What You Always Where Afraid To
Ask About High-End Audio
Ever heard this Question? Yes? So, what did you say? You explained just how much distortion tone controls produce, how they make the music kaput and so on I'd guess. And of course you mentioned how tone controls and equalizers are for Audio Wussies, how real men know "how to take it" and so on.... I always used to say "My system is so good, I don't need no stinking tone controls and equalizers!". Of course, secretly I was hankering for a good equalizer. One that would allow me to correct for flaws in the recording. One that would allow met to deal with that annoying mid-bass hump in my room, one that would help me get the most from my record collection and system. Of course, only Audio Weenies use equalizers or tone controls. And I'm a Man, a real man, no 1990's "New Man" softie. So no damn equalizers for me, or maybe....
But the question is, do we really need tone controls or Equalizers? Well, I think if you are passionate enough about music to spend major bucks on a High End Audio System, you do need them, desperately. Here is why...
The Listening Room - Enemy
Now in case you have been reading the late lamented "Audio" Magazine you already know what I'm driving at. If not, I'm looking here at Audio January 1999, at the review of a pretty respectable Satellite Subwoofer system, capable easily of a frequency response of +/-3dB between 20Hz and 20kHz in the anechoic chamber owned by the reviewer. However, the Reviewer, Mr. D.B. Keele JR had the habit of also printing the response of the reviewed speaker in his listening room, at 3m distance.
Looking at this graph is a revelation. Printed only above 100Hz we see series of major dips and peaks to the tune of as much as +12dB and -20dB compared to the overall mean SPL Level of 80dB. If Mr. Keele had included the sub 100Hz response the graph would have been way of scale. If you are interested in the room resonance's of a real room (from my cyber buddy Brian Steele on Grenada) click here for their website. Can you imagine what a 24dB Peak at 66Hz sounds like?
So, after we have placed our perfect speaker into a normal room the perfect Frequency response of our perfect speaker is so FUBAR (Fu<&ed Up Beyond Any Recognition). And there are so many discrete dips and peaks of quite substantial magnitude that a pretty sophisticated Equalizer is needed to correct that. Some may dispute that this frequency response, the one that includes the resonant field and the direct sound is the one that makes for our perception of tonality, but based on my own experiences in both pro-audio and high-end I beg to differ. The more even the in room, total summed response, the more natural the sound.
Hence any decent equalizer that can flatten out this fever curve most if not all speakers produce in normal rooms is good news.
The Recording Engineer - Public Enemy #1?
Indeed, over the decades we can observe differences of as much as 20dB in tonal balance between top-grade Studio Monitors. And these where the "yardsticks" the Mastering and Recording Engineers used to judge the sound of the recording. If they felt something was amiss, yup, they equalized it out. So even if the sound engineer did his or her best to obtain a sound for the recording that was natural and evenly balanced (not that all many do this in the first place), the result will be colored by the choice of Monitor used during recording and mix down.
The butt-naked truth... ugly, isn't it?
So, here you have it, assuming we want to hear our music as it should be, we need to equalize out the effects of speaker and room and correct for the tonal balance of the recording. Without doing so you maybe get 50% of what is possible. Why is it then that these simple facts are not shouted all the time from the pages of any and all Audio Magazines? How comes these simple truth is so rarely spoken? The reason is simple. Until very recently the best professional equalizers where at the best not all that good (not to speak of the cheap rubbish of the "10-Channel EQ with Analyser" sold in any Circuit City and similar chain). To use a professional 1/3rd octave graphic equalizer or a professional parametric equalizer is not necessarily easy either. In the end, no-one wants to fiddle constantly with halve a million of buttons (never mind remembering what they actually do) all the time. Lastly, the classic Bass & Treble controls are about as useless to correct recording or room response problems as a #9 Ball Peen Hammer to fix a Swiss watch (Editor Steve sez: not near my Patek Philippe collection you don't)
Hence, no equalizers or tone controls in almost all audiophile systems. True, there where notable exceptions, especially the Cello Palette Equalizer and more recently the Z-Systems Digital Preamp/EQ for correcting the tonal balance of recordings and things like the Sigtech and the Tact digital frequency response correction systems. Yet all these tend to cost a very substantial sum of money and especially Digital Units look like a poor investment, considering the upcoming new digital formats. All in all an understandable, but sorry state of affairs. (Editors Steven says: You know Thorsten that i reviewed the Z-Systems unit and had the Cello too. Cello was horrid and colored, yet the Z-Systems rocked)
As us sez in Germany - Vorsprung durch technik...
