Page 1 of 2 12 LastLast
Results 1 to 25 of 30

Thread: Volume Control

  1. #1
    Sophomore Member
    Join Date
    Apr 2010
    Location
    UK
    Posts
    232

    Volume Control

    Thinking; why does this still have to be written into a DAC.

    I don't need any other function from a DAC other than volume control, USB and optical inputs, and balanced output

    Have been thinking about volume controls in software, Is there anything that allows proper dithered adjustment/level control on a PC/MAC and iPod Touch?

    Will Amarra do the job without loss in data?
    Meridian MC200 - DSP6000MKII

  2. #2

    "Will Amarra do the job

    "Will Amarra do the job without loss in data?"

    Not if you dont have any other way of reducing gain. This will reduce resolution.

    The best scenerio to achieve remote control is to use a volume-controlled DAC and set the level for a low-volume track. Then, when a louder track plays, reduce it with Amarra volume. This only requires 3-9 dB.

    Check out the volume control technology on the DAC first. They are all different, with different quality levels.

    Steve N.
    Empirical Audio

  3. #3

    What Steve N. said...

    What Steve N. said... basically, you don't want to use too much digital volume control because that will chop off bits, but a little is ok. Hence combining analogue control in the Dac or pre amp for the ballpark volume and then fine tuning with Amarra or equivalent.

  4. #4
    Sophomore Member
    Join Date
    Apr 2010
    Location
    UK
    Posts
    232

    Cheers guys, so a dac with

    Cheers guys, so a dac with volume is the minimum.

    Amarra claim they have fully dithered volume control but I'm not a user and is why I've asked the question

    regards,

    Raj
    Meridian MC200 - DSP6000MKII

  5. #5
    Sophomore Member
    Join Date
    Feb 2009
    Location
    UK
    Posts
    289

    ... another approach

    The overall gain of many power amps can be easily reduced by appropriate modification of resistors at the input stage. Some (e.g. Krell) have external switches for reducing gain, other manufacturers will do the mod for a small fee.
    Depends on how many bits are used in the digital volume control but if have 24 bits (now common) then you only need to get within 10dB or so of max digital volume.

    David
    Main System
    ALAC iTunes library on Synology DS412+ running MinimServer > Airport Extreme bridge > Sonore Rendu + iPad mini remote > Bel Canto DAC 3 with VBS1 > Krell S-275 > Harbeth SHL5s
    Bedroom System
    ALAC iTunes library on Synology DS412+ > Mac mini (2013, 16GB RAM) OSX Mavericks, Audirvana+ > USB > MF V-link192 > Metrum Octave DAC > LFD integrated > Harbeth P3ESRs and MF X-CANv8p > Sennheiser HD800

  6. #6
    Senior Member
    Join Date
    Apr 2010
    Location
    Finland
    Posts
    4,225
    Blog Entries
    5

    Volume controls

    Any type of volume control reduces SNR. Implementation details define which one gives best quality. There's no clear cut answer on what is the best way.

    If you are playing 44.1/16 data through 24-bit DAC, you can reduce volume by 48 dB without losing any bits. OTOH, when properly used, 24-bits is not the limiting factor, but the analog electronics. And digital signal processing allows various tricks regarding dynamic range.

    Digital volume control with 16-bit output is not very good idea.

    Signalyst - http://www.signalyst.com
    Developer of HQPlayer

  7. #7

    Digital volume control

    I wrote a white paper on the "digital volume control" topic.

    http://www.weiss-highend.ch/computerplayback/Digital_Level_Control.pdf

    It includes some sound examples.

    Daniel

    www.weiss.ch

  8. #8
    Sophomore Member
    Join Date
    Apr 2010
    Location
    UK
    Posts
    232

    Nice one.

    From the man himself! Hello Daniel, I'm just in the process of deming and will hopefully buy your DAC202. Could you tell me something, How does it compare to the DAC1 mkII?
    Meridian MC200 - DSP6000MKII

  9. #9

    DAC202 vs DAC1-MK2

    The DAC1-MK2 is a completely different unit from the DAC202, both analog and digital parts are entirely different. The DAC1-MK2 has a discrete class A output stage, DSP based upsampling (our own algorithm) and DSP based de-jittering. A firewire input can be installed in the DAC1-MK2, i.e. basically the INT202 PCB installed in the DAC1-MK2.
    Which one you would prefer sonically you would have to judge for yourself.

    Daniel

    www.weiss.ch

  10. #10
    Sophomore Member
    Join Date
    Apr 2010
    Location
    UK
    Posts
    232

    Thankyou

    Two more questions , is the pro DAC line-up due an overhaul soon and Which is more Technically advanced in you opinion.

