I haven’t reached out to Gryphon or Vitus, the latter being another amplifier brand I own and with which I’m experiencing similar issues, regarding the use of higher than recommended input voltages. I expect that their advice would be cautious, particularly with the integration of third-party equipment with their products. A Google search revealed that Vitus once recommended a 2V preout for a power amp designed for 2V input sensitivity. However, I was not satisfied with the results when I tried this setting myself.
In examining the unique design of high-output voltage preamps and the low-input sensitivity of power amplifiers, Vitus and Gryphon show similar characteristics. For example, the Gryphon amplifier has an input sensitivity of 0.564V, yet their corresponding preamp can output up to 23V. Similarly, the Vitus amplifier, designed for a 2V input sensitivity, has a preamp capable of reaching a maximum output of 20V. This design suggests that their power amps may be optimized for higher input voltages, but it also raises concerns about compatibility when used with equipment from different brands.
For enthusiasts like myself, keen on further exploration, a methodical trial-and-error approach is the most viable. This process involves gradually increasing the output voltage from the DAC and diligently monitoring for any signs of distortion or clipping. Finding the optimal setting is a meticulous and challenging task, but it’s crucial for achieving the best possible sound quality.
I know what you mean. I go direct from my Vivaldi stack with 2V to my Tenor Amp which has a sensitivity of 1.3V. However, occasionally I feel the need to push the dCS up to 6V for an extra kick.
There’s a downside to that though, superficially the 6V sounds punchier, but I once measured my speakers with both the 2V and 6V settings; with the volume knob to level things, the necessary digital attenuation being higher for the 6V setting, dynamics seems to suffer measurably.
I posted about it back in the “6V vs 2V” thread from 2020.
Yes, you’re right in your observation. From my understanding, if the input stage of the amplifier is fed a voltage higher than it is designed to handle, it can lead to an overload situation. This can cause the amplifier’s input circuitry to clip the signal, reducing dynamic range and causing the sound to appear compressed.
Amplifiers have a certain amount of headroom, which is the ability to handle peaks above the average level without distortion. If the average level is too high (due to a high input voltage), the peaks might push the amplifier beyond its headroom, leading to clipping and dynamic compression.
Matching components from different brands can indeed be complex and sometimes challenging.
At the time, I didn’t really dig deeper, but the reduced dynamics it was not from the Amplifier clipping. If you read my post, you’ll see there was 9dB of headroom.
So, I guess the compressed dynamics are likely a result of the amplifier’s intrinsic response to varying signal levels, particularly considering the Tenor’s zero global feedback design, rather than traditional clipping or a lack of headroom.
I’ve been trying to understand what the actual performance is of my Bartok as a pre into my Bryston 3B3. (Low setting is rated at 100W for 2V 200W at 3.4V). I Listen in near field at .2V out from the Bartok at 0dB for most chamber music and .6V @-5-0dB for orchestral. For some older very dynamic recordings I occasionally bump up to 2V out at -13 to -5dB.
Subjectively, if you asked me what was better: 0dB @ .2V out, -10dB @ .6V out, or -20dB @ 2V out; I’d have a hard time identifying which is better.
However Objectify, I would like to know which volume setting actually measures better.
Long ago when I had this conundrum of 2V vs 6V, I reached out to dCS to find out that the the DAC’s design’s default out is 6V and to bring it down to 2V something is applied. And since I have stuck with the rule of staying with defaults with any piece in my system, wherever possible, I live with 6V. Just sharing my thought process.
Just to point out that the line level output voltage is not a volume setting. Imagine that the output of the two settings fed passive 8 ohm resistors. Each setting will reveal a slightly different measurement. However in both cases the volume will be zero.
As a generalisation the higher output voltage will yield a slightly better measured noise figure. However as the noise figure for the lower voltage with dCS units is inaudible anyway the slight improvement is likely to be insignificant. Coincidentally Hi-Fi news has just published a review of Lina 2.0 ( not AI ). S/N ratios from their laboratory test are 109.1 dB @2V and 117.0 @6v. At bearable listening levels if you any difference on this parameter alone you have , not golden ears, but platinum ones with diamond detailing in a custom design from David Hockney.
