If you think of an old 16bit dac, you are probably right, if the volume control is badly implemented. For a modern 24 or 32 bit chip it looks different. If the wordlenght is 24 bit, you have 8 bits to give away with a 16bit CD, the same for a 24bit download a 32 bit dac.
One bit means 6db, you can attenuate to (8 times 6db) 48 db, lossless.
And there is more. Read the very interesting
"White paper on digital level control"
by Daniel Weiss
ComputerPlayback
Nevertheless I still use analog level control, because of my vinyl....
Andre
It's simply impossible to match such components and expect any constant matching on the time. Its own physical principle wouldn't allow such a behavior... Unfortunately, and you can easily verify this by constantly measuring the values for different attenuation ratios.
It's well known that some "special distortion" will even "improve the audio quality perceived" this is used in several amps, like some AKSA models and other... Then, even distortion can be "tolerated" or moreover "subjectively improve audio", this is not acceptable at least for my usage... And for obvious scientific reasons !
We built a very large panel of attenuators from very expensive pots, to the LDRs controlled by ADCs going through TVC and several other solutions. The last more reliable, distortion free, and musically totally transparent attenuators are very high precision resistors with relays, period.
We probably are not speaking of the "same distortion"... For example, with a very high feedback loop, you could reach very low levels of harmonic distortion (even lower than those of your AP device...). There are several types of distortion. Some are more important than other. But anyway, it is preferable to maintain ALL types of distortion as low as possible !
Unfortunately, it would not work in every case, if you are in balanced mode with long links between the preamp (attenuator) and your active monitors or your amp have insufficient input impedance, etc. you'll face some difficulties... And finally, the argument of listen fatigue is not acceptable, since I know a lot of gear with this facility and high distortion, again, even physiologically acceptable this is not the right way : but I respect every one deciding to follow it. This is just my personal result after several years of "perfect attenuation" research. No more, no less.The “Lightpeed” is ridiculously simple, ridiculously inexpensive, works very well indeed and does not cause any listening fatigue.
ondesx:
Have you listened to a Lightspeed or a clone? What were the test conditions? Sources, power amps, speakers, source material, etc.? What sonically did you observe as the defecit when comparing your LDR attenuator to the product of your "perfect attenuation research"?
When you performed your testing of the LDRs, when you observed variations from test to test, what were your lab conditions? Did you check each LDR at the same temperature each time? George has repeatedly stated the importance of potting the LDRs for thermal tracking between them. Is this what you did? Did you select and match a quad, or a pair, then pot them and re-test at a later time at the same temperature, or did you note the temperature change? Air blowing on the LDRs affects their resistance. Did you encase the LDRs when testing to assure that was not the cause of the variation?
A common claim made in this hobby is the notion of a "distortion" which makes the listening experience more pleasant. That concept used to deride both the individuals using the distorted technology as well as the equipment: i.e., "You guys just like the sound of this distortion without realizing that it's a distortion that you like. We enlightened people don't fall prey to that fallacy." From my own point of view, improved clarity, resolution of low-level detail, a stable "sound field", reproduction of high frequencies without an added "edge" (which is not present in live, unamplified music), bass without boominess, but with excellent definition do not count as distortions.
Whatever your ideal stepped attenuator form is, will you be willing to set up a demonstration, comparing it to the [euphonically distorted?] Lightspeed, or a clone? I'm sure some participants in this thread must be near enough to you to make such a comparison a reality. You could really show us something.
Stuart
Have you listened to a Lightspeed or a clone? What were the test conditions? Sources, power amps, speakers, source material, etc.? What sonically did you observe as the defecit when comparing your LDR attenuator to the product of your "perfect attenuation research"?
When you performed your testing of the LDRs, when you observed variations from test to test, what were your lab conditions? Did you check each LDR at the same temperature each time? George has repeatedly stated the importance of potting the LDRs for thermal tracking between them. Is this what you did? Did you select and match a quad, or a pair, then pot them and re-test at a later time at the same temperature, or did you note the temperature change? Air blowing on the LDRs affects their resistance. Did you encase the LDRs when testing to assure that was not the cause of the variation?
A common claim made in this hobby is the notion of a "distortion" which makes the listening experience more pleasant. That concept used to deride both the individuals using the distorted technology as well as the equipment: i.e., "You guys just like the sound of this distortion without realizing that it's a distortion that you like. We enlightened people don't fall prey to that fallacy." From my own point of view, improved clarity, resolution of low-level detail, a stable "sound field", reproduction of high frequencies without an added "edge" (which is not present in live, unamplified music), bass without boominess, but with excellent definition do not count as distortions.