Okay, why did I go through all this spiel, dissing the poor room, the poor sound engineer and all? Because I have what I think is the closest to a solution to the abovementioned problems yet. And it does not even cost the earth.
Now That Is Something To SHOUT
So What The Heck
Is The "Ultracurve Pro", I Hear You Crying
It comes complete with Crystal 24-bit analogue to digital and digital to analogue converters on board. You connect it to your system and start equalizing..... ;-) The general build quality is on par with better mid-fi gear of the Cambridge Audio and Rotel kind, adequate though not exceptional. If it where a car it be a probably a fully loaded Volkswagen Golf GTI Convertible, with loads of extra tuning and an acceleration that would beat most short of real hotrods or a corvette. The sound, well read on for that. Sufficient to say that I'm buying the review sample.
Oh yes, I forgot, the U.S.A. list price is $917 including a measurement microphone and the digital input/output option. I have seen the unit advertised on the internet in the same configuration for less than 2/3rd of the price, so don't pay the list price. Go to your nearest Pro-Audio shop and buy one. NOW! Don't bother reading the rest of the review, it's boring.... ;-)
Some Technical Notes On The Ultracurve Pro
Over the holidays I spend some time replacing ALL electrolytic Capacitors in the Ultracurve Pro with Sanyo Os-Con's (around 15pcs 10uF/25V, 10pcs 47uF/16V and 5pcs 330uF/6.3V), with the exception of the main PSU reservoir capacitors, for which I used Panasonic FA Series (3pcs 3300uF/35V). I also replaced the JRC 4580 Op-Amp's with LM6172 (secret tip - these Op-Amp's SMOKE anything else for Bitstream AD/DA Circuits). For this I had to make DIL to SIP adapters from some SIP Sockets, tedious but very well worth it. To ensure stability of these superfast Op-Amp's I included directly in contact with the chip a 1uF Siemens stacked film Capacitor for each Op-Amp in the Adaptor, connected between the positive and negative supply. I also replaced most regulators with LM1085 (1pcs Adjustable, 3pcs 5V fixed), the LM1085 is the equivalent of the LT1085, but halve the price. My ears hear no difference between the LT and LM.
Comparing the modified and 24 Hours burned in Ultracurve Pro in the bypass test (no EQ selected, the 5db attenuation corrected) against the straight plugging together of the cables into each other left only the slightest remaining impact on the sound. I'm impressed. With these mods in place the Ultracurve Pro would score easily 95 - 98 on the analogue inputs/outputs. The total Parts cost for the modification came to around $130 US, parts source was RS components UK.
Now for level setting story promised earlier. The Behringer processor needs to have a level of +21dbu rms. for digital full scale (about 8.7V). The true resolution of ADC chip in the analogue input is around 20Bit (the notional Bitrate is 24 bit). This means if a normal Poweramp with fully opened level controls where used then for absolute clipping in the resolution from the converter would be barely 15 bit, with an additional 20db attenuation less than 12 bit. Now this would sound harsh and distorted as has been reported. Ergo, feed the Ultracurve Pro input with a lot of Level and you get good sound.
In my system the power amplifier has level controls and they are matched with the Preamp so that power amplifier clips when the preamp Volume is near 3 O'clock. I found that at this volume I can usually trigger digital overload in the Ultracurve Pro, so with the Preamp Volume at 2 O'clock the Peaks just avoid overloading the equalizer. Setting the power amplifier to deliver then it's full subjectively undistorted output (around 108db Peak at the listening position) gives a matching of the Digital processors dynamic range with analogue systems dynamic range and voila, no problems. Running at conventional "-10dbV" Consumer audio levels shaves around 30dB of the total available dynamic range thus severely compromising low level performance. Seeing as it is, the comments in rec.audio.pro and similar groups about how bad the UC sounds and so on; I suspect that these self-styled audio "Professionals" do not even know how to correctly set levels in digital systems....
Anyway, being digital with an analogue performance more in line with 20 bit digital audio for the input and 18 bit digital for the Output it is IMHO essential to get the correct levels with UC, otherwise the sound will be pretty bad. Considering however just how little the sonic degradation of the EQ circuitry itself is I to have to take some more care in level setting is a small price to pay.