This is an interesting example. With the Bryston set at either 23dB or 29dB gain, all three settings (0dB at 0.2v, -10dB at 0.6v, and -20dB at 2v) indeed produce the same output voltage, resulting in similar perceived loudness. However, theoretically, the final perceived sound quality, which depends on factors such as frequency response, distortion, dynamic range, and noise floor at these three settings, cannot be identical. Practically speaking, I’m not sure if I can distinguish between these settings when listening to them.
In a dCS DAC that features separate controls for volume and output voltage, it seems that these two components contribute differently to sound quality. The dCS employs advanced techniques in its digital volume control to minimize the impact on effective bit depth and reduce quantization errors at lower volume levels. However, I believe maintaining a higher digital volume level (closer to 0 dB) is generally better for preserving the resolution and dynamic range of the digital signal.
As for the output voltage, a higher level can improve the interaction with the amplifier, enhancing both the signal-to-noise ratio and potentially the amplifier’s performance. Yet, in this scenario, the three settings - 0dB at 0.2v, -10dB at 0.6v, and -20dB at 2v - should theoretically produce the same voltage output. Thus, the differences in how the DAC’s output voltage interacts with the power amp, and any impact on the amp’s performance, are effectively neutralized as variables.
This leads to the conclusion that the primary factor affecting sound quality is the digital volume control. The critical aspect, therefore, becomes how well the digital signal is preserved before its conversion to analog, a function governed by the digital volume control.
Given these points, opting for a higher volume setting on the dCS DAC seems logical, as it should theoretically yield a purer and more detailed representation of the audio signal. Of course, this is an educated hypothesis in the absence of actual measurements, and the real impact may vary.
I use my Rossini APEX direct into MBL monoblocks. The output voltage is set to whatever allows me to keep the output between 0 and - 10 on the Rossini. Usually 0.2v but when having a wee headbang at 2v or even 6v at times.
Was at 0 and 2v today listening to some Reference Recordings (the Eiji Oue Stravinsky) with a DR of 25. Much fun.
It’s probably obvious to everyone here, but when the DAC is plugged in direct to the amp and say the speakers are playing at 70dB the voltage coming out of the DAC is the same regardless what output voltage on the DAC is selected. However the digital volume will be different. The ideal voltage output may requiring changing depending on the source. I listen via streaming from the upsampler and via toslink from a Sony TV. Because the TV is much louder than the stream I use 0.2 volts for TV and 0.6 or higher for streaming. For example, if I use 0.2 output for both TV and streaming then with TV volume control is set to -5 to 0 for a comfortable level. But stream would be -20 for the same speaker output. It would be great if DCS allowed saving different output levels per input. If this isn’t possible in the hardware it is possible for mosaic to send multiple commands, much like the Logitech Harmony remotes.
That would be extremely valuable, especially for those of us driving our power amps directly. I had a scary moment before Christmas where one of the Strutslets inadvertently started streaming from Spotify (using Connect) at full volume which although it only lasted a split second is probably one of the loudest things I have ever been exposed to. Luckily neither my guests nor my speakers seemed to suffer any permanent damage much, I suspect, down to the incredibly undistorted power delivery of my Nord Three amp. I will add this to the feature requests thread.
Like many others, I use Rossini as a digital hub with various inputs and often encounter variations in loudness. Having an input gain adjustment for each input would be beneficial. From a technical standpoint, a global volume normalization feature might be a more straightforward solution compared to individual gain adjustments for each input. Additionally, even within the same input, such as when switching between different Netflix and YouTube content, there are noticeable variations in loudness. Therefore, a very useful feature would be the quick access to a few different preset volume levels.
I think I was slightly confused, so I deleted my earlier reply. After checking further on Wiki, I can confirm that global volume normalization and dynamic range compression manipulate audio levels in fundamentally different ways. Here’s a summary of what I read and understand: Volume normalization focuses on maintaining consistent loudness across different tracks without altering the internal dynamics of each track. In contrast, dynamic range compression aims to reduce the range between the loudest and softest parts within a track.
What I liked, a long time ago, while reading this thread, was watching the pictures of the stereo of the other members of this forum…But it was a long time ago…
). S/N ratios from their laboratory test are 109.1 dB 