A usefully intelligent discussion by Digital Audio Review on the topic of digital volume control and the (non) use of traditional preamplifiers.
http://www.digitalaudioreview.net/2013/12/kih-3-can-digital-volume-controls-keep-it-honest/
Thanks. Between the source, the amp, and the speakers there is a lot variation out there in terms of system matching, gain etc. Add to that user preferences, room sizes, domestic circumstances, and it is hard to see how there is a one-size-fits-all response to this issue. I run direct to an integrated b/c I cannot use the PWD much above 30, so definitely into a truncation issue. 50 W of high quality, paralleled SET into benign, 93db load can produce a hell of a lot of sound.
If there were an option of bringing the output of the PWD down by half, without sonic compromise (the big IF), that would be my preferred option. Weiss might be onto something…
Weiss, Wadia and others have long had that ability to set analog gain to various levels for the purpose of system integration.
A number of us have pushed this as desired feature on the PWD.
Yes sounds good to me too. One of my headphones amp I need it turned up all the way while mid to 3/4 . So I guess it’s fine for me.
Al. D
I found it interesting that the PWD II is certainly not alone in having an output that is “too hot”, and this seems to be an industry wide issue. Also interesting was Bel Canto reportedly looking at 64-bit processing to solve the issue (will that work?). And lastly, the example quoted suggests that in a modest system, even a cheap-ish preamp can produce the type of improvements Paul has noticed in his system with the Aesthetix.
A 64-bit data path would be delicious overkill, with over 240dB attenuation available prior to 24-bit data truncation.
A good friend has a MSB stack and he paid about 3k for some kind of stepped attenuated volume control .
I cannot imagine what that is.
Al. D
@stereophilus The ultimate gain of the PWD is determined by the gain of its output stage which we could set anywhere we want. I chose the gain we have because for our system and for most systems with standard power amp gains of 26 to 28, standard speaker efficiencies of 87 to 90dB, a Reference Recording plays at the proper level between 70 and 80, giving you a bit more room to go if you have less efficient speakers or a less efficient amp.
It’s never easy find a perfect gain for everyone.
Having said that, a pure DSD engine could theoretically have a digital volume control without any drawbacks at any level or incoming sample rate. The 64 bit solution that Bel Canto is contemplating is an interesting approach for PCM.
Weiss, Wadia and others have long had that ability to set analog gain to various levels for the purpose of system integration.
A number of us have pushed this as desired feature on the PWD.
And Paul continues to reject the idea as this would produce a compromised output for the PWD. Paul continues to strive for perfection in the PerfectWave series. This is a valid argument with a perfectly matched system - source level, amp gain, speaker sensitivity and room acoustics. The problem with this argument is that such a perfectly matched system is rare. A significant percentage, if not the majority, are not so perfectly matched. Many systems seem to require some gain reduction between the PWD and speakers in order for the PWD volume setting to be above 50 for anything other than a loud listening level.
Somehow I suspect that a properly designed and integrated output level adjustment would have only a subtle effect. Perhaps a tiny loss in the perfectly matched system, but probably an improvement in many if no most systems.
If we keep pushing, perhaps the Mk-III version will have a switch selected 20dB level reduction on the output to allow for better matching with almost all systems. With this addition most systems would have normal playback levels above 50 on the PWD volume scale. That combined with an output that does not require running the signal through a preamp to sound its best would be spectacular! Back to Paul's mantra of 'the best preamp is no preamp'.
J.P.
What timing… Paul posts while I am composing my own. 
J.P.
After having no preamp in my system for a while, then getting the preamp back (with a slightly different volume control revision) I can really say that it is quite beneficial to have the PWD at full volume, at least in my system.
One of the points in the article is that digital domain volume attenuation still has a way to go. Perhaps even 64-bit implementation will cause some “loss” - who knows. Analog volume attenuation is not perfect, but the there are far fewer (if any) unknowns. One school of thought might be that a new preamp/DAC includes an analog volume attenuation device, rather than digital. I know this has not been the PS Audio direction for a while now, but then (sorry to remind you Paul) there was the preamp crow eating exercise…
J.P. +1 for output gain switch 
I tried the Pwd mkii as a preamp alone , but to be honest it sounded fine . But I kinda got spooked with it having total control of a couple of thousand watts , and I now have a preamp after it. I just had recurring images of the speakers flying back into walls behind them. Lol. Not that it ever did anything wrong and does look as good as it sounds. Thanks ps audio. Happy holidays.
Having said that, a pure DSD engine could theoretically have a digital volume control without any drawbacks at any level or incoming sample rate. The 64 bit solution that Bel Canto is contemplating is an interesting approach for PCM.
Here is something to consider for those running Wolfson chips:
http://www.computeraudiophile.com/f22-networking-networked-audio-and-streaming/lumin-audiophile-network-music-player-14280/index5.html#post198574
I’m sure this is question that might have been answered some where else , but if digital volume is a possible problem , what is wrong with analog volume . I have a krell preamp amp I use it for headphones. I’m pretty sure it has a digital volume control. As the know just goes around with no stoping point , is this bad for the sound as I cannot hear any chNges other than the volume . I can through a attenuation done with resisters on the output control where my normal listening position on the dial is. I go for around mid as 0 to 150 changes on the readout . Is this ok?
Al. D
A digital volume control would probably be just fine if we had even just a true 32 bit DAC including true 192dB dynamic range AND S/N. That would give us a lossless 96dB reduction range for 16 bit files and 48dB for 24 bit files. System wide level matching would be less of an issue. Since 24 bit encompasses more dynamic range than our ears and even the best pres and amps can support then a true 32 bit DAC with even just 32 bit digital volume control should suffice for almost any system we can assemble.
Since few DACs can manage better than 20 bit S/N (120dB) and very few systems can even manage 20 bit dynamic range end to end and few are level matched from source to speakers then we need some sort of analog level adjustment somewhere in the signal chain to extract the maximum possible resolution from our systems.
Oh, would it be if we had true 32 bit DACs and perfectly level matched systems… Life would be easier and no doubt sound incredible!
It is great that Paul is matching the gain of the PerfectWave Amp to the output level of the PerfectWave DAC. Unfortunately that does not account for speakers of various sensitivities, much less using something other than all PerfectWave electronics in the signal path. Some sort of level adjustment on the output of the PWD and/or input of the PWA would make them easier to integrate into various systems with better system signal quality overall, even just a switched 10 or 20dB gain choice…
I will continue to believe in and present such concepts until the idea is acknowledged or proven invalid. Even then…
J.P.
As an aside, it isn’t uncommon for headphone amps to have several switched amounts of gain. This makes signal matching much easier as the ‘goal’ is to always keep the (analog) volume control at the 12 o’clock position (≈± 2hrs) for best channel balance…
In my setup I need to ‘burn off’ ≈ 3-12dB (depending upon how hot the source is) just to keep the volume control in its ‘sweet spot’. I only wish the MC network player would allow me to do this, instead of having to use the PWD.
Alas, maybe when wavestream becomes available.
JJ
@johnjen that video was fascinating. Very useful to the topic here. I am left feeling somewhat pessimistic about keeping my DAC volume at 50 despite the internal chip architecture being 32-bit. It would seem that the SN ratio suffers progressively with each dB of digital domain attenuation… Gonna have to review those figures a few times! :-B