[Rob Watts]

I wish. Davina project has been on the back burner a bit. The decimation filters (768>48 kHz) is done and tested and work perfectly. I am currently doing the 104.25MHz > 768/705.6. This has proved mighty tricky, as the decimation is 135.75; the tricky part has been arranging a filter to do non integer decimation, without the possibility of any added THD and noise, nor aliasing; but I have managed to do this, at least at the top level design. I set myself the target of zero aliasing from 104 MHz; it's actually better than 250 dB from 384 kHz to 104 MHz, and uniquely is a linear phase FIR filter not the usual cheap and nasty moving average decimation filter. Should have this finished when I get back. Then it's coding for the pulse array, then hardware testing will start. I have set myself the target of actual recordings for next year, as the project has been swamped due to work commitments.

 

[514077]

I am on holiday in Florida - taking the kids to the theme parks.

I was left with two feelings; on the one hand thinking that considerable progress has been made in reproduced sound - but that there is still a long way to go, to mach the visceral beauty and tangibility of live unamplified sound.

And I will need to come back to Florida to record this with the new ADC....

I'm wondering what your recording chain would be, along with the Davina? i.e. Mics, preamps, etc?

 

[Rob Watts]

Davina comes equipped with it's own mic pre-amp, with adjustable gain and phantom PSU. So only mics to worry about, and I probably will make up my own cable too.

 

[514077]

Davina comes equipped with it's own mic pre-amp, with adjustable gain and phantom PSU. So only mics to worry about, and I probably will make up my own cable too.

Oh, I didn't realize that Davina was self-contained. Well, I'll be curious about your mic and self-made cables when the time comes. Meanwhile, enjoy Florida, and I'll try not to bug you while you're there.

 

[Tims Fong]

There are two elements to MQA - firstly the conversion back to 88.2/96 from 44.1/48 - the so called unfold - and the the second part the low tap length minimum phase interpolation filter. Let's deal with the second issue first, the interpolation filter - I have compared the MQA interpolation filter to my WTA filter and it sounds very much worse than the WTA - soft, fat bloated bass, poor instrument separation and focus, with a flat ill defined sound-stage. So there is no way I would ever offer this as an option.

As to the compression from 88.2/96 to 44.1/48 this is seriously flawed with major sound quality and measurement issues. For one, it has a massive notch at 22.05 kHz or 24 kHz that is introduced, which will have transient timing repercussions; secondly the system has completely unacceptable aliasing issues, which means distortion at 20kHz is a massive 1% - and aliasing has a huge consequence to the sound quality too, as again it degrades transient timing; thirdly the system is lossy, and converts a 24 bit signal into something like 17 bits. This is again unacceptable.

My advice is to ignore MQA and always go for the unchanged original file as the WTA filter will do a much better job of reconstructing the transient timing information to a much higher ability than MQA can.

Thank you for your reply and I believe you are right.

 

[JaZZ]

I'm glad you are progressing with DAVINA, Rob, after the effort with the M-Scaler! Wish you recreative vacations with your family and a lot of fun with programming the Pulse-Array code.

 

[MacedonianHero]

I am on holiday in Florida - taking the kids to the theme parks.

Yesterday we were going to Animal Kingdom, to do the new Avatar ride (the 2 hour wait was worth it!) and see the light show afterwards. Anyway, a thunder storm interrupted us; so rather than drive in the rain, I sat on the balcony watching the approaching storm, and waiting for it to pass. I thought I would spend half an hour relaxing and listening to it.

It was a cracker of a storm; huge lightning strikes, and aural explosions as the thunder cracked.

But the sound was incredible - and the placement accuracy that human hearing and perception is simply immense. As the thunder cracked, you could hear exactly where the thunder originated, as it rippled and moved through the sky - you could perceive exactly where it was coming from to simply extraordinary accuracy - height and depth. Also, each thunder crack sounded completely different, with remarkable low frequency pitch and texture. Sat on the balcony of the seventh floor one can hear pool sounds, the hotel's music system, and place them to pinpoint accuracy; then this usual ambient sound was punctuated by the huge dynamics of the thunder, with a simply immense soundstage extending for many miles.

I was left with two feelings; on the one hand thinking that considerable progress has been made in reproduced sound - but that there is still a long way to go, to mach the visceral beauty and tangibility of live unamplified sound.

And I will need to come back to Florida to record this with the new ADC....

We were there in May (5th time at Disney World in 7 years) and we all absolutely loved the new Avatar ride! Got 2 fast passes for it over a week...simply amazing ride! It's to amusement park rides as the DAVE is to DACs (simply the best). .

 

[hmartin]

I wish. Davina project has been on the back burner a bit. The decimation filters (768>48 kHz) is done and tested and work perfectly. I am currently doing the 104.25MHz > 768/705.6. This has proved mighty tricky, as the decimation is 135.75; the tricky part has been arranging a filter to do non integer decimation, without the possibility of any added THD and noise, nor aliasing; but I have managed to do this, at least at the top level design. I set myself the target of zero aliasing from 104 MHz; it's actually better than 250 dB from 384 kHz to 104 MHz, and uniquely is a linear phase FIR filter not the usual cheap and nasty moving average decimation filter. Should have this finished when I get back. Then it's coding for the pulse array, then hardware testing will start. I have set myself the target of actual recordings for next year, as the project has been swamped due to work commitments.

I have always wondered why you picked 104.25 MHz instead of 112.896 MHz, maybe you also wonder this now.

Edit: Joke, if that was unclear.....

 

[514077]

I may be having a brainf**t, but wasn't TT2 supposed to use 208MHz instead of 104? Or was it just the number of taps that were doubled? I seem to remember 208 MHz being mentioned recently, but not sure.

 

[ecwl]

I may be having a brainf**t, but wasn't TT2 supposed to use 208MHz instead of 104? Or was it just the number of taps that were doubled? I seem to remember 208 MHz being mentioned recently, but not sure.

208MHz is the clock speed of the FPGA I thought.

 

[Currawong]

It's down to the quality of the Toslink tx and rx devices; they are simply not fast enough. There is nothing in principle using glass fibre to run at 768 kHz if the tx and rx were fast enough.

Why not use the same TX/RX units used for high-speed fibre networking instead?

 

[GryphonGuy]

Why not use the same TX/RX units used for high-speed fibre networking instead?

My uneducated guess would be that it would not be Toslink after you did that because Toslink is a standard by which you can purchase any device and cable and it will talk to another Toslink device. So you'd need to change the standard for Toslink and retrofit all toslink devices. Not an exercise most people would entertain who manufacture to the Toslink standard.

Regards
GG

 

[Currawong]

I probably should have been more specific in that my idea is NOT to use Toslink at all, as it is too outdated. What we really need is a new optical S/PDIF physical standard based around better hardware.

 

[jayz]

To add to what @Currawong says, I don't see a need to stick to any standard between MScaler and DAC - it is not like other manufacturers will want to support that link. There could even be a commercial advantage in having proprietary interface there and in fact, the dual BNC link may already be non-standard. Having said all that however, the most optimal solution might be to one day integrate the MScaler with the DAC with an internal optical coupler?

 

[Triode User]

And depending on cost it might even be that the equivalent of monoblocs next to each speaker with MScaler, dac and digital amp all in the same box will be possible. Then add in the possibility of digital crossovers and it could get very interesting.