Setting levels

August 11, 2019
 by Paul McGowan

Where we set our system-level can make a tremendous amount of difference.

One of the problems I often see in digital audio is people hell-bent on setting the volume level in the app they are streaming from. In almost every case, this is exactly the wrong thing to do. Changing the playback level at the source is a really bad way to reproduce high-performance audio. For example, if you’re using Audirvana, iTunes, or Bit Perfect, the last place you want to adjust the playback level is in those programs. The moment you do that any chance of bit-perfect performance flies right out the window.

The exceptions to this are when you’re using a program like Roon, or our upcoming program Octave. There, the levels can be adjusted right in the music management program because the source remains bit perfect. What you’re actually doing, in that case, is controlling the DAC itself through the interface. Thus, it looks like it’s happening right in the app when actually you’re doing it just right.

It’s a common mistake to make and one we see all the time. Your preamp or DAC is where you should be adjusting the level.

Keeping digital audio at its bit-perfect best is always going to sound best.

Subscribe to Paul's Posts

40 comments on “Setting levels”

      1. If the GDC reacts as the DS, the volume of the GDC changes with short delay simultaneously with the modified Jriver volume vice versa. So it doesn’t matter where you change the volume, it’s both optimal.

  1. I’ve always been on the lookout and avoided this problem. If you use an integrated streaming system like Linn there’s nothing to worry about. I then used Auralic Aries that had built in volume control, but they actually recommended you to disable it and control the volume pre-amplification. As Paul says, with Roon Ready devices the app is changing the volume in the receiving equipment and is one of its key features.

  2. Thanks for the tip, Paul. On my main system with its Arcam all-in-one there is a volume control, but I always thought it was linked to the app. On the desktop system the latest (but almost certainly not the final) renderer now has its own volume control, so I can play with the app, the renderer and the Sprout100 to see what gives the best results.

    1. I found with the Auralic Aries Mini into a Quad 909 using the dodgy Auralic software volume control that at under 50% the sound was almost unlistenable. As a temporary fix I got some Rothwell attenuators so I could run the volume control at about 80% and it was OK. Auralic improved the volume control in the software and I changed to an integrated amplifier. So it is not a black and white issue and it can be partially remedied, but with a lot of this software like jRiver and Audirvana you don’t have much idea what’s going on.

      1. After a bit of fiddling on the desktop system I found out what I had probably forgotten, that the volume settings on ‘app’ and renderer have no effect. I just have to do what I normally do, which is to adjust the volume using the Sprout100 remote. At least it saves on a lot of experimental optimizing.

  3. I had totally forgotten about this particular issue. I think early on I was forced to control source levels, perhaps back in the Logitech Squeezebox days or even what I might’ve had before (wow it’s been a long time ago I totally am forgetting all of the solutions I employed over the years to get my computer music to my sound system), when using my old vintage equipment that did not contain DAC or had no remote/app control. Now I have everything tweaked to the point where the system sounds on my computer are generated on the speakers built-in to my monitor while everything from JRiver or Roon goes out through the optical cable to my DAC and the software controls my kit. How things progress.

  4. I always found setting the volume level in the app very imprecise.
    One minute you almost hear nothing and the next minute, turning up the volume, the speakers almost explode if you are not very, very careful.
    So instead I always use(d) the volume control knob on my pre amp (after all I paid for this knob, so I’d better use it).
    And with 0,5 db steps it’s very precise and a breeze to work with.
    I’m a bit puzzled by the fact that some people seem(ed) to think the DAC/pre volume is controlled by the app.
    If you look(ed) at the volume setting on the DAC/pre you see nothing happens by turning up the app-volume.
    Or is my thought too simple and am I missing something ?
    So far I have no Roon experience. I’ll wait and see what Octave can do.

  5. I am old and easily confused, but I want to pay attention to this tip.

    So the answer on making the best choice in my system of Tidal app —> iPhone —> Sprout100 is what?

    Put the iPhone volume setting mid way, or full open, and leave it and adjust room volume off the Sprout100 volume knob?

    ASAP of an answer please before I forget I asked this.


    1. Well, sorry, but this doesn’t apply to Sprout or any product using Bluetooth. There, you’ve little choice but to use the volume on the device.