Whatever your ideal stepped attenuator form is, will you be willing to set up a demonstration, comparing it to the [euphonically distorted?] Lightspeed, or a clone? I'm sure some participants in this thread must be near enough to you to make such a comparison a reality. You could really show us something.
Stuart
We build several models of this attenuator even a balanced version and an auto-calibrated one... For the measurements, we used an AP. For the subjective audio comparison, I'm afraid we only have our ears... But this is the same for everyone isn't ? Unfortunately, the distortion observed with the LDRs was expected since it's one parameter of the component you can easily retrieve in the DS...
I think the measurement condition are standard. Casing or not for an iterative measurement of the currents needed for a given attenuation show variable values what ever you do. Moreover, these components work on a statistical basis and by construction don't have a correct linearity. With the calibration we have observed that whatever the time used for doing it (from one hour to one day...) the audio results AND measurements are more or less... the same !
Unfortunately we aren't speaking of the same things... Subjective considerations are one point, but have nothing to do with scientific measured parameters. Both ways of development are respectable. I just recall the drawbacks of the LDRs and give here our own experience after a couple of years playing with them... It's normally the way a forum will work. Isn't ?
I'm afraid most of the participants are far from me. I'm in France as shown by the banner... Don't misunderstand my input. I'm not saying that the audio result with LDRs is "bad", but it's just better doing another way...
I explained that in some situations (for example low input impedance of the amps) the audio results will be poor. And as a direct consequence this preamp wouldn't give pristine result without a good buffer... But anyway, with the figures of distortion observed with the LDRs I couldn't derive a high quality audio preamp... Again, this is my personal point of view. Obviously, everyone could have another opinion !
Well to illustrate this point I'll just use a very well known experience in the visual domain. If you take two squares of the same color or the same gray level and included each one in a different environment (one in a darker one and the other in a lighter one) you can have "subjectively" the illusion that these two squares are of a totally different color...
But the truth is that they are of THE SAME color whatever you are observing !
Scientific colorimetry measurement of the squares will show that both are of the same color or gray level, even if nobody will believe such an evidence...
what about the linearity of the mentioned circuitry, especially for larger amplitude (say >1V)
This is definitely an unacceptable behavior in a decent and well administrated forum. I ask the moderators to recall to you the rules of this forum. I don't try to prove any superiority, but I report here our own results experimenting with this component.
You may have different results and in this case you can report here your own experience, but please avoid in the future sensitive posts like this one.
I posted mine here...
Here...http://www.diyaudio.com/forums/analog-line-level/176847-warpspeed-optocoupler-volume-control.html
Can you explain the specifics? What have you taken into account in the various designs? I have listened to some commercially available LDR, they certainly are much more clean and detailed in sound lack of hash that normally is associated with pressure contacts.
Thought digging up the past would be interesting. It was...
A few quotes from a very satisfied LDR listener:
Probably the most intersting in light of more recent posts is this:
Finally and as recent as the last 24 hours is this
A few quotes from a very satisfied LDR listener:
Probably the most intersting in light of more recent posts is this:
Stop the match please ! Everyone interested in the best solution must do the comparison by himself...
It's very easy to do btw. You just have to use a couple of CDs (same of course) and fix a listening level (of course again you need some stuff to work with, here a sound level meter will be useful). Then you have to measure the series value of the LDRs. You can do two experiments :
- the first is to replace the "free" LDR with a LDR fed by the current giving the "same" series value.
- the second is to replace the series LDRs by the equivalent resistors (you can use the highest precision you may find).
Thats all folks !
If you guys still didn't hear any difference, stop high quality audio and buy any Chinese low cost integrated amp after go drink all the saved money !...
My Very Best to you all,
Thank you James. I did a try before going to bed yesterday night with a very simple operational amplifier mounted in follower mode (a couple of LME 49720) and the observed phenomenon at low level remain... I concluded that there isn't an impedance issue !
Moreover, the impact and intensity of the lows are just amazing, that wouldn't be the case with an impedance mismatch, or this mismatch just appears for low levels ?
Perhaps somebody knows what happens if one use a too high or a too low power supply ? Anyone tried batteries instead ? With a charger, it'll be perhaps the best solution isn't ?
I'll try with different PSU now and we'll see...
For an enough high level, all is pretty good, the sound is just incredible, better than with my third opus of auto-transformer (which still reached a very very good audio quality indeed, much better than most active preamps, even some swiss made costly gear...). We have more presence, transparency, the sound stage is very large, the voices are more natural, with a very good stability for high frequency at high levels, the dark side is the mixing errors are more evident too !...
Enclosed a picture of my first balanced LS opus !
Finally and as recent as the last 24 hours is this
Last edited: