[ZappaMan]

I think wazzaps commentary is useful - this stuff is expensive - all too easy to get carried away - ‘needing’ this or that....

It really is expensive - comparatively or otherwise - it’s really expensive.

 

[musickid]

In Mandarin and Cantonese a person expresses oneself by three determining factors. The first is the content of what they are saying, the second is the posture of the individual and finally the tone of their voice. When conveying the same information this can take a dramatic change for the worse or better by a simple change in tone or posture which is ultimately the determining factor. To come onto an mscaler thread full of proud mscaler owners and then describe it as an expensive box can only be construed in two ways. Either an attempt to grab attention or secondly provoke a response. A little more tact is called for from now on i think??

 

[Whazzzup]

Ok a high strung thread, could be more sensitive, while I appreciate the etiquette of mandarin and Cantonese, perhaps a third option is available. Sorry for the use of box, but it was upkeeping with the three mentions of box used leading up to the final.

 

[musickid]

Many thanks Whazzzup

 

[SunWarrior]

I think wazzaps commentary is useful - this stuff is expensive - all too easy to get carried away - ‘needing’ this or that....

It really is expensive - comparatively or otherwise - it’s really expensive.

Very true, worth regular reminders.

And because expensive, that's why I bought the Hugo M Scaler.
Because it's not one of audio's endless tweaks.
It's a true -- and for many of us immediate -- major improvement (if you have one of the four fully compatible Chord DACs).
FWIW, I got away with spending $5K on the HMS, which wasn't the best budgetary choice for us, partially because my music-lover-but-not-equipment-lover Significant Other heard better music right away.

Dave, who might categorize ferrules as tweaks and that's cool but I'm just tired of audio tweaks and want to go big or go home

 

[Rob Watts]

For Rob and everyone else who can help, @Rob Watts

On a more interesting point i understand the full benefits of a 1M tap interpolation filter and the principle of reconstructing an analogue signal to greater than 16 bit accuracy in real time. I am struggling however with understanding the benefits of upscaling to 705/768 kHz. Is this fully tied in with the physics concerning interpolation filters as i've read the human ear cannot hear beyond 192 kHz (if that) as it lies beyond the ultrasonic capability of human hearing? It would be great to get a real explanation and deeper understanding of this point as that can only be good for curious mscaler owners such as myself! many thanks again MK.

As you know, ideally to reconstruct a bandwidth limited signal you need a sinc function; but this is infinitely oversampled and an with infinite number of samples. With a finite filter we have a balancing act to do; either running at a high oversampling rate, and working with less data, or running at a lower oversample rate but the filter processing more data. Seeing more data affects the filter turnover frequency; and the closer we get to FS/2 the better the sound quality. But also, the output resolution is key too; doubling the oversampling rate with the same number of samples improves SQ too. There was a massive increase in going from 8FS filters to 16FS filters, with the number of samples the same. So we do have a resolution issue.

If you had asked me what oversampling was OK 20 years ago I would have answered that the requirement was 4uS - that is 250 kHz, as this is the smallest figure I have seen for the interaural delay - that's how fine the brain measures time difference between the ears. But hearing resolution and perception are not the same things; also this only relates to sound-stage left to right processing, it tells us nothing about perception of pitch, timbre or instrument separation and focus - things we have little understanding about - that is how the brain does this processing. I now know for certain that 4uS resolution is not good enough. Also, when I replaced the 16FS to 256FS filter I got a big difference in SQ; so the resolution required is very much finer than 4uS.

So to answer your question, it's about how accurate the timing of transients actually are, not being able to hear 768k.

 

[mlxx]

For Rob and everyone else who can help, @Rob Watts

On a more interesting point i understand the full benefits of a 1M tap interpolation filter and the principle of reconstructing an analogue signal to greater than 16 bit accuracy in real time. I am struggling however with understanding the benefits of upscaling to 705/768 kHz. Is this fully tied in with the physics concerning interpolation filters as i've read the human ear cannot hear beyond 192 kHz (if that) as it lies beyond the ultrasonic capability of human hearing? It would be great to get a real explanation and deeper understanding of this point as that can only be good for curious mscaler owners such as myself! many thanks again MK.

What comes out of the mscaler is actually less accurate (made worse) than what went in. The mscaler is not creating new frequencies that you can't hear, it is filling in data in between the original samples but it is doing so with imperfect filters. The reason the mscaler exists is because we have imperfect dacs that need higher frequencies to do their job better. To get to these higher frequencies you need to upsample and therefore need to calculate the data at these new samples. It is still a representation of the same 44.1k waveform but now with 705k samples (the waveform looks the same but with more samples). If you did have a perfect dac, it would be better to feed it the original 44.1 than the output of the mscaler.
(Although I say the mscaler makes it worse, do not take offense to this it will always be the case, the mscaler does a better job at this than anything before.)
Rob beat me.

 

[JaZZ]

On a more interesting point i understand the full benefits of a 1M tap interpolation filter and the principle of reconstructing an analogue signal to greater than 16 bit accuracy in real time. I am struggling however with understanding the benefits of upscaling to 705/768 kHz. Is this fully tied in with the physics concerning interpolation filters as i've read the human ear cannot hear beyond 192 kHz (if that) as it lies beyond the ultrasonic capability of human hearing? It would be great to get a real explanation and deeper understanding of this point as that can only be good for curious mscaler owners such as myself! many thanks again MK.