      I was referring more to the use of servers and computers feeding external DACs.

      1. Paul, I am currently playing music on my iMac via TIDAL/Audirvana with a USB connection to my stellar gain cell dad. What volume level should Audirvana be set at to maintain the bit perfect play back? Should it be set at 100%, which means the SGCD is playing at a relatively low volume? Or should the Audirvana volume be set to a mid-level, allowing the SGCD to play at a higher volume level? I guess my question is is the SGCD designed to give better performance at a low volume level or a higher volume level?

        1. Audirvana should be set to 100 and SGCD should then be the controlling entity. I know it can be a little confusing as I have said in the past you want your preamp volume level higher than 12 if possible. However, that preamp comment was meant for those preamps with potentiometers, which SGCD does not have one.

          SGCD uses a Gain Cell which is a variable gain amplifier and that will sound essentially the same at any volume setting.

          Better to have Audirvana at 100.

  6. Once you are in the analog domain, the rule of thumb is to adjust the gain at the earliest point to the highest level you’ll need without overloading the next stage. The reason is that this gives the best signal to noise ratio because any noise in the signal is amplified by later stages including residual or baseline noise. Another advantage is that with a conventional potentiometer this gives the lowest series resistance between the source and the input to the next stage. This resistance is added to the source resistance in analyzing the ratio of load resistance to source resistance which should be as high as possible. In the digital domain S/N ratio is probably so high to begin with that it doesn’t matter. What counts is that you don’t overload the DAC with the loudest signal. In the loudness wars exactly the opposite is done. The ADC is deliberately overloaded causing massive distortion. The mastering engineers don’t care, all they want is the loudest signal when you are hunting for a radio station to listen to. Given the state of music as I posted a link to yesterday it probably doesn’t matter. A fuzz box overloads the sound of a guitar deliberately causing massive harmonic distortion. Then it’s amplified to the point of being deafening while a pubescent voice shrieks two inches from a microphone to overload that too. To get the full benefit of the intent only the best state of the art reproducing equipment capable of ear shatter loudness without any inherent distortion of its own will do.

    1. Totally off topic today: check out Steve Gutenberg’s latest video. He and Andrew Jones discuss the future of sound reproduction, and I am pretty sure that you will enjoy it.

      1. I’ve been watching it. I’m partway through. I like Gutenberg. I don’t agree with anything he says but I do like him and find him … well… entertaining. But then I don’t agree with anyone ever. Not just here, EVERYWHERE 🙂

  7. I am frugal and have had good results with the free app Foobar2000 and use it to rip (to WAV files, getting metadata and using Accuraterip) and as a player, feeding my DAC via USB. I also enjoy using the Foobar2000 app on my phone to as a remote to select tracks from my listening chair (or anywhere else within WiFi range on my phone).

    I had been running the Foorbar volume control at maximum and using my preamp to control volume. Then I saw Paul’s post about a volume control’s sweet spot being between 10 and 2 – my system is very loud with Foobar at maximum volume and the preamp volume at 9 to 10 o’clock. So after reading Paul’s post I now run the preamp volume at 12 o’clock an control volume from the app, which is also nice because I can optimize the volume from my chair, since my preamp doesn’t have a remote.

    I think it sounds “better” (clearer, more impact?) with the preamp volume up and the Foobar volume down. Anyone else out there using Foobar?

    1. I agree with R. Murrison and what he wrote at 10:33. In my setup, FB running in 24 bit mode and Wasapi, there is no degradation of SQ down to about -15dB on the FB volume control. At -20 dB it is still ok for casual listening or podcasts. My amp potentiometer is at 10 normally, and at 12 if I need headroom up to 110 dbspl. This is rarely nessesary; only for some ECM classical editions, A. Pärt for example.

      1. Thank you. I was not familiar with Wasapi and will make sure it’s functioning and will also confirm 24 bit mode. -15dB is pretty far down on the range so I should be OK. Are you basing the -15dB comment on measuring or listening?

        1. On listening. Because I can not measure clarity and resolution;-). But I forgot to add that I use high efficiency speakers in this setup. They grant me 98dB with one watt/1m. So most jazz or pop is quite loud at -15dB and would be deafening at 0dB.
          P.S. Wasapi is a official FB add on which you can install in your components folder.

