I hope this thread starts a lively discussion including the pros. I've been tediously rebuilding my electronics for years including using LM opamps, naked Visay resistors, premium tubes, eliminating caps in the signal path or worst case replacing with teflon, or VitQ, etc. Preamps, EQs, crossovers, power amp front-ends, etc. Every time there is a marked improvement in areas that I don't know how to measure: openness, clarity, texture, smoothness, space, snap, etc. A radio shack 10k carbon film and a naked Vishay can both measure exactly 10k ohms but they "sound" very different, the Vishay being significantly better - clearer, less hazy, less electronic. I don't use those terms to be esoteric, I just don't know how else to describe what I hear. As a concert pianist I am well aware of what a 9' Steinway should sound like and most electronics completely reduce the sound of a piano to complete crap! Everyting that makes a live piano sound beautiful is missing. Playing in a live acoustic jazz trio has created my reference when I listen to electronics.
My guess is that measurements taken before and after mods would probably be almost the same. 20 - 20k + or minus a few tenths of a dB one way or another is NOT significant (a piano sounds quite different day to day depending on humidity, temperature and more!) yet the "sound" after modding is completely different. Distortion (according to the manufacturer) can measure in .000s of a percent before the mod yet the original sound can be sterile and lifeless. Upgrading the power supply can greatly improve a units sound yet there is no "specification" to show why a preamp sounds better with better filter caps. If the manufacturer says 20 - 20k +-.1 dB and 110dB signal to noise then what is not being shown on the specs that make the unit sound so poor by comparison in stock form. How about a tube device with much worse specs but that sounds much more like a real piano? What changes when I upgrade parts that improve the sound even though the original specs showed noise and distortion below the theshhold of audibility? why do the worst measuring devices (tubes, class A, no feedback) sound most like real instruments?
I hypothesize that the gross distortions of loudspeakers can be less audible than the minute distortions of electronics based on the way our hearing actually works. A concert piano has TONS of noise from the key action, pedal board, probably a signal to noise ration of 20dB or so yet it sounds magnificent. Are manufacturers measuring all of the wrong things?
My guess is that measurements taken before and after mods would probably be almost the same. 20 - 20k + or minus a few tenths of a dB one way or another is NOT significant (a piano sounds quite different day to day depending on humidity, temperature and more!) yet the "sound" after modding is completely different. Distortion (according to the manufacturer) can measure in .000s of a percent before the mod yet the original sound can be sterile and lifeless. Upgrading the power supply can greatly improve a units sound yet there is no "specification" to show why a preamp sounds better with better filter caps. If the manufacturer says 20 - 20k +-.1 dB and 110dB signal to noise then what is not being shown on the specs that make the unit sound so poor by comparison in stock form. How about a tube device with much worse specs but that sounds much more like a real piano? What changes when I upgrade parts that improve the sound even though the original specs showed noise and distortion below the theshhold of audibility? why do the worst measuring devices (tubes, class A, no feedback) sound most like real instruments?
I hypothesize that the gross distortions of loudspeakers can be less audible than the minute distortions of electronics based on the way our hearing actually works. A concert piano has TONS of noise from the key action, pedal board, probably a signal to noise ration of 20dB or so yet it sounds magnificent. Are manufacturers measuring all of the wrong things?
Last edited:
Loudspeaker distortion/interaction with the room is significantly more of what we hear than the electronics...
Could be a more pleasing frequency response to you?
Could be a more pleasing frequency response to you?
I believe that the reason that changes in the sound of electronics are nevertheless noticeable is that the signature of the acoustic environment (loudspeaker-environment interaction) does over time become transparent. Thus the success of 100-hour burn-in of loudspeakers.
What do you mean transparent?
Your ears might get used to it, but the room interactions and higher level of distortion from the speaker remain.
Your ears might get used to it, but the room interactions and higher level of distortion from the speaker remain.
Interesting posting - thanks. It reflects my current line of tinkering too, though I'm not retrofitting 'premium' parts, rather correcting layout errors (mainly insufficiently thought-through grounding).
