Watts Up...?

Nov 14, 2019 at 12:33 AM Post #1,546 of 5,069
I am surprised to see many recent preamp / integrated amp reviews casually state that all analog inputs are digitized. Given the budget constraints, are these internal ADCs any good?

I suspect you have answered your own question when you mentioned ’budget constraints’. But what are these amps you are considering? I have never seen any pre amps which convert the input to digital. Names please?
 
Nov 14, 2019 at 1:05 AM Post #1,547 of 5,069
I suspect you have answered your own question when you mentioned ’budget constraints’. But what are these amps you are considering? I have never seen any pre amps which convert the input to digital. Names please?
The Devialet all-in-ones used to do the digitalisation, and then could do all sorts of dsp for loudspeaker correction.
 
Nov 17, 2019 at 5:53 AM Post #1,550 of 5,069
Hi Rob,

I have another little question that has been in my mind. I've been looking at crystals used by various gear and was curious about your decision to use that 104MHz oscillator. It doesn't divide down to any regular sample rates and I know your approach to sampling isn't so frequency dependent on the industry standard rates, but is there any story behind the choice or was it just in stock that day when you started your first FPGA prototype and it met the timing constraints? :)

The actual number is an integer multiple of 44.1 kHz and 48 kHz - and the reason I needed an integer multiple is for generating the clocks for USB, which need to be 44.1kHz and 48 kHz, and integer division (or rather frequency synthesis) is very much easier to do.

I am surprised to see many recent preamp / integrated amp reviews casually state that all analog inputs are digitized. Given the budget constraints, are these internal ADCs any good?

No they are not. But then neither are state of the art ADCs either. All suffer from large levels of aliasing, poor small signal accuracy and measurable noise floor modulation.
 
Nov 17, 2019 at 6:06 AM Post #1,551 of 5,069
No they are not. But then neither are state of the art ADCs either. All suffer from large levels of aliasing, poor small signal accuracy and measurable noise floor modulation.

Oh, have you ever thought of developing a true state of the art ADC yourself? :laughing::stuck_out_tongue_closed_eyes::stuck_out_tongue::wink::wink::wink:
 
Nov 17, 2019 at 12:34 PM Post #1,553 of 5,069
You should see some more positive news about the pulse array ADC next year.
That would be great! Are you aware how many people are waiting since how many years for your Davina?! :wink:
 
Nov 17, 2019 at 1:57 PM Post #1,554 of 5,069
The actual number is an integer multiple of 44.1 kHz and 48 kHz - and the reason I needed an integer multiple is for generating the clocks for USB, which need to be 44.1kHz and 48 kHz, and integer division (or rather frequency synthesis) is very much easier to do.
Hi Rob,

I must be missing something as I don't see the integer number based on standard rates.
AUDIO:
104254200 / 44100 = 2364.040816
104254200 / 48000 = 2171.9625
USB:
104254200 / 12000000 = 8.68785
104254200 / 48000000 = 2.1719625

Sorry, I am clearly missing something that you understand much better! I guess maybe this is a PLL synthesized value inside the FPGA? I remember you saying you avoid PLLs, but that is for the audio section because they are always hunting for the frequency and introduce jitter. It would be much less important in a USB PHY due to the ability to buffer data and process it through your WTA filters. I'm still completely confused as to how that 104MHz clock gets used to generate a 2048fs signal on the pulse array based around our industry standards of 44.1 and 48KHz as fs (but I am loving the challenge of trying to figure it out!). :)

No they are not. But then neither are state of the art ADCs either. All suffer from large levels of aliasing, poor small signal accuracy and measurable noise floor modulation.

I absolutely cannot wait to hear what the next result will be in your progress with the custom ADC. The recording industry is totally ready for a make over when it comes to audio conversion (assuming they will accept the initial costs) and there's such a hard limit with creating modern DACs because it is all based on playing back music made with conversion technology that is decades old!
 
Nov 17, 2019 at 2:07 PM Post #1,555 of 5,069
That would be great! Are you aware how many people are waiting since how many years for your Davina?! :wink:

I suspect a certain amount of “if you have to ask how much it is then you can’t afford it”. On the other hand it is difficult to envisage why I would want an ADC and in any case I tend to think of it primarily as a test piece of equipment for Rob before it is released to recording studios.
 
Nov 17, 2019 at 3:30 PM Post #1,556 of 5,069
Just in case: I'm seriously looking forward to it. Personally I may not have any use for it (well, who knows...), but every single Head-Fier and even more people outside would be able to benefit from better recording quality – provided that it will establish a new studio standard.
 
Nov 17, 2019 at 3:35 PM Post #1,557 of 5,069
Just in case: I'm seriously looking forward to it. Personally I may not have any use for it (well, who knows...), but every single Head-Fier and even more people outside would be able to benefit from better recording quality – provided that it will establish a new studio standard.
I wonder if it will run on batteries intentionally, for live music recording out in the wild.
 