          1. …And in the extended sound settings of the OS choose 24/96 or 24/192 and check both boxes.
            This will bypass the MS mixer, volume and up/down sampling. Only FB’s vol will be functional. These are the basics for a halfway decent computer sound.
            There is a lot more to do to if you want to reach the level of a 3000$ to 4000$ CD player: PSU, 2nd internal HDD, cables (preferably good optical, a pity your sprout has only a coax input), ferrites, mains filtering and so on.

    2. Yep…Me. Love Foobar2000 but like yourself, I set my PC to “100”. And run it through USB into my SGCD. I’m gonna try setting my SGCD to “50” and set reduce the Win PC Volume to 1 or 2. Hopefully it fattens the the sound a bit. Currently USB playback of PCM/DSD2.8/5.6 is very thin sounding. I very rarely use my PC USB out /SGCD USB input. I access my DSD & PCM music files from my PC via the Sony UBP-X1000ES which is connect to one of the S/PDIF input. PCM sounds awesome this way and DSD playback is through the Sony’s internal True DSD DAC (up to 11.2MHz) and 5.1 DSD (only 2.8MHz) via the STR-1080’s internal DAC.

  8. I was under the impression that digital volume control must have access to the DACs internal data path for the best results, and that in lots of instances that option hasn’t been implemented. So I have always made a concerted effort to ensure that is the case in the DAC, especially if I plan on controlling the volume with an app.

    Does that sound correct or am I out in left field again?

  9. I’m afraid I don’t go along with that. While on the surface it makes sense, in reality, like a lot of audio , when you dig in deeper a different picture emerges.

    First, the concept of “Bit Perfect” playback. In principle it sounds like a fine objective. But in reality, as soon as that signal goes into virtually any real-world DAC it undergoes a plethora of digital processing stages far more complex and far more destructive than a simple, well-designed, volume control function in a source App. Virtually no playback chain is truly “Bit Perfect”.

    The term arose in the early days of computer-based audio when it was observed that when computers play back the audio signal it is usually routed through a mixer which is designed to meet the needs of a computer user, rather than an audiophile. Additionally, and most particularly using iTunes on the Mac, this could involve spurious sample rate conversion steps. And all of this processing is done using a low-quality DSP engine (again, because that’s good enough for a computer). To get around these constraints, software was developed which would route the audio data directly to the digital output without any spurious processing, and the term of art chosen was “Bit Perfect” to indicate that the bits delivered to the audio device were exactly those in the original audio file.

    In reality, the problem was not that fact that the audio signal was modified, but rather the nature of the modifications done. Simple operations like digital volume control can be done at the source with essentially zero loss of fidelity. Even highly complex operations like filtering, if done right, can be performed with no audible loss of sound quality.

    I think we have long since gone beyond the point where “Bit Perfect” data delivery is the gold standard. And I have no qualms about implementing some pretty serious signal processing at the source – with the proviso that it has to be done right.

    1. It would be interesting if you mean that, if nothing”s really bit perfect/identical within certain stages of a DAC and there’s just no ‘audible loss’ in the best case, the signal is just put together again to another good sounding alternative? How do you then define ‘lossless’ in terms of sound quality? Compared to what if we don’t know how the bits would have sounds when they went in?

      Man, then you’re destroying a big part of the digital argumentative bible of ages of digital vs. analog debates 😉

      1. It would take a lot of space to properly describe the inner workings of a DAC, and rather than go over it here I would refer you to a few of my columns in Copper Magazine. For a basic explanation of why modern DACs rely so heavily on digital processing, try my column in Copper 11. For a more detailed dive into the nature of this digital processing, refer to my columns in Copper 15, Copper 16, and Copper 17.

        1. Thanks Richard, I will dig a little deeper.

          At least without technical knowledge it was always hard to imagine how little gets lost in such complex processings and easier to realise that something must get lost, even if not immediately audible.

  10. Having no experience with streaming I really can’t comment on digital audio. But I do agree with the concept of setting the volume just so for good sound. Vinyl recordings do not have the complications different types of digital streaming has. What they do have is different levels of recording for different records which can be noticeably greater for some or softer for others. This require the volume to be turned up or down to achieve just the correct loudness. Here a remote really comes in handy. No fear of bits going haywire since none are involved. Life is so much simpler with results which can be so much better. Of course recordings on vinyl can be poor as they are on digital. Tone controls can rectify matters to some extent and sometimes to quite an extent. How well they work for streaming I do not know. Is digital still suffering from growing pains ? It certainly sounds very complicated. Regards.

