[Reactcore]

One day I hope I will get it so that it is accurate enough, in that doubling will result in no or insignificant improvement in sound quality - but we are not there yet.

This means we will be looking at a Chord dac with triple or more coax data inputs to incorporate the bandwidth for multible Mtaps

I'm also wondering, since the Mscaler gives a delay big enough to give synchronisation problems for watching movies.. how many 'future' samples does it count ahead for say redbook? I can't project a trajectory covering too many original samples before the level changes from rising to falling slope or visa versa.

 

[ecwl]

I'm also wondering, since the Mscaler gives a delay big enough to give synchronisation problems for watching movies.. how many 'future' samples does it count ahead for say redbook?

Since it runs at 16fs with 1 million taps, roughly 1 million divided by 16 is 62500 so at 44.1kHz, that’s 62500/44100 = 1.417 seconds. But the filter is symmetric so divided by 2 then you’re getting 0.708 seconds of future samples (and 0.708 seconds of past samples) in the filter. Of course, movies are usually at 48kHz so the math would be slightly different.

 

[Reactcore]

Since it runs at 16fs with 1 million taps, roughly 1 million divided by 16 is 62500 so at 44.1kHz, that’s 62500/44100 = 1.417 seconds. But the filter is symmetric so divided by 2 then you’re getting 0.708 seconds of future samples (and 0.708 seconds of past samples) in the filter. Of course, movies are usually at 48kHz so the math would be slightly different.

So if i calculate further it will be 44100*0,708 = roughly 31222 original samples /2 is 15611 for future and 15611 backwards. But as i stated a signal level changes slope from rising to falling or via versa in a random number of samples.

This flaws my theory. Unless its always taking 4 samples and made of 'visualized' straight lines (as shown in my drawing) where the edges get rounded off by a formulated rendered shape.
Which with more taps get closer to the ideal intersection (transient) making the timing error smaller as the tap number increases.

 

[feddar]

Rob, sorry to ask, but I cannot find this anywhere: What is the current draw for the MScaller?

 

[Christer]

Since it runs at 16fs with 1 million taps, roughly 1 million divided by 16 is 62500 so at 44.1kHz, that’s 62500/44100 = 1.417 seconds. But the filter is symmetric so divided by 2 then you’re getting 0.708 seconds of future samples (and 0.708 seconds of past samples) in the filter. Of course, movies are usually at 48kHz so the math would be slightly different.

The latency/synch problem with Mscaler is a big problem for me with music videos. And especially so since I started playing the piano last autumn and suck at reading sheet music. I want to learn new music both from watching pianists fingering in YT videos and also a shortcut called Synthesia for lazy guys like me who find it too difficult to read complex piano works with lots of sharps and flats directly from the sheet music.
Is there a way to keep the FULL 1M taps SQ which I want to hear?
Even YT videos can sound quite good with Mscaling.

Or is my only choice to keep using the inbuilt really bad speakers on my mbp while learning new music, or resort to a non Chord dac without the latency problem to avoid the bad mbp speakers SQ?
With Mscaler the pianist's fingers are sometimes already at the other end of the keyboard before I actually hear the notes or chords I expected to hear.
It can be dizzying to see one thing happening and hear what you see so much later.
Cheers CC
PS I am aware of the video setting on Mscaler.
Cheers CC

 

[AndrewOld]

The latency/synch problem with Mscaler is a big problem for me with music videos. And especially so since I started playing the piano last autumn and suck at reading sheet music. I want to learn new music both from watching pianists fingering in YT videos and also a shortcut called Synthesia for lazy guys like me who find it too difficult to read complex piano works with lots of sharps and flats directly from the sheet music.
Is there a way to keep the FULL 1M taps SQ which I want to hear?
Even YT videos can sound quite good with Mscaling.

Or is my only choice to keep using the inbuilt really bad speakers on my mbp while learning new music, or resort to a non Chord dac without the latency problem to avoid the bad mbp speakers SQ?
With Mscaler the pianist's fingers are sometimes already at the other end of the keyboard before I actually hear the notes or chords I expected to hear.
It can be dizzying to see one thing happening and hear what you see so much later.
Cheers CC
PS I am aware of the video setting on Mscaler.
Cheers CC

My wife plays the piano quite well, I play it extremely badly. We use Pianoteq on a Mac, a Kawaii keyboard and a Focusrite Scarlett 2i2 USB DAC/midi interface with very low latency and some small powered speakers. High latency DACs are not really suitable to play through, unless you are used to playing large church organs. The 2i2 is not expensive, and quite adequate.

 

[ecwl]

The latency/synch problem with Mscaler is a big problem for me with music videos. And especially so since I started playing the piano last autumn and suck at reading sheet music. I want to learn new music both from watching pianists fingering in YT videos and also a shortcut called Synthesia for lazy guys like me who find it too difficult to read complex piano works with lots of sharps and flats directly from the sheet music.
Is there a way to keep the FULL 1M taps SQ which I want to hear?
Even YT videos can sound quite good with Mscaling.