    Regards, Raj


    Meridian MC200 - DSP6000MKII

  11. #11

    overhaul

    we are working on a new analog part for the DAC1-MK2 and the Medea DACs. This brings them to a level (technically) above the DAC202.

    Daniel

    www.weiss.ch

  12. #12
    Sophomore Member
    Join Date
    Apr 2010
    Location
    UK
    Posts
    232

    When will the (i assume) MK3

    When will the (i assume) MK3 be out, will it have remote controlled volume. As i understand it, the DAC1 has a volume control already but is manual?
    Meridian MC200 - DSP6000MKII

  13. #13

    DAC1 / Medea update

    We will show the upgrade at CES in a Medea DAC.
    A remote controlled volume control is available with the Firewire input, which is also an option for both DAC1-MK2 and Medea. That is already available. The volume control is for the Firewire input only, though.

    Daniel

    www.weiss.ch

  14. #14
    Sophomore Member
    Join Date
    Apr 2010
    Location
    UK
    Posts
    232

    Excellent, thankyou.

    Excellent, thankyou.
    Meridian MC200 - DSP6000MKII

  15. #15

    volume control via FIrewire on DAC2

    Hi Daniel,

    I am a happy owner of a DAC2 and won't be able to upgrade for a while. Is it possible to control the volume via Firewire on the DAC2? I control my headless Mac Mini from my Macbook, and would love to eliminate the preamp and control the volume from afar. When I've got my pup on my lap, I don't want to get up, but controlling the DAC2 via a screen-sharing laptop would be great.

  16. #16

    volume control via firewire

    That won't be possible until we write our own firewire driver, but that is not very high up in the list, as the standard driver for the Dice chip works fairly well.
    I assume you are using iTunes and maybe Amarra or Pure Music for playback? Then you could control the volume via the player software, or?

    Daniel
    www.weiss.ch

  17. #17
    Sophomore Member
    Join Date
    Apr 2010
    Location
    UK
    Posts
    232

    Are the volume controls in

    Are the volume controls in the above programs lossy or will any of them cause compression? I'd also like to control the volume via my ipod touch, itunes remote ap does this but I'm not sure if its a well designed feature.


    Meridian MC200 - DSP6000MKII

  18. #18

    volume control

    I am not quite up to date re the quality of the volume controls in the various programs. With Amarra I think you are on the safe side.

    Anybody else knows more?

    Daniel

    www.weiss.ch

  19. #19

    Pure Music & AyreWave volume controls

    AyreWave: If the volume control of the DAC-202 appears in Audio MIDI Setup, then I believe AyreWave should be able to control its volume.

    Pure Music has software volume control with dither parameters configurable by the user. Here's the explanation from the Pure Music user guide:

    "At Monitor volume settings below 0 dB, dithering may optionally be applied:

    "A 64 bit pseudorandom number generator (PRNG) with an extremely long sequence repeat length is used to generate the dither in real-time, instead of using a "canned," repetitive dither sequence. The Uniform dither option is the suggested choice, and the optimum setting of the slider is 0.75 LSB. The LSB word length should be set to 24 Bit, unless using a DAC with only 16 bit resolution (unusual) or capable of 32 bit integer word length (very unusual). (Many audio interfaces simply report the canonical Core- Audio 32-bit float format in Audio MIDI Setup, instead of the true internal integer format of the DAC.

    "The Amplitude word length setting (16, 24, 32) only affects the dither amplitude. All volume adjustments in Pure Music are made at 64 bit resolution, regardless of the Amplitude setting.

    "Other options are provided for those who prefer noise shaped dither. Moving from the Uniform to the 12th Order Noise Shaping options, the dither energy becomes shifted to increasingly higher frequencies, and a narrower frequency range. More CPU cycles will be consumed when using noise shaped dither.

    "Stereo (uncorrelated) dithering uses dual PRNGs to generate the dither independently for each channel. Deselecting this option causes the same dither to be applied to both channels. The Stereo setting is recommended."
    Mac Mini (2010 Core2Duo, OS X Mtn Lion) > Audirvana Plus, BitPerfect or Pure Music > Metric Halo LIO-8 > Parasound JC-1 > Thiel 3.7

  20. #20
    Senior Member
    Join Date
    Apr 2010
    Location
    Finland
    Posts
    4,225
    Blog Entries
    5

    Software volumen controls

    In HQPlayer volume is applied together with all other processing in 64-bit floating point format. Audio device is always used at the maximum bitdepth and dither/noise-shaping is applied at final stage using algorithm selected by the user (6 possibilities currently).