Bring On The Boring
Anyway, I ended up with system levels set to get the best dynamic range (takes a bit and you do need a power amplifier with build in attenuators) from the processor and the Ultracurve Pro between the amplifier and preamplifier. I initially used some homemade XLR to RCA Adapters of good but not excessive quality. Simply plugging the XLR output and input plugs together allowed me the ultimate bypass test, eliminating the Ultracurve Pro totally from the signal path. While I did spend some time using Ultracurve Pro in pure Digital Mode between Transport and DAC I rarely listen to digital, so most of my evaluation was done with Ultracurve Pro between pre-amplifier and power amplifier. Linking the Ultracurve Pro into the tape loop of my pre-amplifier did not prove to give very good sound, the Ultracurve Pro does like a lot more level than available there.
Testing the bypass against the Ultracurve Pro via the analogue IO's I found I needed to boost the gain on the Ultracurve Pro by 5db, as there is a 5db overall attenuation. With that done I found that the sound was somewhat altered and to the worse, yet not so terribly that I yelled "take it out". Comparing the bypass against running the signal through the Ultracurve at 48kHz sample rate showed that the soundstage was somewhat flattened and the sound acquired a slight edge to the tone, non too terrible. I expected much worse. Just for fun dropping the Level an extra 20dB and turning the power amplifier up fully sounded very much worse. Now so far all I was assessing was the sound of the Ultracurve Pro doing nothing. I would class the degree of loss from running through the Ultracurve Pro as slightly larger than a really good High End Preamplifier compared to passive Volume control and the slight edginess made the solid state nature of the unit well known.
This by the way held true for both Vinyl and CD being played. I think considering that the Analogue signal from Vinyl is first digitized, then converted back to analogue the fact that only minimal sonic degradation happened is most impressive. As an interesting experiment I dropped the sample rate from 48kHz to 44.1kHz. Not good. The sound now turned outright harsh and had all the hallmarks of what is classically wrong with CD. In the digital loop between CD transport and DAC the Ultracurve Pro was essentially sonically neutral, without audible contribution unless the EQ is employed. The only thing I reliably noted in the digital loop was a reduction in "speed" and "pace" and a ever so slight reduction in soundstage depth. My next stop was the quite substantial (over fifty A4 pages) Manual, to see what I would need to do to make this box of tricks jump through the hoops.
I found something called Auto-Q, a function that will automatically equalize the response of your speakers and room flat. You simply place the supplied measurement at the listening position, point it forward and engage Auto-Q. Within around one minute the system will equalize both channels as flat as it can. Of course, the microphone needed a long XLR Microphone Lead and I had non at hand. After a trip down the Pro-Audio store on the next day to buy one I simply hit the "frapee" button, stepped out of the room (I hate listening to pink noise) and let the box do it's tricks. When all the different noises had stopped changing I stepped back in the room, turned the system over into normal play mode and started playing music.
Looking at the equalizer, I found that the EQ curves for both speakers where quite different, as I had expected in a somewhat asymmetric room. I also found that I had as much as 14db attenuation in the 50Hz and 63Hz bands showing the rooms major room modes. The "floor bounce" induced suck out around the 100Hz...160Hz range was also filled in. In the upper midrange a around 4dB lift in the 2 to 6.3kHz range (I knew about this one of course) was also flattened and the HF substantially lifted.
The result? Flat in room response from 30Hz to 16kHz. 20Hz and 20kHz where not fully boosted and hence where a bit down. I boosted the 20Hz and 20kHz slider a bit, to make sure the Auto-Q function would sort these out as well and repeated the process, this time watching. I was both astonished and pleased to see that the system now managed 20Hz to 20kHz at +/-0.5dB at the listening position! Beat dis!
How did it sound afterwards? Wow! What a difference!!! Initially I was not quite sure if I liked this new "flat from 20Hz to 20kHz" system, but quickly the sound grew on me. It is a dramatic change, no beef about that. The bass gained definition, speed and "deep impact", in the midrange Voices suddenly sounded all that more real and the soundstaging showed much more solid images.
My next stop was putting some of my knowledge about using equalizers into practice. The Ultracurve Pro has in addition to the graphic equalizer, three parametric equalizers per channel. I programmed these in such a way that the bulk of the heavy and wideband changes (like the deep notches around 60Hz, the boost at 100Hz and the wide and shallow drop in upper midrange) where carried out by this parametric EQ. The bandwidth of the parametric EQ is programmed in the initially unintuitive unit of X/60 Octaves. Well, it is unintuitive until you realize that 20/60 is equivalent to one slider on the graphic EQ. So transferring the settings from the graphic EQ to the parametric ones is deadpan easy.