Nov 18, 2019 at 4:07 AM Post #1,558 of 5,069
Hi Rob,

I must be missing something as I don't see the integer number based on standard rates.
AUDIO:
104254200 / 44100 = 2364.040816
104254200 / 48000 = 2171.9625
USB:
104254200 / 12000000 = 8.68785
104254200 / 48000000 = 2.1719625

Sorry, I am clearly missing something that you understand much better! I guess maybe this is a PLL synthesized value inside the FPGA? I remember you saying you avoid PLLs, but that is for the audio section because they are always hunting for the frequency and introduce jitter. It would be much less important in a USB PHY due to the ability to buffer data and process it through your WTA filters. I'm still completely confused as to how that 104MHz clock gets used to generate a 2048fs signal on the pulse array based around our industry standards of 44.1 and 48KHz as fs (but I am loving the challenge of trying to figure it out!). :)



I absolutely cannot wait to hear what the next result will be in your progress with the custom ADC. The recording industry is totally ready for a make over when it comes to audio conversion (assuming they will accept the initial costs) and there's such a hard limit with creating modern DACs because it is all based on playing back music made with conversion technology that is decades old!

The numbers are rounded, and the rounding error is much smaller than the 50ppm accuracy of the clock, and so inconsequential.
 
Nov 30, 2019 at 12:39 AM Post #1,559 of 5,069
So November was a busy month - Shanghai CanJam, Chord distributor conference in Phuket and finally Singapore for the AV show (which was very busy). And designing in-between events, with a huge workload on at the moment.

But the highlight was definitely getting to hear Mr. Speakers (now Dan Clark Audio) Aeon 2 and being at the Aeon 2 and new naming launch. You can see the thinking behind Aeon 2 here:

https://www.head-fi.org/threads/the-new-aeon-2-its-the-end-of-mrspeakers.918309/

Personally I have been a big fan of Dan Clark Audio, in particular with the original Aeons. When I first heard them in early 2017 I was knocked out by the accuracy, transparency and refinement - and on my next listen, at Singapore CanJam, I ordered a pair. The production units arrived in June, and I was struck by two things - they didn't need any break-in, and they sounded identical to the early prototypes I had heard. The only downside was the bass was too lean, but adding the tuning pads helped that.

I ought to state my needs for headphones - I do a lot of listening on flights, and a lot of my flights are 12 or 13 hours. So I need headphones that are comfortable, that you can use for ten hours or so without fatigue. Also, I absolutely need acoustic isolation. Finally, I need transparency, accuracy and refinement - and above all musicality (that is the ability to get emotionally involved with the music). The Aeons firmly ticked all of those boxes. What impressed me so much was the transparency - very convincing sense of depth (cross-feed must be on) and pinpoint imagery.

Fortunately, I managed to blag an Aeon 2 after Shanghai (thanks Andy Regan) and so I thought I might post my personal impressions on the Aeon 2.

Firstly, it's a radical re-design, with a clever folding band, and with the magnets flipped to the closed side. The SQ differences were not small either - this is no marginal change.

The first thing you noticed is the tonal balance - gone is the lean sound of the originals. Bass is much more extended, with a heavier more meaty and loudspeaker like tonal balance. Moreover, distortion seems lower - as the volume increases, the sound does not harden up or change - and I was surprised at how much better this aspect is against the original, as low perceived bass distortion was a strong point on the Aeon 1. This aspect is very important to me - low bass distortion - and it's a major failing of many respected high-end headphones.

Lateral imagery is similar to the originals, but depth - that ability to hear sounds outside your head - is better than before - but this was always a major advantage of the Aeon 1. Also, because the Aeon closed is closed, early reflections from the rear gives it an upward tilt for me to the soundstage - I guess about +20 degrees. This doesn't happen with open backed, and it was never a huge problem. With the Aeon 2, this effect is still there, but it's much less noticeable.

Another issue is the clamping force is stronger on the Aeon 2s - this is good, as it improves bass sealing, and acoustic isolation - but I certainly would not want the clamping force any greater! This might be an issue for big heads though...

So the $64k question is is the improvements worth buying an Aeon 2 when you already own an Aeon 1? In my view absolutely yes. Moreover, if you are in the market for a closed headphone, you need to listen to the Aeon 2 - even if your budget is very much larger than the asking price - they are that good.

Congratulations to Dan and his team at Dan Clark Audio - you have made my flights more enjoyable - and enjoying music is what this hobby is really about.
 
Nov 30, 2019 at 11:31 AM Post #1,560 of 5,069
very convincing sense of depth (cross-feed must be on)

Thanks for the review. Just curious, which crossfeed setting do you prefer?
 

Users who are viewing this thread

Back
Top