My current hypthosesis is that noise floor modulation might be one of the missing aspects that's not being tested for. But equally, I'm sceptical of current measures of 'threshold of audibility' - to my understanding they deal with tones and don't take account of stereo imaging effects. I remember at an AES long long ago before the widespread acceptance of mp3 Michael Gerzon mentioning that one of the areas not addressed by perceptual coders at the time was 'stereo unmasking'. I have no idea if its taken account of in more recent algorithms.
Glowbug, perhaps you make my point better than I can. What intrigues me is that despite the distortions of the speakers and speaker/room combo I can still hear very audible changes when I swap components that SHOULD be below the threshold of audibility.
In the fairness of full disclosure the speakers i am currently using consist of lacquer soaked Eminence 3012HO and JBL2407 (BMS4540nd) on CD horn crossed biamped at 1200hz 24 dB/octave plus a seperate subwoofer. I use a highly tweaked White instruments EQ and 1/6 octave analyzer to set the response to mostly flat at my listening position. My room is NOT perfect but I have a large array of decorator pillows on the walls to absorb the worst reflections. I would guess that the thd of the speakers is below 1% at my domestic listening levels.
On a side note there is a fascinating level of microdynamics that exist in many recordings which I believe most people never get to experience and I am having fun discovering them!
In the fairness of full disclosure the speakers i am currently using consist of lacquer soaked Eminence 3012HO and JBL2407 (BMS4540nd) on CD horn crossed biamped at 1200hz 24 dB/octave plus a seperate subwoofer. I use a highly tweaked White instruments EQ and 1/6 octave analyzer to set the response to mostly flat at my listening position. My room is NOT perfect but I have a large array of decorator pillows on the walls to absorb the worst reflections. I would guess that the thd of the speakers is below 1% at my domestic listening levels.
On a side note there is a fascinating level of microdynamics that exist in many recordings which I believe most people never get to experience and I am having fun discovering them!
Last edited:
Yes. And when this happens, it becomes of less significance than the numbers (distortion) might suggest. Consider that the ears are only transducers; we actually hear with the brain which, I believe, is responsible for processing what we hear. Over time it could very naturally neutralise the sonic signature of the loudspeaker and acoustic environment. When we talk about threshold of audibility we are considering the information reaching the ears, but the function of the brain in the process upsets the equation.
Is this true of the direct mike feed in the studio or concert hall recording? Can I assume that every recording in my collection (save one or two exceptionally boring "audiophile" specials) will sound like complete crap regardless of the care and work I've put into designing and building my system?
A few days ago I was listening to a radio program which dealth with the composition details of a classic music piano concert.
The program producer-a pianist himself- played part of the piano bars in a piano that was in the radio station studio and later on he had the same part been played from a DG renowned recording.
The dynamics, the full sound and the liveness of the radio station piano was some classes above the professional recording of the DG record.
The detrimental effect that all the equipment - from the radio station studio up to my ears (included) - imposed on the sound, could not mask such a huge difference.
I have to live with the castrated sound of the recordings I can afford to buy.
Best Regards
George
It's generally the recording itself that is at fault. I used to think that I would never hear a stereo system reproduce a grand piano accurately until one day I bought a cd of Georges Pludermacher playing Debussy. The quality of the recording was astounding and even my mediocre homemade system produced the sound of a real piano. I've since come upon some other good piano recordings, but not many. It's the recording engineer's and producer's fault that you can't get a good piano sound.
http://www.amazon.com/Debussy-Child...=sr_1_5?ie=UTF8&s=music&qid=1265902709&sr=1-5
http://www.amazon.com/Beethoven-Com...=sr_1_4?ie=UTF8&s=music&qid=1265903492&sr=1-4
John
Last edited:
Jan Didden gave an excellent presentation and demonstration last year about the major culprit: compression. A few small labels avoid it, but it's pretty much the rule these days. The frustrating thing is that it's deliberate.