  11. This comment touches tangentially what is discussed in this post, but touches it anyway.

    As a premise: there is a single comfortable volume for a person’s audio installation, which depends on their sensitivity, the type of music and of course the whole: amplification / speakers / acoustics of the room.

    Taking into account how extremely variable the recordings are, it is imposed that the volume control act in the sector closest to the source, in order to avoid the distortion that may occur later when the signal is being amplified.

    In my case, in one of the sound rooms, I have as a source: CD & laser discs, for the former, I use the AES / EBU outputs of Theta Jade transport using the Kimber TGDL that goes to the balanced digital input of the BenchMark DAC-1 (configured to use the variable outputs) from here the balanced analog signals, go to a pair of Crown Studio Reference One amplifiers, arranged in vertical bi-amplification that with its 20,000 damping one channel feeds the 3 woofers of 12 “and the other channel to the mid-hi section of the Legacy Focus, the potentiometers of the Crown are in full CW position so that they are practically out of the circuit.

    So I don’t use preamplifier. Why preamp a previously preamplified signal?

    Using a preamplifier in this case implies increasing the distortion without reason, or someone can argue that any audio circuit does NOT have distortion, in my case, enough with the distortion produced by the complicated passive Xover of the Focus.

    Only when a DAC has no volume control do I use a preamp when the power amplifiers do not have potentiometers.

    If someone who has a DAC with volume control (which generally acts within the digital domain) uses a preamp (supposedly to make the sound of their installation more euphoric) in fact this pre-amp is acting as an equalizer, this means that there is something abnormal in the rest of the chain from the DAC. It’s that simple

      1. Each potentiometer that is moved on the side of the recording chain, technically adds distortion to the signal, so that the efforts made not to increase it on the side of the playback, will be welcome to not pervert it further.

        It’s as simple as that.

        1. In the analog domain the interface between two components is itself a “thing”. If the output stage of the DAC and the input stage of the power amplifier are not well matched the result may be less than ideal. However, by inserting a preamplifier whose input stage matches well to the output stage of the upstream DAC, and whose output stage matches well to the input stage of the downstream power amplifier, the net result can sometimes be an actual improvement in sound.

          Agreed, the additional preamplifier in itself cannot “improve” the signal as it passes through, yet the net result of inserting it into the signal path can yield an unexpected overall improvement.

          And, for clarity, I am not suggesting for a moment that this is fundamental to all combinations of DACs, preamps, and power amps. But it can (and does) occur.

          1. The match between components has been a problem solved since the days when the standard for coupling within professional audio was 600 Ohm, strictly.

            Today this has changed, you can see that the vast majority of professional amplifiers have input impedances of 10 K Omn or more.

            I would not buy any DAC that cannot handle a device with input impedance of 10 K Omn, or more. everyone I’ve had (and there are many) have been able to do it.

            Now, if your position is to appreciate the subjective differences between the DAC-Amp coupling, that is a different story.

            But you have not been explicit in saying what these mismatches consist of, so I deduce that these are subjective.

            It has been interesting talking with you.

  12. To popularize use of DAC/pre-amp volume control, I would make your products more broadly compatible with universal remote controls. It’s a pain to always keep my separate PS Audio remote on-hand just for volume control; it would be great if my TV remote could control my DS DAC… so, I think app-based volume control is primarily a convenience issue.

  13. While many server software products give you the ability to control the volume it is not the only place you can do it.

    Here’s what the developers of the BubbleUPnP control point had to say (which I use with Jplay Femto):
    “Every UPnP/DLNA renderer has a command to set its volume (generally between 0 and 100).
    BubbleUPnP uses that command and it is up to the renderer change the volume how it wants. In most case it changes the volume of the hardware and do not modify the audio data to do so.”

      1. I just discovered Femto does not pass volume commands to the dac. There is BitPerfect Volume control feature in JPLAY Settings but that’s not useful for adhoc volume changes.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram