restorer-john
Grand Contributor
I started this thread so as to not derail the Yamaha R-N803 review thread.
There's a lot of people who think the "direct", "pure direct", "CD direct", buttons guarantee a pure analog experience. Often, this is not the case. Sure, hopefully, the manufacturers will bypass an A/D-D/A stage if you are lucky, but often the poor little signal still has to stumble through a labyrinthine series of semiconductor switches and is still subject to either a digitally controlled VCA or stepped active attenuator and several gain/buffer stages before it gets to the power amplifier. It's hardly a pure path.
We see people saying things like this (no criticism intended):
People also might draw nice direct lines on a signal path, but remember, that is merely a simple block diagram of the IC and shows nothing of the actual active componentry in the IC itself.
What should considered with these so called "pure" signal paths is the fact they are usually capacitively coupled before going into the IC and again at the output. So we have RC time constants all along the path and consequent non-linearities in the lowest octaves. They also use cmos bilateral or equivalent switches internally, with their associated R losses and "feature" one or more gain stages to offset those losses along with a pretty ordinary VCA (voltage controlled amplifier) and yet another buffer stage on the output of the IC. Phew.
Another problem with IC input switching is they often perform well at high levels (~2V) but also overload easily. Yamaha for instance have had a series of issues with their amplifiers overloading with high level inputs (>2V). I have a few in my collection with "CD direct" active buffer stages that overload with signals slightly above 2V. The input switch array ICs used also exhibit skyrocketing distortion at low levels. As much testing is done at high input levels, people are not aware of how bad some of these "soft" input switch/volume arrays actually are.
The only two benefits I see is are definite improvements in L/R channel tracking with digitally controlled volume implementations and fewer contact related issues down the track. Channel tracking is where a traditional pot cannot compete. Take the current Yamaha A-Sxxx series. They use a nice fat Alps motorized volume pot, but only use one gang to drive an A/D converter on the main uP, which in turn digitally controls the volume in the input switching IC! What a joke. All that money for an "audiophile" motorized dual gang Alps pot which may as well be an up/down button or a rotary encoder. But they know audiophiles love a proper knob. They are taking those audiophiles for fools.
Electronic input switching is also fragile. I have lost count of the number of failed/intermittent input multiplexing ICs I have had to source and replace. Often gear is written off due to bespoke parts not being available. (Yamaha I'm looking at you). They are also subject to way more transient events due to all the sources being plugged into them. In my opinion, these ICs are a poor choice. But they are extremely low cost and manufacturers can use last year's obsolete AVR ICs in this year's "audiophile" 2 channel offerings.
They know a single "pure direct" button will sell more products, regardless of how it is implemented. In reality, they have thrown the 2 channel pure/direct concept into the trash.
This is a 1982 Sansui AU-Dx series direct signal path. Notice there is not a single capacitor, active buffer or anything in the signal path from a line source (tuner/aux) right through the power amplifier to the speaker. This amplifier had a rated signal to noise of 110dB, for a 65wpc integrated amplifier, nearly 40 years ago.
The power stage is a high gain design with a FET front end, able to dispense with cap coupling and the consequent frequency response was specified as:
Power amplifier stage:
Harman Kardon pursued similar topologies through the 1980s with their fabulous PM-xxx range of amplifiers. Most "main direct" functions were truly straight from the input socket to a volume pot and into the power amplifier- nothing in between.
I have many vintage amplifiers with absolutely straight "pure/direct" connections from the input sockets to the main volume pot and directly into the power stage. Preamplifiers with a bypass function that routes a single source directly to the volume pot and a single buffer stage. Pretty much all of them exhibit a significantly lower noise floor, better frequency response and fewer non-linearities in general.
It seems at one end of the audiophile spectrum we have a March audio style power amp/buffer and direct source using digital attenuation or you have commercial offerings with dubious "direct" signal paths offering a very average journey for the fragile signal. The products of the past that combined both general approaches into one unit seem to have unfortunately disappeared in all but the absolute TOTL models.
What is your experiences and can you tell the difference?
There's a lot of people who think the "direct", "pure direct", "CD direct", buttons guarantee a pure analog experience. Often, this is not the case. Sure, hopefully, the manufacturers will bypass an A/D-D/A stage if you are lucky, but often the poor little signal still has to stumble through a labyrinthine series of semiconductor switches and is still subject to either a digitally controlled VCA or stepped active attenuator and several gain/buffer stages before it gets to the power amplifier. It's hardly a pure path.
We see people saying things like this (no criticism intended):
I've marked pure analog lines with red, so that it is clearly seen the route is pure analog when DIRECT mode is chosen on RX-V583.
People also might draw nice direct lines on a signal path, but remember, that is merely a simple block diagram of the IC and shows nothing of the actual active componentry in the IC itself.
What should considered with these so called "pure" signal paths is the fact they are usually capacitively coupled before going into the IC and again at the output. So we have RC time constants all along the path and consequent non-linearities in the lowest octaves. They also use cmos bilateral or equivalent switches internally, with their associated R losses and "feature" one or more gain stages to offset those losses along with a pretty ordinary VCA (voltage controlled amplifier) and yet another buffer stage on the output of the IC. Phew.
Another problem with IC input switching is they often perform well at high levels (~2V) but also overload easily. Yamaha for instance have had a series of issues with their amplifiers overloading with high level inputs (>2V). I have a few in my collection with "CD direct" active buffer stages that overload with signals slightly above 2V. The input switch array ICs used also exhibit skyrocketing distortion at low levels. As much testing is done at high input levels, people are not aware of how bad some of these "soft" input switch/volume arrays actually are.
The only two benefits I see is are definite improvements in L/R channel tracking with digitally controlled volume implementations and fewer contact related issues down the track. Channel tracking is where a traditional pot cannot compete. Take the current Yamaha A-Sxxx series. They use a nice fat Alps motorized volume pot, but only use one gang to drive an A/D converter on the main uP, which in turn digitally controls the volume in the input switching IC! What a joke. All that money for an "audiophile" motorized dual gang Alps pot which may as well be an up/down button or a rotary encoder. But they know audiophiles love a proper knob. They are taking those audiophiles for fools.
Electronic input switching is also fragile. I have lost count of the number of failed/intermittent input multiplexing ICs I have had to source and replace. Often gear is written off due to bespoke parts not being available. (Yamaha I'm looking at you). They are also subject to way more transient events due to all the sources being plugged into them. In my opinion, these ICs are a poor choice. But they are extremely low cost and manufacturers can use last year's obsolete AVR ICs in this year's "audiophile" 2 channel offerings.
They know a single "pure direct" button will sell more products, regardless of how it is implemented. In reality, they have thrown the 2 channel pure/direct concept into the trash.
This is a 1982 Sansui AU-Dx series direct signal path. Notice there is not a single capacitor, active buffer or anything in the signal path from a line source (tuner/aux) right through the power amplifier to the speaker. This amplifier had a rated signal to noise of 110dB, for a 65wpc integrated amplifier, nearly 40 years ago.
The power stage is a high gain design with a FET front end, able to dispense with cap coupling and the consequent frequency response was specified as:
Power amplifier stage:
Harman Kardon pursued similar topologies through the 1980s with their fabulous PM-xxx range of amplifiers. Most "main direct" functions were truly straight from the input socket to a volume pot and into the power amplifier- nothing in between.
I have many vintage amplifiers with absolutely straight "pure/direct" connections from the input sockets to the main volume pot and directly into the power stage. Preamplifiers with a bypass function that routes a single source directly to the volume pot and a single buffer stage. Pretty much all of them exhibit a significantly lower noise floor, better frequency response and fewer non-linearities in general.
It seems at one end of the audiophile spectrum we have a March audio style power amp/buffer and direct source using digital attenuation or you have commercial offerings with dubious "direct" signal paths offering a very average journey for the fragile signal. The products of the past that combined both general approaches into one unit seem to have unfortunately disappeared in all but the absolute TOTL models.
What is your experiences and can you tell the difference?