Exactly. And if you don't think you can get superb performances from audiophile recordings, take a look at the Caro Mitis label from Russia:
Caro Mitis
SA-CD.net - Beethoven: Complete Piano Sonatas Vol. 1 - Igor Tchetuev
SA-CD.net - Beethoven: Complete Piano Sonatas Vol. 2 - Igor Tchetuev
SA-CD.net - Beethoven: Complete Piano Sonatas Vol. 3 - Igor Tchetuev
SA-CD.net - Schnittke: Complete Piano Sonatas - Igor Tchetuev
Tchetuev has become my new favorite living pianist.
John
Compression seems to be where everyone is placing the blame but I think it's distracting people from what the real problem with most CDs actually is. The spectral balance imo more often than not is way too bright. It's as if they are EQing for low level listening by boosting up the bass and the 2-7kHz area.
There are plenty of recordings on vinyl that use compression which don't suffer from the same fatigue inducing problems. There are a handful on CD. And I'm sorry it has nothing to do with jagged digital waveforms or records having more warmth or any number of audiophile nonsense. It's all about the spectral balance being goosed.
There are plenty of recordings on vinyl that use compression which don't suffer from the same fatigue inducing problems. There are a handful on CD. And I'm sorry it has nothing to do with jagged digital waveforms or records having more warmth or any number of audiophile nonsense. It's all about the spectral balance being goosed.
Every time I changed speakers, I experienced a quantum leap in sound improvement. Not so in electronics.
An example. When I first hooked up my electrostatic panel concoctions, I used a lousy 50Hz power transformer for a step-up tranny, powered by a cheapo solid state amp (transformer ratio was not optimal, and the panels needed the power). My previous speakers were not exactly "budget" and were powered by a much nicer amp. Still, the improvement in sound was massive.
An example. When I first hooked up my electrostatic panel concoctions, I used a lousy 50Hz power transformer for a step-up tranny, powered by a cheapo solid state amp (transformer ratio was not optimal, and the panels needed the power). My previous speakers were not exactly "budget" and were powered by a much nicer amp. Still, the improvement in sound was massive.
Sy, I've learned to never assume anything! If you have the opportunity, go to a piano store and pretend you want to buy a grand piano. Any quality piano (Steinway,Bosendorfer, Baldwin and several others that cost at least $50,000) that is 7' or larger will do for the audition. Stand next to the piano with the lid open while it is being played to demonstrate. Chances are what you hear and experience will be very different that what you hear from recordings on a stereo system. There is tremendous detail, texture, power, nuance and attack. The sound may be less bright than what you are used to. (the result of engineers "improving" the piano sound) A hard quick power chord will actually jar your head for a moment! The low notes should be felt throughout the room. (think of the size of the soundboard compared to most woofers...)
Cheap pianos from the East will not work for this demo. I have played many $10,000 (the yugo price range of pianos...) grand pianos that sound terrible and completely lack any richness, color or texture.
Last edited:
Oh, I've heard quite a few first-rate pianos in a live setting, and it's a wonderful sound. That's one of the key things that bothers me about compression in recordings. What I'm taking issue with is your assignment of that issue to the electronics as opposed to the person operating the electronics (as well as the transducers!). Live mike feeds and recordings made without (or with very limited) compression sound, to my ears, VASTLY better, despite those mike feeds passing through the self-same sorts of electronics.
The other issue is miking- a piano sounds vastly different twenty feet away in a reverberant space. Many engineers do an excellent job of capturing that sound. The noises you speak about are highly audible sitting on the piano bench, but not from the audience. And good recordings are generally NOT trying to reproduce the sound of the piano from your perspective, but from the audience's.
The other issue is miking- a piano sounds vastly different twenty feet away in a reverberant space. Many engineers do an excellent job of capturing that sound. The noises you speak about are highly audible sitting on the piano bench, but not from the audience. And good recordings are generally NOT trying to reproduce the sound of the piano from your perspective, but from the audience's.