Simply re-running Auto-Q now left me with an EQ where most sliders did not boost or attenuate more than 6dB. There where some "outliers", especially above 10kHz and below 50Hz, but not all that much. I saved this Curve to the first Memory, wow, that was easy. I then changed the EQ settings automatically arrived at by taking all "outliers" down to a maximal 6db boost or cut. Then I used the very useful A/B comparison function to convince me that on 95% of the musical material the two different EQ curves had only minimal sonic differences and non that would give ether setting the nod over the other. This slightly less extreme EQ setting was saved in the memory position 2. In the end the second Equalizer Program left me with the ability to add at least 10dB cut or boost to any of the frequency bands, more than enough to correct the source.
Each memory can be named (this is a bit tedious but immensely helpful). The Memory system also allows to overlay additional EQ Curves upon the existing one or to subtract from it. I used that to program myself a few basic Curves for source correction, nothing outrageous just some bit of BBC curve, a bit of "tilt" as found in old Quad Preamplifiers and other miscellaneous curves I thought where good ideas. It barely took ten minutes. Then I finally sat down to listen seriously.
Playing "Solid Air" from the Q Magazine chill out sampler (CD) the whole image was more solid, more palpable than before, small details like the handling noise of the acoustic guitar was more distinct and notable. Some cut's from the Windham Hill Jazz Sampler also showed more air for cymbals, more evenhanded and sweet vocals and in all cases a fast, dry and precise bass with serious bottom end reach. Moving on to massed Vocals of Haendels Messiah on Decca again, the vocals sounded much better, but I quickly cut in the additional "BBC" EQ curve, and off I was correcting the source problems as well.
As for the general sonic impact on my system, once I had listened for a few hours "flat" and switched the EQ out, the system suddenly sounded like poor plastic boombox, with overblown, slow midbass, lacking upper bass and a not so nice but rather tinny midrange. The shock from doing this comparison after a few hours acclimatization to "flat" for the first time made me realize just how much the room messes up the sound of even rather good speakers. And my room is actually quite decent sounding and arranged quite sensibly for general acoustics. Many who have heard my system classed it as certainly in the rarified upper ranges of Ultrafi too. Yet even so the step of going form a very good Ultrafi System sans Ultracurve Pro to the same with was no small matter.
Listening for over a week this way, with the Ultracurve Pro in the hi-fi rack I found that I really wanted to correct more often than I could bring myself to get up. What to do?
Really Getting Into It!
Having the EQ placed there and using the "shelf tool" of the UC allows extremely quick and effective equalization of the Music when required. The "shelf tool" for the Ultracurve Pro is really about four aces and a few jokers up the sleeve of the Ultracurve Pro. The "shelf tool" allows the 31-band graphic EQ to be used very much like a the Cello Palette by simply selecting a 3dB/oct slope for the "shelf" (Setup Menu) and by selecting suitable center frequencies of 20Hz, 125Hz, 500Hz, 2kHz, 5kHz and 20kHz for starters. The curve can be selected as the classic "Bell" curve or as "shelf" where below or above a certain corner frequency all remaining sliders are boosted equally, with a slope as programmed towards the rest of range. The level control is not as fine as with the Cello Palette, but more than sufficient in practice. Ahead of the Palette is the ability to shift the center or corner frequencies, so that one can center straight in on any sonic anomaly.
The four left buttons control level and frequency in an arrangement that is quite intuitive. I find I'm always having my eyes closed when adjusting these, making it easier to hear what is going on. With a little practice finding the correct balance for a given recording takes very little time. Using the A/B comparison function the before/after comparison is a breeze and will quickly tell if the EQ is alright yet or even too much. Having a 0.5dB step volume control (limited to +/-16dB) directly at hand is also quite good, I find myself fine-tuning the Level by half a db or a bit more quite often.
Comparing the ergonomics with the Cello Palette equalizer in a similar application I would still feel the Cello Palette to be easier to use (side note: Cello is no longer in business). The requirement to step through menus on the Behringer is distracting and the menus are not as ergonomic as they could be, but flicking through them quickly becomes routine. For corrective scope and especially delicacy and pinpoint accuracy of correction the Behringer Ultracurve smokes Cello big time (Steve says: i have owned the Cello and found it uninspiring). Sonically I suspect that there will be not all that much between the two either, but I have no recent experience with the Palette in my system, so I cannot be too sure. Of course, even at knockdown 2nd Hand prices the Palette is out of my league anyway, the Behringer Ultracurve Pro is distinctly within it though.