Oversampling is the interpolation filter (also called «digital filter»). The finer the steps between the original samples, the more effective the filter. So high (over)sampling rates are called for.

Actually that's a simplified display and explanation. In fact Rob's WTA filter doesn't even preserve the original samples.

(I see Rob has already answered.)

 

[AndrewOld]

I think wazzaps commentary is useful - this stuff is expensive - all too easy to get carried away - ‘needing’ this or that....

It really is expensive - comparatively or otherwise - it’s really expensive.

Absolutely, we should never forget how privileged we are to be able to even consider such things. It is an expensive box, by any reasonable standards, and as a passionate owner I am not remotely offended by anyone describing it as such.

 

[racebit]

What comes out of the mscaler is actually less accurate (made worse) than what went in. The mscaler is not creating new frequencies that you can't hear, it is filling in data in between the original samples but it is doing so with imperfect filters. The reason the mscaler exists is because we have imperfect dacs that need higher frequencies to do their job better. To get to these higher frequencies you need to upsample and therefore need to calculate the data at these new samples. It is still a representation of the same 44.1k waveform but now with 705k samples (the waveform looks the same but with more samples). If you did have a perfect dac, it would be better to feed it the original 44.1 than the output of the mscaler.
(Although I say the mscaler makes it worse, do not take offense to this it will always be the case, the mscaler does a better job at this than anything before.)
Rob beat me.

You are absolutely right that if we have a perfect DAC there is no need for MScaler. But the reason is, as I said a little before, the MScaler is a component of the DAC, the component that upsamples to 705K, the WTA1. So using the MScaler we are just replacing the component and making the DAC more perfect, or less imperfect as you prefer. A perfect DAC will have an MScaler inside (or a better MScaler than the current one) instead of the WTA1 filter, the MScaler is a WTA1 filter really.
MScaler seems near perfect right now, but maybe in 20 years it will look far from perfect compared to Chord products of that time.

 

[mlxx]

You are absolutely right that if we have a perfect DAC there is no need for MScaler. But the reason is, as I said a little before, the MScaler is a component of the DAC, the component that upsamples to 705K, the WTA1. So using the MScaler we are just replacing the component and making the DAC more perfect, or less imperfect as you prefer. A perfect DAC will have an MScaler inside (or a better MScaler than the current one) instead of the WTA1 filter, the MScaler is a WTA1 filter really.
MScaler seems near perfect right now, but maybe in 20 years it will look far from perfect compared to Chord products of that time.

It is not part of the dac, technically the dac is the pulse array in this case... but we are basically saying the same thing.

Describing the mscaler as an expensive box when servers can cost double as much is really something. Describing the WTA algorithm as something that can be formulated in one lunch time break when it is actually the culmination of 30 years of dedicated work and research and the pinnacle of one's career is really something. Funnily enough these statements always seem to come from those who don't own the hardware being commented on. I noticed that trend by cross checking profiles everytime such comments are made. Please do go ahead and implement a new WTA algorithm if you can do it...then do it. That's really something.

Please do explain how WTA is relevant for the mscaler because as a non owner I don't get it....
https://images2.imgbox.com/12/84/sjhgwZK0_o.jpg

 

[musickid]

I have no technical training beyond the age of 16-18 at school. My background is in military history as i stated before. That's why when i reach a mental block in my quest to understand the basics here on headfi i go straight to Rob for clarification. This is my hobby since 17 please let me do what i can. At 17 i had a NAD amp, celestion 3 speakers with the titanium domes and a German dual turntable. For 4 years i had a yulong u100 with akg k550. After tube experimenting i discovered hugo 1. Now i probably have the world's 3rd most advanced headphone rig with an emphasis on the word probably.

 

[musickid]

Many thanks for the response Rob. On occasion i do hit a mental block trying to understand some of the scientific principles here on headfi.

 

[Amberlamps]

For Rob and everyone else who can help, @Rob Watts

On a more interesting point i understand the full benefits of a 1M tap interpolation filter and the principle of reconstructing an analogue signal to greater than 16 bit accuracy in real time. I am struggling however with understanding the benefits of upscaling to 705/768 kHz. Is this fully tied in with the physics concerning interpolation filters as i've read the human ear cannot hear beyond 192 kHz (if that) as it lies beyond the ultrasonic capability of human hearing? It would be great to get a real explanation and deeper understanding of this point as that can only be good for curious mscaler owners such as myself! many thanks again MK.

192khz human hearing ?

I thought the best a human with perfect hearing can hear is 20khz.

And that gets degraded with age to a point where most folk who are middle aged hear 15/14khz or lower.

You ever walked past your local newspaper shop and noticed something that resembles a burglar alarm but isn’t ? Those are high pitched speakers that only kid’s can hear and which is so annoying to them, they choose to go congregate elsewhere.

They use 17.5khz, us adults can’t hear it. They have even made ones for every age group that emits a 8khz tone.

192khz is way above human hearing.

https://www.compoundsecurity.co.uk/security-equipment/mosquito-mk4-anti-loitering-device

https://en.m.wikipedia.org/wiki/The_Mosquito

 

[rgs9200m]

I used to have the Celestion SL6 speakers back in the 90s. Sweet sound.