Its been a while since I visited my own thread... Sy the point I'm trying to make (but not doing a very good job of!) is that to MY ears, anything less than a "premium" signal path is highly audible, even though measurements suggest that the differences are minute. Additionally, the sound of a cluttered signal path is highly offensive to my asthetic, even though measurements suggest that the differences are minute. Yes speakers are generally highly inaccurate but my comments do assume a certain baseline as I listen in calibrated spaces to speakers with accurate frequency and phase response.
It's incredibly easy to fall into the belief that electronics makes a huge difference. I suppose whatever bothers us the most can be considered huge, but in the grand scheme of things those electronic differences are dwarfed by other factors.
The stark reality is that once frequency response is taken out as a variable, blind testing has shown most electronics to be indistinguishable. I'm still in favor of reducing odd harmonics in particular, and reducing all harmonics to as low a level as possible, but speakers and room acoustics are where it's at.
I grew up with my mom playing a rather old Knabe piano made in nearby East Rochester NY. It has a soft wonderful sound when in tune. Some time ago I listened to a brand new and very expensive Japanese piano and was horrified by the thing. No doubt someone who grew up with it would find the Knabe unlistenable.
Conrad
The stark reality is that once frequency response is taken out as a variable, blind testing has shown most electronics to be indistinguishable. I'm still in favor of reducing odd harmonics in particular, and reducing all harmonics to as low a level as possible, but speakers and room acoustics are where it's at.
I grew up with my mom playing a rather old Knabe piano made in nearby East Rochester NY. It has a soft wonderful sound when in tune. Some time ago I listened to a brand new and very expensive Japanese piano and was horrified by the thing. No doubt someone who grew up with it would find the Knabe unlistenable.
Conrad
I haven't found that electronics makes a huge difference, but the difference electronics makes isn't always dwarfed by other factors. The difference is different from those other factors - orthogonal to them.
I take it you are implicitly excluding digital electronics from your claim. Otherwise, its trivially easy to see that the deficiencies of, say the mp3 codec aren't primarily in the realm of frequency response.
Why the fascination with the minutiae of harmonic distortion reduction? I'm not in any way denying that room acoustics and speakers are hugely important btw.
Sorry, I barely acknowledge digital electronics. I think I heard an MP3 once, but found it too awful to try twice. Ok, I'm a dinosaur.
I'm not so fascinated with reductions in harmonics to senselessly low levels, and a single THD number is basically worthless, but those are the numbers remaining that separate modern electronics from perfection. Once an amplifier replicates its input in terms of time and voltage, there isn't much left to talk about.
Now, it is important to test under real world conditions, with real world loads, and few people do that. I love my test bench dearly, but audio equipment has to perform in a different setting and I admit to rarely dragging everything to the bench and confirming it works as a full system in the same manner that the individual components do. IMO, there may be some discoveries concerning RF pickup, ground loops and cable interactions awaiting those willing to do more of that.
I do agree that where differences can be heard, they tend to be of a different nature between transducers and the rest of the signal chain. There are still pitfalls. I compared several very different microphones some time ago. When I started I thought the differences were obvious. When I tried to really describe and characterize each one, they started to sound more and more alike. The ear/brain adapts itself very rapidly, so we're often chasing a moving target.
CH
I'm not so fascinated with reductions in harmonics to senselessly low levels, and a single THD number is basically worthless, but those are the numbers remaining that separate modern electronics from perfection. Once an amplifier replicates its input in terms of time and voltage, there isn't much left to talk about.
Now, it is important to test under real world conditions, with real world loads, and few people do that. I love my test bench dearly, but audio equipment has to perform in a different setting and I admit to rarely dragging everything to the bench and confirming it works as a full system in the same manner that the individual components do. IMO, there may be some discoveries concerning RF pickup, ground loops and cable interactions awaiting those willing to do more of that.
I do agree that where differences can be heard, they tend to be of a different nature between transducers and the rest of the signal chain. There are still pitfalls. I compared several very different microphones some time ago. When I started I thought the differences were obvious. When I tried to really describe and characterize each one, they started to sound more and more alike. The ear/brain adapts itself very rapidly, so we're often chasing a moving target.
CH
- Status
- Not open for further replies.