[Slim1970]

I also got the Burson Swing in house. Right now I have it paired up the Cavalli Liquid Carbon 2.0 and the HeadAmp Gilmore Lite Mk2. I have the V6 Classics in the I/V stage and the V6 Vivids in the LP stage. All I can say is good grief is this DAC musical, dynamic and detailed.

Against the Hugo 2 it loses out in transparency and ultimate resolution. But sometimes you just wanted to enjoy the music instead of analyzing it. The Swing does a fantastic job of delivering a musical experience. I like this DAC a lot.....

 

[Joong]

It is very interesting that single Ess9038 q2m of 2 channel chip can compete against quad or octa core channel chip version dac.
It reminds me of Schiit's claim while ago that 32 bit resolution of dac would be a joke when they introduced Yggdrasil dac of 21 bit supported by their special closed form filters, those combination of which surpassed many top-class dacs (of 32bit) of that time. This indicates that we are not hearing dac chip alone but every cascaded connections .ie. system.

 

[Slim1970]

It is very interesting that single Ess9038 q2m of 2 channel chip can compete against quad or octa core channel chip version dac.
It reminds me of Schiit's claim while ago that 32 bit resolution of dac would be a joke when they introduced Yggdrasil dac of 21 bit supported by their special closed form filters, those combination of which surpassed many top-class dacs (of 32bit) of that time. This indicates that we are not hearing dac chip alone but every cascaded connections .ie. system.

Agreed, having a good chip is only part of the story. It’s the implementation with component and system integration that net you good, quality sound. The Swing is very well built and designed.

 

[raoultrifan]

It is very interesting that single Ess9038 q2m of 2 channel chip can compete against quad or octa core channel chip version dac.
It reminds me of Schiit's claim while ago that 32 bit resolution of dac would be a joke when they introduced Yggdrasil dac of 21 bit supported by their special closed form filters, those combination of which surpassed many top-class dacs (of 32bit) of that time. This indicates that we are not hearing dac chip alone but every cascaded connections .ie. system.

In reality this 21-bit DAC may not even have 16-bit of measured resolution, although not sure this actually matters so much for us, but this is happening with PLAYMATE and most other DACs discussed here, mostly because of a higher THD+N. But the most important thing is linearity and accuracy of the output sound, how our ears is perceiving the final songs coming out from the speakers/headphones and a completely dark background (free of hum or background noise when music's not playing).

 

[Joong]

In reality this 21-bit DAC may not even have 16-bit of measured resolution, although not sure this actually matters so much for us, but this is happening with PLAYMATE and most other DACs discussed here, mostly because of a higher THD+N. But the most important thing is linearity and accuracy of the output sound, how our ears is perceiving the final songs coming out from the speakers/headphones and a completely dark background (free of hum or background noise when music's not playing).

It is ambiguous in the definition of the resolution. When the resolution of 21 bits in 1 hz, one sinusoidal wave can be composed of 2 million points. 20khz gives 100 points for the sinusoidal. A 32 bit dac has 200 k points in 20k hz, and this must be joke due to too much resolution at this almost limit zone of audible frequency.
If this is ok interpretation, then 21 bit is more than enough with good low pass filter.
Can anyone give some explanation for the definition of the resolution in audio?

 

[raoultrifan]

I was referring to ENOB: https://en.wikipedia.org/wiki/Effective_number_of_bits, which is measurable and it actually represent the no. of bits of an audio device.

 

[Joong]

I was referring to ENOB: https://en.wikipedia.org/wiki/Effective_number_of_bits, which is measurable and it actually represent the no. of bits of an audio device.

That resolution or dynamic range is for a measure for geometric or spacial one. When each value of the digit, there is corresponding discernible level of potential or voltage.
It is easily understood for geometric mapping between digit and analog value.
However audio dac is dynamic one so that Dac should be able to match any numeric value at given speed to a voltage for example.
In this case, the high speed information like numeric value at speed of 20khz has no more fine value like 32 bit on static case.
I don't know how to define the dynamic resolution for a given dac.