Just for fun I recorded my correction settings for a few randomly selected LP's I played. Seeing the Selected EQ slope is easy by using the preview of the Memory subtraction function, otherwise things are hard to spot as the Graphic EQ shows the Room/Speaker correction overlaid with the Source correction. Anyway, listening to Eryka Badu's 1197 LP "Baduizm" I found that I took a "bell" shape EQ with -2dB centered around 3.15kHz, a low boost shelf at 100Hz with 2.5dB boost and a high frequency cut shelf with -2dB. The difference between the before and after was surprisingly large, considering the modest level of equalization. Gone was the overly incisiveness on the vocals, gone the overly hot and crisp cymbals, the soundstage gained depth and the Drums just kicked that little bit more convincing.
Turning to Billy Paul "Live in Europe" (1974 on Philadelphia) I corrected a lot more, a 3dB cut in bell shape around 6.3kHz, a low boost shelf at 200Hz with +3.5dB and a further 3db bell EQ boosting at 80Hz to give the kick drum more "kick". Moving on to Emmylou Harris "Bluebird" (1989 on Warner/Reprise) I needed to work real hard. A +3dB Shelf boost at 160Hz, a +2.5dB bell boost at 50Hz and a -3dB shelf cut at 3.15kHz made the overly jangely, very AM sounding country rock mix much more listenable. Now the vocals lost a little definition, so I put a little +2db bell boost at 8kHz in. Perfect, still janely, hot C&W, but listenable, kick the volume up 3dB, lovely.
For a final EQ setting note, playing Vivaldi's Gloria in D-Minor (on Loiseu Lyre/Decca) with Emma Kirkby on Vocals I found a +2dB shelf boost at 250Hz gave the tome more warmth, while a -2dB bell cut at 5kHz took the edginess out of the Voices. All in all impressive and I would say that I could think of few commercially available equalizers that would allow the "remastering" of tonal balance in such precision and scope so easily.
In the end I would rate the impact of the Ultracurve on my system in the same region as that of high quality full range drivers or single ended triodes. It brings me closer to the music and makes it easier to forget about the mechanics of reproduction. For the money it's a clear bargain if there ever was one. Even better is that due to being a Pro-Audio unit you may find it likely quite easy to hire a unit for a few days to evaluate at home. A number of the Pro-Audio shops I frequent will hire you the demo / rental Unit for a modest cost and will credit this money towards a purchase of the unit while still giving decent discounts. Go and haggle, pro-audio is highly competitive marketplace, you'll be surprised about the deals possible. And quite frankly, at the price I do not care the tiniest bit if the Ultracurve Pro becomes obsolete two years from now as it will not support 96kHz or 192kHz sample rate (or SACD). By then we should have a 192kHz/24-bit capable model which I will buy unseen and unheard!
So really, this EQ is here to stay. Of course, equalizers are an absolute heresy - so go ahead and burn me at the stake. As usual, good tunes to you all, until soon, and remember...
Enjoy the Music!
Analogue Inputs: XLR Balanced & 1/4" Jack; +21dBu/8.7V max.; 50k balanced & 25k unbalanced Input Impedance
Analogue Outputs: XLR Balanced & 1/4" Jack; +16dBu/4.9V max.; 60 Ohm balanced & 30 Ohm unbalanced Output Impedance
Digital Input (optional): XLR Balanced; AES/EBU Transformer Balanced
Digital Output (optional): XLR Balanced; AES/EBU Transformer Balanced
Sample Rates supported: 32kHz, 44.1kHz, and 48kHz
Digital Processing: 24 Bits
Frequency Response: 20Hz to 20kHz +0/-0.5dB
Signal to Noise ratio: 103dB unweighted 22Hz to 22kHz, Analogue In/Out
Harmonic Distortion & Noise: 0.004% @1kHz/+4db Input
Display: 240 X 64 dot matrix LCD, backlit
Power Supply: 100 - 120V AC or 200 - 240V AC selectable, 50 - 60Hz
Power Consumption: 21 Watt
Dimensions: 3.5" x 19" x 12" (HxWxD)
Weight: 11 lbs. (net.); 13.2 lbs. (shipping)
Price: $699, plus $129 for the AES/EBU and $89 for the ECM8000 measurement microphone