    There are maximum three level controls (all optional), main volume, speaker configuration (multichannel case) and gain compensation for convolution engine. Control range for the main volume is adjustable.

    Signalyst - http://www.signalyst.com
    Developer of HQPlayer

  21. #21

    "Any type of volume control

    "Any type of volume control reduces SNR. Implementation details define which one gives best quality."

    Not true. There are volume control technologies that have the opposite characteristic. Lower volume delivers lower S/N.

    Steve N.
    Empirical Audio

  22. #22
    Senior Member
    Join Date
    Apr 2010
    Location
    Finland
    Posts
    4,225
    Blog Entries
    5

    SNR

    Not true. There are volume control technologies that have the opposite characteristic. Lower volume delivers lower S/N.

    Sure if it's poor to begin with, but I didn't count those. If the volume control is proper, noise floor is thermal noise of low-valued resistor. Thermal noise gets lower only by lowering resistance or temperature. SNR is ratio of the signal level compared to this thermal noise, as the signal level goes down, distance to the thermal noise decreases.

    With a good digital volume control, amount of noise is equivalent to DAC's self noise, in best cases equivalent to around 100 ohm resistor and usually from 500 ohm to 1 kohm.

    Of course, with really fancy volume control the temperature of the volume control circuit would go down as volume is decreased, reaching absolute zero at mute. In this case, SNR would stay the same or increase as volume is lowered.

    I'm yet to see a circuit with >120 dB SNR at full level and SNR to increase from this value as volume is lowered.
    Signalyst - http://www.signalyst.com
    Developer of HQPlayer

  23. #23
    Senior Member PeterSt's Avatar
    Join Date
    Mar 2008
    Location
    Netherlands
    Posts
    3,207

    Lossless

    To my own experience and infinite trying, there is no analogue volume control possible which doesn't -one way or the other- degrade. And I may have tried more possibilities than anyone may have thought of (including the most wild variations and combination in the I/V area).

    The most important what will happen is non-linearity at attenuating. At best a VC may incur for only 3dB worse figures (let's say THD+N but it can be N only as well), but next at attenuating the non-linearities start (not in all cases, but when *this* is not happening, THD+N is e.g. 12dB worse to begin with).

    Notice that this is all to be compared with no analogue VC in the chain, and not with a random other. So, no pre-amp there as well.

    What *always* happens is a more dead sound. Not necessarily to be perceived as less dynamic (although in most situations that will be the case), but just not "alive" anymore.

    For the situations where it was applicable, I (as well) used expensive resistors (think 10$++) of various kinds (paper foil etc.). It just can't work.

    Let me add that my attempts all were meant to be "in-DAC", which in my case is about "more transient" than common. This will matter. I mean, if anything has the chance to filter, it will filter high transients as the first thing. That's also why it is always audible, and why "alive" turns into dead (those transients being at micro-micro level, not necessarily leading into the perceivement of "less dynamical").

    If Steve thinks there is a way to do it, I will be glad to hear it. But expect an immediate answer how *that* solution exhibits. But one never knows ...

    Are the volume controls in the above programs lossy or will any of them cause compression?

    Since the term "lossless" for a digital volume control will have come from me anyway (not that I know otherwise), the digital VC from XXHighEnd is lossless. That is, for 16/44.1 material played onto a 24 bit DAC and no more than 48dBFS attenuation. This means that after the attenuation has been applied, out of the attenuated file the original can be recreated; I call that "a lossless digital volume".

    Normally this isn't all that much important. I mean, when people like to use a pre-amp anyway (reason unimportant for now) then you just don't use the digital volume and all is okay. However, once you know what *any* analogue volume does to the sound as I explained above, you will know better, and then the importance of the (quality of the) digital volume comes into play.

    As I think implied in earlier posts, the digital volume can be in the DAC chips used (depending on the chip), which theoretically allows for remotely controlling that (say, on the screen). However, these means are again not lossless.

    Also, all is relative;
    If you use a D/A converter of the heavy oversampling type, all is a kind of moot, because that mangles with the bits far more than any lousy digital volume is capable of. Thinks me.
    It still will matter somewhat for theories, but I guess the way Miska explained about his application of the digital volume will work out as good as mine.
    Similarly, an analogue volume will be less harming in the heavy oversampling situation (which doesn't contain those transients anyway).

