Do tube amps smooth out digital signals?

  • [Blocked Image: https://scontent.fitm1-1.fna.fbcdn.net/v/t1.15752-9/42497072_474940556338298_7632889825456029696_n.png?_nc_cat=109&oh=fd8017164e3f892403748b027f7b4160&oe=5C632DEE]
    I know how digital signals are choppy compared to analog, so I was wondering if/how running a Kemper through a tube amp smooths out the digital signal. Running my Kemper through my Fryette Power Station seems to do just that, but I'm wondering if what I'm hearing is a smoothing out of the signal or just an amplification of it with the added gain only seeming to smooth out the signal. If this has been discussed before I can't find it, so please do enlighten me if there has been a previous discussion of this or the technical side of how this does or does not occur.


    TIA

  • First of all, these illustrations do have a major flaw ... they just don't show you the actual number of samples (steps) per cycle.


    If your analog curve is meant to be a 1000Hz sine wave, then you'll have 88.2 samples for the given 2 cycles at 44.1kHz sample rate. The digital illustration only consists of 20 samples. So just by this, the actual precision of the digital representation of an analog sine wave is 4.4 times as high as the graphic shows.
    On top of that, DA converters typically have a smoothing filter (anti-aliasing) to reduce the stairsteps and even out the resulting analog curve quite a bit.


    Also keep in mind that a technical drawing is an idealized and highly simplified visualization of something that is way more complex in reality. You'll likely never send a perfect sine wave to an amp ... and even less to the actual speaker. Analog audio content like a guitar signal already contains lots of harmonic content and unevenly shaped waveforms. A purely analog signal chain already modifies the audio content more than the AD and DA stages of a decent digital device would introdude to your audio.


    Bottomline:
    I wouldn't worry about "digital arifacts" at all. Of course a tube power amp will sound different to a Class D amp. But that's not due to the audio signal having passed a digital stage (e.g. the Profiler) or not.

  • OK, thanks for the response. So is the general consensus that, due to anti-aliasing, digital signals are already as smooth as analog ones and so the question is rendered invalid?


    The illustration is meant to be a generalized example of a stairstepped digital wave vs a smooth analog one, not of any specific sine wave, which is how I'm led to believe is the state in which they generally exist. Is this an incorrect depiction?


    One of the main complaints AFAIK with all things digital concerns this stairsteppedness and since the Kemper produces strictly digital output, if that output (regardless of the sample rate or sample size) is smoothed by a tube amp or if it is merely amplified. That is the question I'm looking to answer.


    I have many tube amps as well as digital ones and of course they all sound different, not only based on class but also based on circuitry, design, and the speakers they employ. To my ear, there is a world of difference between digital and analog amps and I will choose analog every time. The Kemper into the Power Station and/or into a Tube Amp seems to maintain the warmth and feel that even the best tube simulation plugins can't match. Am I missing something here?

  • "Stairsteppedness" doesn't exist. Really, it DOES NOT EXIST! It's one of the biggest, yet most hard to kill myths about digital audio. The maths is difficult to understand, granted, but if you're interested in an easier to understand explanation, watch this (from 3.40 is specifically about stair steps) :


  • Yeah, this video from Monty practically went viral across the audio forums back in the day.


    OK, thanks for the response. So is the general consensus that, due to anti-aliasing, digital signals are already as smooth as analog ones and so the question is rendered invalid?

    I believe so, SJ.


    With regards to the Kemper -> tube-based power amp or preamp, whilst it has been asked about many times here, especially in the case of amplification for an FRFR setup, the consensus as always unanimously been that the Kemper is a very-accurate tube "simulator" and that it wouldn't make sense to whack an additional layer of tube/s on top of it. After all, this would compromise the integrity of the Profiles.

  • The tube amp sounds warmer not because it is "smoothing" the digital signal back to a more accurate reflection but because it is introducing new distortion of the signal. Tubes sound "warm" and "full" because they distort and compress the signal. Therefore, I believe you are correct that the question is invalid.

  • Concerning digital harshness... I have been a high end vinyl playing audiophile for many years, and build my own single ended tube power amps and tube preamps\phono stages. I am sensitive to and despise digital harshness, but fortunately with higher bit rates and higher quality DACs and ADCs, digital harshness is mostly gone from high end audio unless an equipment builder ignores the issue or cheaps out.


    The main reason I love the Kemper so much is it has zero digital harshness. Zero. IMHO the use of a Kemper goes beyond profiling and modeling and effects, in the end it just sounds and feels great. Its lack of digital artifacts is by far the best I have ever heard in any guitar digital equipment. I would ditch the Kemper immediately if it had a trace of harshness, and in several years of live and studio use I've never heard a trace from it.


    There are modelers available that if you play them through a neutral Class D power amp, it will rip your head off with harsh tone and you need a tube power amp to tame the harshness and feel issues. But the Kemper is specifically designed to run through as neutral a power amp as possible as it has the harshness and feel issues already eliminated.

  • "Stairsteppedness" doesn't exist. Really, it DOES NOT EXIST! It's one of the biggest, yet most hard to kill myths about digital audio. The maths is difficult to understand, granted, but if you're interested in an easier to understand explanation, watch this (from 3.40 is specifically about stair steps) :



    Nice, great vid! 8)

  • We once send my Kemper signal through some Telefunken V76 tube preamp. First I didn´t like the result because I wasn´t used to the sound. But then I liked it. And after all to me this tube preamp complemented the Kemper signal.

  • OK, thanks for the response. So is the general consensus that, due to anti-aliasing, digital signals are already as smooth as analog ones and so the question is rendered invalid?

    The general consensus is that when you are removing high frequency content from the digital curve, it becomes as "smooth" as the other curve.
    The thing is, if you can't hear above 20khz (which you can't do if you are a human and not a bat), you will never be able to hear those "stairsteps".

  • Tube poweramp sounds like it does not because its tube but because of the way the output transformer responds with the speaker. SS amps have a super high dampening factor which keeps a tight control over the speaker. This is why people say they sound sterile because you get very little of the speakers character. All the profiling or modelling tricks in the world cant change that fact.
    A tube amp lets the speaker 'loose' at certain frequencies depending on where the resonance and presence controls are (if fitted). The lower they are set the flatter the response of the amp. You can absoluey build a tube amp with tight speaker dampening if you wanted to. Hifi boys have been doing that for years.


    I love the sound of my kemper through my Synergy 5050 poweramp.
    Diffence in feel is night and day.

  • The general consensus is that when you are removing high frequency content from the digital curve, it becomes as "smooth" as the other curve.The thing is, if you can't hear above 20khz (which you can't do if you are a human and not a bat), you will never be able to hear those "stairsteps".

    Please, do yourself a favour and watch the video I linked to further up the thread. There are no stair steps. The filtering above 20 kHz stops the analogue to digital converter from "folding" the frequencies above Nyqvist back down in to the audible range creating artefacts, a process known as aliasing. You can think of it as unmusical harmonics, in that they don't harmonise naturally with the rest of the audible frequencies, often causing "digital harshness". This hasn't been a worry for manufacturers of AD/DA converters since the late 80s. When people say that digital audio doesn't sound as good as analogue, they are mistaking the shortcomings of digital audio for the shortcomings of analogue; more often that not, it is the inherent extra noise and distortion (i.e. less than perfect audio reproduction) of analogue that gives the audio "warmth". Unless of course we're talking about lossy digital audio formats (mp3, mac etc.), but that's a whole other chat ;-)

  • Please, do yourself a favour and watch the video I linked to further up the thread. There are no stair steps.

    Thank you, i have not watched the video you have linked since this is something i'm already familiar with. What part of my previous answer did you not agree with ?

  • Thank you, i have not watched the video you have linked since this is something i'm already familiar with. What part of my previous answer did you not agree with ?

    “Stair steps”. One won’t be able to hear them, not because of the band pass filter, but because they don’t exist in the first place. The filter isn’t there to smooth out the digital curve, either, as it doesn’t need smoothing.
    Maybe you were just trying to put the science into layman’s terms, but I think notions like these should just be debunked, otherwise people will still be getting confused years from now.

  • “Stair steps”. One won’t be able to hear them, not because of the band pass filter, but because they don’t exist in the first place. The filter isn’t there to smooth out the digital curve, either, as it doesn’t need smoothing.

    How is the signal looking before the reconstruction filter ?

  • How is the signal looking before the reconstruction filter ?

    There is no signal before the reconstruction filter, just binary sample point values.
    Those sample points are just waymarkers for audio waves to travel through. The value at a sample point when considered with the preceding and subsequent points determines the shape, or frequency, of the wave. Mathematics laws mean there can only be one solution, one resulting wave, that runs through those points. Hence the reconstruction of the audio wave perfectly, within the bounds of the Nyqvist law and the sample rate. Audio waves don’t have straight edges and sharp corners, not even ‘true’ square waves are actually square, it’s physically impossible.