs are 'negligible' and which ones are really important that *look* negligible. I don't think I am in a position to really judge exactly what that might be, because like everyone else, I have my blind spots too. In the case of EQ, we can spec the EQ components within 0.05% pretty easily, so I don't think hitting 0.01% should be all that much harder.
Dear Eldartford: +++++ " original point was that 0.1 dB is readily obtainable and more than adequite.... " +++++

adequite?, for whom?. Certainly your " goals " are really different that the ones in other people like me.

How do you know what is adequite and what is not if you never had the opportunity to compare about?, your statement make no sense.

IMHO the mediocrity that surround the audio industry has its origins ( between other things ) in the mediocrity of the whole audio industry goals, there is almost no attitude to be better and why to worry in be better when the customers don't ask for more.

Fortunatelly there are designers like Ralph, my self and others that still care to growing up making better designs out of the mediocrity of the whole audio industry in where the customers are the main part.

I respect your attitude in the RIAA subject, it is an easy one where you don't really cares about where you don't ask for anymore where you don't ask for a better audio items and that attitude is fine with me and fortunatelly does not affect my spirit to improve and be better.

Regards and ejoy the music,
Rauliruegas...There are more important things for you and Ralph to worry about than 0.01 dB RIAA accuracy. Neither you nor Ralph have suggested a reason why high accuracy RIAA is important when other elements of the system are so much worse with respect to frequency response. As an engineer, I know it's so much fun to see how good you can make your design that it's easy to loose sight of the real performance requirements.
Eldartford, that one is easy! We use a linear recording method, even if recording digitally. By that I mean linear as opposed to logarithmic.

The linear system requires flat frequency response. So we adhere to getting the RIAA correct in the same way that we make the linestage and amplifier absolutely flat frequency response.

IOW, its easy to get 0.1 or better flat frequency response from the AUX input, if we can do that then the phono should have the same benefit. Perhaps I should say *especially* the phono should have that benefit, since, for many, like myself, it is the primary playback source.
Funny, as I follow this % back and forth, I think of cartridge specs. - quoted cartridge specs - also in 0.2dB region say of channel balance.
Then we test this baby - ha, ~1,8dB........ and all say wow, sounded marvellous. Brinkmanship?

Who I ask, of any end-user will be the judge that these specs are actually the case?

Sorry, to say that testing of some of these claimed specs, made me highly critical or non-believing.

As the saying goes: "Paper is patient..."
Maybe I aught to mention, I'm well acquainted with:

- MIL spec requirements
- Hi-Rel component testing and selection
- Hot handling (testers)
- Burn-in of components (re.: child-mortality)

i.e. Component assembly and Testfield procedures, and MIL spec QA requirements.

It could well be the reason I have a healthy respect of what is required to achieve superiour - long term - stabilty, never mind 1ppm failure rates, 0.01% tolerances.

All of this hi-spec stuff has to come at a PRICE!
I am hearing it loud and clear, when Eldartford is weary of over-engineering, just because 'we can' and not because it is a clear requirement in the first place.
I should have mentioned that I am/was the Aerospace type of engineer and worked on military missile guidance systems since 1961. Back in the day we overdesigned everything. What the hell...we were saving the world from the USSR. Cost was not an issue. It was technically interesting work. (I guess for high end audio cost is not an issue either). But when you think about it putting an H-Bomb 100 feet from the target is not really better than 200 feet.
Eldartford, its my opinion that in high end audio you build the equipment to do the job, then figure out how much it costs when you are done. It is a bit different from targeting a bomb :)
If tighter tolerances make a difference...why not design 100% accurate phono?
...Don't encourage him!! :-)
Ultimately, the accuracy of the inverse RIAA equalization on a phono preamp is only meaningful to the point it is equal to the accuracy of the equalization used on the cutter in the production process.

Look at it this way. I can't add "about 2" plus 2.00000 and expect an answer of 4.00000. The "about 2" in the left hand side of the equation limits my answer to "about 4."

If one looks back to the "golden age" of LP production - the 1960s - which produced in particular jazz and classical recordings that are still revered to this day, I seriously doubt that much of the production equipment was much more accurate than 1 dB.

One has to be careful to not chase one specification to the exclusion of others. That happened in the 1970s when amp designers went after ultra low distortion numbers. They overused feedback which gave good looking static test numbers but gave audibly poor results with dynamic music.

Yes, a flat frequency response is a worthy goal. But then so are lots of other characteristics. The goal is not how each one tests by itself in a static environment but rather how they operate in unison when playing music.
Dear Mlsstl: +++++ " Yes, a flat frequency response is a worthy goal. But then so are lots of other characteristics. " +++++

absolutely right, I agree.

+++++ " The goal is not how each one tests by itself in a static environment but rather how they operate in unison when playing music. " +++++

I agree here too.

Both statements are ( between other things. ) an almost a " routine " in the eletronic design whe the designer really cares to improve the quality performance of what already are in the market. IMHO almost no one designer works as if his " unit " lives or will live in a stand alone " stage "/aisle/ static, everyone know the multiple relationship that exist through the whole audio chain and the synergy that is a must to have to otain/achieve a top quality performance.

That's why anyone: customers and designers have to take care that in each link on the audio chain not only exist synergy but that in each link we lose the less on the signal and add the less too trying to preserve the source signal integrity with minimum degradation. IMHO tight RIAA accuracy help to meet those goals.

Regards and enjoy the music,
I might have a different viewpoint here- I feel that it is not a good idea to excercise synergistic effects. Instead, I prefer that each component operate out of its own strengths, that you are not compensating for a brightness in one component by using a dullness in another. IOW there are no synergies, just strengths.

That may be the same as Raul was saying, not sure, but I thought it might be useful to clarify this. People ask us if our equipment is 'voiced' to work together, and it certainly is not, other than our amps and preamps are balanced and all-tube: that's the extent of it.

So this is why we take care with the RIAA- we have no idea what amp or cartridge will be used with our preamp; whatever coloration that might appear as a result will not something that is occurring because we were sloppy about EQ or didn't care. The response will be flat whether its phono or CD. If a cutter from the 60s had a problem, its not an excuse on our part to be negligent, all that results is now you have an instrument with the resolution to really know what is on the various LPs!
Dear Ralph: Yes, I'm refering to that kid and real synergy and not the one that is commun on the reviewers: " if your speakers are to bright don't use this cartridge that is overbright ", this kind of synergy is not a real synergy but a sum of " errors ".

Regards and enjoy the music,