Or is my only choice to keep using the inbuilt really bad speakers on my mbp while learning new music, or resort to a non Chord dac without the latency problem to avoid the bad mbp speakers SQ?
With Mscaler the pianist's fingers are sometimes already at the other end of the keyboard before I actually hear the notes or chords I expected to hear.
It can be dizzying to see one thing happening and hear what you see so much later.
Cheers CC
PS I am aware of the video setting on Mscaler.
Cheers CC

Hmmm...
First of all, why would you need a non-Chord DAC if you are aware that you can just put M-Scaler into video mode running at 600,000 taps?
Second, some of us watch YouTube on our big screen TV with the nVidia SHIELD and program the SHIELD to delay the audio by about 650ms (you'll have to time it with your TV). I have the nVidia stream the audio and video via HDMI to the TV and the TV to send the stereo Toslink signal to M-Scaler to get the full 1 million tap.
Third, I don't know if JRiver for Mac can stream a YouTube video and insert a A/V sync delay so that the video is delayed to sync with the audio for the 1 million taps. Perhaps somebody else using JRiver for Mac here can answer this question. Or you can just download it and try it for 30 days.
I have to admit, even though 1 million taps sound better, I don't always turn on my nVidia SHIELD for YouTube as I just live with the video mode.
But you bet I'll be watching the new episode of The Mandalorian with 1 million taps though... No spoilers please

 

[technobear]

VLC Media Player can also stream from Youtube and configure audio/video sync.

 

[AndrewOld]

Hmmm...
First of all, why would you need a non-Chord DAC if you are aware that you can just put M-Scaler into video mode running at 600,000 taps?

Because if, say, you have a keyboard connected to that DAC you want to hear an instantaneous response when your finger hits a key, as you would with a real instrument. Or if you are watching a video and trying to figure out what someone is playing by watching them and listening to them it is much harder if the sound is out of sync with the picture.

 

[Rob Watts]

I am afraid we can't escape latency issues and being able to reconstruct transient timing accurately enough - the filter must be able to see the future in order to reconstruct things correctly - it's just a mathematical certainty.

The Mandalorian sounds fantastic with 1M taps via nvidia Shield.

Rob, sorry to ask, but I cannot find this anywhere: What is the current draw for the MScaller?

It's 12W whilst operating, so just under 1A with the 15v supply.

 

[lojay]

Not really. 4 ohms is too much for Dave - unless your volume is at -6dB or below.

Hi Rob,

I came across your post and hope this is the right place to ask the question. I am considering driving the Cessaro Mini Wagner with the Chord Hugo TT2, but it sounds underpowered with the RCAs outputs. I then found out that the Wagners are 83 dB at 4 Ohm, which means that I will need to use the XLR outputs (with the Pin 2 and 3 configuration) with its full 18W per channel glory. Before I order a custom cable for this purpose, do you think this is feasible?

Also, I have been driving the Voxativ 9.87 speakers with the Chord Dave directly with wonderful results. I wonder if this is recommended? The drivers are field coils with the following characteristics.

Voxativ AC-Xp

Frequency response	20 - 20,000 Hz
Efficiency	max 108 dB in recommended housings
Capacity	50 W nom
Impedance [R]	10.7 Ohm
Complex impedance [Z]	16.0 Ohm
Resonance frequency	29.12 Hz
Mech. Q-factor [Qms]	3.1 / 9V
Electr. Q-factor [Qes]	0.4 / 9V
Q-factor [Qts]	0.354 / 9V
Equivalent volume [VAS]	148.0 Ltr.
Xmss	10 mm
Depth	166 mm
Install diameter	7.5" / 19.0 cm
Weight	12 kg

Many thanks and have a great day.
Jason

 

[Rob Watts]

Unfortunately, TT2 balanced output into 4 ohms may thermally trip, obviously depending upon how loud and the music you use. Once TT2 cools down, it will re-start automatically. Apart from thermals, 4 ohms is perfectly OK. Single ended into 4 won't trip.

Dave would have no problems with the Voxativ though.

 

[lojay]

Unfortunately, TT2 balanced output into 4 ohms may thermally trip, obviously depending upon how loud and the music you use. Once TT2 cools down, it will re-start automatically. Apart from thermals, 4 ohms is perfectly OK. Single ended into 4 won't trip.

Dave would have no problems with the Voxativ though.

Thanks Rob for the ultra fast response. Since the speakers will be used in a near field computer desktop setting (about a feet away from my ears), would it be a safe assumption that the TT2 balanced output can handle the load?

Yes, the Voxativ sounds marvelous with the Dave indeed.

Best,
Jason

 

[Rob Watts]

Yes near field is much more likely to work fine.

 

[flyte3333]

Thanks Rob for the ultra fast linear phase response.

Fixed that up for you !