    Back to the analogue volume briefly :
    What I learned to use as a measure, is looking at the noise level only. Regarding this : notice that noise is never noise as such, but are just waves comprising of complex forms, all with an origine (which could be a nice sine for that one origine). Because they are complex, they incur for harmonics all over the place. Read : this is a test signal comprising of multiple frequencies (think about the origine) creating all those harmonics. Now, I have never seen an injected analogue volume control that would let the noise level straight, up to the analyser's capabilities. To me this will show that the frequency response will not be linear as well, which will *not* show at performing a sweep (which is ultimately about one frequency per "step" only).
    In the end all is logic, because whatever element is used for attenuating analoguely, it will filter, and it won't filter all the frequencies evenly.

    Just look at the noise line, and compare it without VC.

    Peter
    XXHighEnd (developer)
    Phasure NOS1 24/768 Async USB DAC (manufacturer)

  24. #24
    Senior Member PeterSt's Avatar
    Join Date
    Mar 2008
    Location
    Netherlands
    Posts
    3,207

    Nits !

    With a good digital volume control, amount of noise is equivalent to DAC's self noise, in best cases equivalent to around 100 ohm resistor and usually from 500 ohm to 1 kohm.

    I think this is far more complex than one sentence can ever explain. It may be nit picking from my side, but interpreting the effect of a digital volume control already is difficult.

    For example, if you'd attenuate 3dBFS, THD figures will show worse (I didn't say THD+N). This is just inherent to measurement (applications), because theoretically (calculating from the disappearing resolution) this is so.
    Still it is not true when my original signal was 16 bits, and is to be interpreted as "24 bits capable" which the analyser won't know. I mean, that signal will be 24 bits from the start, and next I will be loosing one bit ... thinks the analyser. -> THD is worse. Hard to explain without using a couple of pages more *and* I even may be wrong on it. :-)

    Next example is the ultimate THD+N of the chip, which just might be at -0dBFS. Or better, at -20dBFS. When the latter, chances are high that at showing all at -0dBFS the noise level rises with it. What next will matter is that THD will rise (worsten) with it too, and lastly it would be the question whether THD was under the noise level to begin with.
    Here too, much more to explain, but you will get the idea.

    All 'n all I only want to say that interpreting these figures can be a tough job, and the higher the self-noise is, the more difficult it can be.
    But it is also the other way around : the lower the self-noise is, the better shows THD, and theoretically, the better it can be heard. Thus, THD under the noise level won't be audible (false harmonics) because the noise comes first (nicely white), while lowering the dBFS may drop the noise as well (see above) but may not drop inherent harmonics from outside the DAC (cables etc.). This means that a fair amount of digital attenuation will make the signal to THD worse, while in the mean time THD itself will worsten also (loosing too many bits).

    As said, nit picking, but can be important depending on the situation.

    Peter
    XXHighEnd (developer)
    Phasure NOS1 24/768 Async USB DAC (manufacturer)

  25. #25
    Senior Member
    Join Date
    Apr 2010
    Location
    Finland
    Posts
    4,225
    Blog Entries
    5

    THD and volume

    For example, if you'd attenuate 3dBFS, THD figures will show worse (I didn't say THD+N). This is just inherent to measurement (applications), because theoretically (calculating from the disappearing resolution) this is so.

    THD is interesting and complex topic in itself... But no, there's no THD involved due to dithering and noise shaping.

    For the sake of reducing complexity, we should assume proper digital volume, thus no THD above quantization noise.

    Some DACs have lowest THD at 0 dBFS some at lower levels like -20 dBFS. In these cases, the optimal THD figure may be when the digital volume is not at maximum. Some, especially traditional multibit DACs, easily suffer from poor low level linearity, thus giving rise on THD percentage as signal level gets lower (this is one of the particular reasons for using delta-sigma DACs).

    Another case depends on type/cause of THD. If it's normal linear ratio, thus some percentage of the signal level, it's level also reduced together with the volume thus the percentage remaining the same. In some cases the percentage at lower level could be even lower (class-A). If cross-over distortion is causing the THD, then it's level will likely remain the same regardless of signal volume, thus increased percentage.

    Now if analog volume control has low enough self-noise it may or may not be beneficial in terms of overall signal quality, depending on the case. Rough rule could be that with DACs having good low level linearity, digital volume may be beneficial to keep noise down. For DACs with less good low level linearity, analog volume may be beneficial to keep THD down.

    My reasoning is that for high quality sources with volume control, THD is dominated by the power amp, which is in most cases always running at full gain anyway. Thus, any type of volume control doesn't affect THD of the power amp. Same goes largely for noise too. Both processes are anyway additive, so the amount can only increase.
    Signalyst - http://www.signalyst.com
    Developer of HQPlayer

Page 1 of 2 12 LastLast