We are on the same page, kudos.
When I say noise I mean deviation from ideal...this means dynamically modulated white/pink/1-f system noises in addition to THD errors.
Dan.
Last edited:
The whole issue RNM was pointing out is that energy well beyond the "audible" range still screws up the stages that follow. Hand wavy? With actual measurements of the noise, a before and after listening test and measurements that show increases in audio THD from this sort of noise.
You also don't seem to understand logic levels. The power supply may be 3.3 or 5 volts. The actual levels are a bit less.
Enormous depends on the probe used. My sniffer give me several hundred millivolts in some chassis.
No one here listens to CDs that are constantly -10 dB nor are they made for any genre except one. When discussing issues of fidelity one does not include that genre. (I doubt you can name it.) I and others have looked at the spectra of music live and recorded with various media. It is an important part of the design issues.
I do not agree with this. This would be true only for old fashioned poorly engineered slow designs, typically from seventies. Well engineered, internally fast solid state amplifier, with some LP filter at the input at about 100 - 150kHz/-3dB will have no problems. The HF noise of some tens and hundreds of kHz at a moderate or low level that is coming from digital sources will not "screw up the stages that follow". I also have a plenty of such measurements and my experience is that only poor designs are sensitive to this.
For those who disagree, please try clipping test of your amplifier at 100kHz. If it passes with a nice plot, I guarantee to you that no worries about HF from digital sources is needed.
Last edited:
Been listening to this today Linn Records - cantus Now its has to be said that the Marimba is hardly the most harmonically complex instrument but the recording is a nice example of how you can do things in digital that Vinyl just doesn't have a low enough noise floor for as the reverb dies away to silence.
haven't decide if its simplicity makes it a good or poor test record, but it's good for late night listening with a glass of ones preferred poison to unwind before sleep.
haven't decide if its simplicity makes it a good or poor test record, but it's good for late night listening with a glass of ones preferred poison to unwind before sleep.
Pavel,
I think you just reinforced what RNM is saying. But my looking at many amplifiers and designs of what you may consider mid-fi show a lack of ability to handle direct input of super band signal.
I won't try a 100 kHZ clipping test on most amplifiers around here because it probably would result in damaging the output stage.
Besides I would rather not have any out of band content.
Perhaps, but that's what keeps many of us in this hobby. If it was nice and easy, we might wind up wuth gear we don't like, despite its technical prowess and brilliant measurements. That was, is and will be so until we come up with a measurement system which can adequately take into account our subhective perceptions of music.
It hans't been once I had in my system a say power amp which measured very well, had oodles of power and was brilliantly made using high quality parts which somehow didn't sound right to me. And, just to complicate thuings a bit more, my loudspeakers neasure as an exceptionally easy load, so they can't be blamed for significant differences.
But the opposihe has also happened to me when an unassimung integrated amp had exactly 3 items changed inside that made it behave in an unexpectedly better way, as if it that was a magic formula of how to modify a standard mass produced item into a de luxe sounding item in three easy moves. It's measurements remained the same but its sound improved like it took me months to really believe was even possible. And given my experience with amps, as well as my very solid collection of them from various sources and price points, I am not easy to impress.
It hans't been once I had in my system a say power amp which measured very well, had oodles of power and was brilliantly made using high quality parts which somehow didn't sound right to me. And, just to complicate thuings a bit more, my loudspeakers neasure as an exceptionally easy load, so they can't be blamed for significant differences.
But the opposihe has also happened to me when an unassimung integrated amp had exactly 3 items changed inside that made it behave in an unexpectedly better way, as if it that was a magic formula of how to modify a standard mass produced item into a de luxe sounding item in three easy moves. It's measurements remained the same but its sound improved like it took me months to really believe was even possible. And given my experience with amps, as well as my very solid collection of them from various sources and price points, I am not easy to impress.
Simon, do you think that would be so if those amps were initially designed as wide open loop bandwidth models? Your opinion is just fine.
There is small signal bandwidth and large signal bandwidth. Many professional amplifiers limit slew rate which limits the large signal bandwidth to under 40,000 Hz. Small signal bandwidth would be at least three times that.
Hi fidelity reproduction amplifiers should have less than 5 degrees of phase shift at 20,000 Hz. So an amplifier meeting that would have no problems with a bit of high frequency leaking into the signal. However as the leakage get into 5 MHz. or more it may cause problems.
Last edited:
Agreed, and I note that H/K for example gives the said spec as a matter of course, and it is indeed <5 degrees suqare wave tilt at 20 kHz.
What they and some other manufacturers (e.g. Studer/reVox, etc) do is to extend the closed loop response to above 400 kHz, but then limit the rate of change by means of an input filter ao 200 kHz or so. Andrew did that on his 200W amp by using a filter inside starting from 300 kHz and upwards. Which again will limit the amp at around that frequency. There's nothing you want up there anyway, just trouble. And, assuming modern BJT output devices which easily hit 30 MHz and above, if you don't watch out there can be trouble.
What they and some other manufacturers (e.g. Studer/reVox, etc) do is to extend the closed loop response to above 400 kHz, but then limit the rate of change by means of an input filter ao 200 kHz or so. Andrew did that on his 200W amp by using a filter inside starting from 300 kHz and upwards. Which again will limit the amp at around that frequency. There's nothing you want up there anyway, just trouble. And, assuming modern BJT output devices which easily hit 30 MHz and above, if you don't watch out there can be trouble.
yes, what PMA said is exactly what one needs to do....However, I would not say people design poor amps due to lack of design capability/ability. Rather they design for the stated inputsignal/conditions.... which rarely, if ever, includes HF from digital sources. That is just missing information. That is reinforced by testing results using a 30KHz BW filter on the DUT. But being aware that you may frequently have HF coming in as well as audio means you filter that input OR have really low distortion to HF freqs.
Without input filtering, designs with low HF distortion sound better when the input signal includes HF.
THx-RNMarsh
Last edited:
Very old news, but relatively uncommon.
Show me the measurements.
That's a joke. I've been working with digital gear since 1968, formally trained as an IBM field engineer.
Like you said, less.
Measurements are relatively abstract, and does't mean that there is necessarily and audible problem.
Of course not, and its not my issue.
I love people who think that they can accurately read other people's mind, expecially for the purpose of being gratuitously insulting.
This hardly makes you unique. For example I was a professional recordist of live performances for over a decade just before I retired.
Since my real world identity has been discussed here in the past few days including a photograph, I can't imagine where one might get this idea.
To help alleviate any possible confusion, the guy on the right is me:
You can read more details here: The Great Debate...and Then Some | Stereophile.com
"5 degrees at 20 kHz" sounds like a "order of magnitude platitude"
Psychoacoustics suggests it takes a lot more to be DBT ABX at 20 kHz
an interesting factoid is that Sony changed the SACD "Scarlet Book" reconstruction filter corner downwards from their original 100 kHz to 50 kHz - reputedly because some "Audiophile" amps were provoked into self destruction by the rising 5th order shaped noise above 20 kHz
and I highly recommend against 100 kHz clipping test of most amps "at home" - likely to let out the magic smoke by burning up any output Zobel - they are seldom(never) rated for continuous full amp power
kind of a "chicken or egg" question - if the Zobel resistor loading stabilizes the amp against 100 kHz+ oscillation then it doesn't need to be sized to dissipate the power
Psychoacoustics suggests it takes a lot more to be DBT ABX at 20 kHz
an interesting factoid is that Sony changed the SACD "Scarlet Book" reconstruction filter corner downwards from their original 100 kHz to 50 kHz - reputedly because some "Audiophile" amps were provoked into self destruction by the rising 5th order shaped noise above 20 kHz
and I highly recommend against 100 kHz clipping test of most amps "at home" - likely to let out the magic smoke by burning up any output Zobel - they are seldom(never) rated for continuous full amp power
kind of a "chicken or egg" question - if the Zobel resistor loading stabilizes the amp against 100 kHz+ oscillation then it doesn't need to be sized to dissipate the power
Last edited:
5 degrees 20 kHz is from both Manfred Schroeder and Rupert Neve reports of their tests. If you note that is enough to change the peak amplitude of many waveforms by more than several dB that follows many hearing models for detectable thresholds. However Schroeder used complex tones that had the same amplitude. I have not been able to replicate those results.
But if you look at high end audio transformers they go for the high bandwidth and low phase shift. As it is easier to accomplish in electronics there really are few good reasons to set the bar lower.
But if you look at high end audio transformers they go for the high bandwidth and low phase shift. As it is easier to accomplish in electronics there really are few good reasons to set the bar lower.
Wow an IBM field tech! Forgive me if I was not clear enough to only be gratuitously insulting.
In college there was a dorm floor where they really had a ******* contest for distance. The result was a disgusting sticky floor, to say the least.
If you ever have any valid measurements or tests and want to discuss things with out subterfuge maybe folks would pay some attention.
And I was pretty sure you have never done the real work or you would know about the special case of that music type.
Last edited:
- Status
- Not open for further replies.