Anybody knowing this for how to define an audio dac resolution or dynamic range?
This understanding gets us better appreciation of humble dual channel 9038q2m being similar range of dynamic resolution that might matches upto multi channel version of chips like 2 9038q2m or the pro version which has 8 channels.

 

[Joong]

I ordered linear power supply for Swing and Fun, which are worth supported by good DC feed.
By my experiences with audio gears, good power influences sound quality by providing stable ground and power at the same time.
The linear power might replaces the weakest link in Swing and Fun which have switching power supply in stock form.

https://www.ebay.ca/itm/50VA-HIFI-U...24V-LPS-PSU-/273372400529?txnId=2174090054017

 

[raoultrifan]

With PLAY v1.6 there might be some differences, but the new and improved power design from PLAY v2.2, PLAYMATE and SWING will, most likely, not be influenced by the input power supply. Worth mentioning that included SMPS power brick from BURSON is really low noise, so spending money on an external power supply might not be so worthy.

PLAY v2.2 with default SMPS

PLAY v2.2 with my DIY Linear PSU​

The only difference is on the 3rd harmonic that decreases from -102dB to about -120dB (which is a, good thing actually), but the THD is barely improved. Of course, this improved will not be audible in any way.


Around 400uV of AC RMS & noise on my DIY Linear PSU on 12V/3A resistive load. PLAY/PLAYMATE/SWING will only take about 0.6A when playing, so LPSU's noise will even be lower, barely measurable by my scope.

I personally do not recommend swapping the default SMPS power brick provided by BURSON, my main concern being related to loss of the warranty, combined with zero audible improvements (although, slight improvements might appear on dedicated measurements).

Please, do understand that inside PLAY/PLAYMATE/SWING/FUN there are dedicated power regulators that care little to none about the main power supply used! Worth telling that these regulators might not like so much the Linear PSU used by yourselves and the possibility of oscillations are raising some concerns (it happened to me with FUN where inside regulators didn't liked my DIY linear power supply).

L.E.: Actually, Alex from BURSON told me that swapping the power brick will not void the warranty, so feel free to knock yourself out, just not expect any miracles.

 

[raoultrifan]

[...]
Anybody knowing this for how to define an audio dac resolution or dynamic range?
This understanding gets us better appreciation of humble dual channel 9038q2m being similar range of dynamic resolution that might matches upto multi channel version of chips like 2 9038q2m or the pro version which has 8 channels.

It's all stated here: https://en.wikipedia.org/wiki/Effective_number_of_bits, just need to find the SINAD. STEPS/ARTA software is able to that for you, all you need is a decent ADC to use it and do some measurements on 1V or 2V RMS 1KHz sinewave. Several DACs fail to achieve 16-bits of audio resolution, but most are passing above 16-bits, although not sure this is audible by our ears and headphones. After all, we mostly care about background noise and transparency in our day-to-day listening.

 

[Joong]

It's all stated here: https://en.wikipedia.org/wiki/Effective_number_of_bits, just need to find the SINAD. STEPS/ARTA software is able to that for you, all you need is a decent ADC to use it and do some measurements on 1V or 2V RMS 1KHz sinewave. Several DACs fail to achieve 16-bits of audio resolution, but most are passing above 16-bits, although not sure this is audible by our ears and headphones. After all, we mostly care about background noise and transparency in our day-to-day listening.

It makes sense in that the range refers to the speed of the signal, and this relation reveals temporal resolution.
In 1khz, the resolution of 16bit covers one cycle of sinusoidal signal by 65k points, and hence in 20khz the resolution covers by 3k points, which is way beyond discernible level of human ears in even 16bit dac. Because 2 points per cycle in one cycle of sinusoidal can be reconstructed back to the original signal by proper filter or integrator.
With 16 bit being way more than enough, 32 bit should be joke!
In some sense, 21 bit dac may be also a joke too.

 

[raoultrifan]

I think you are missing the point: a 21-bit DAC can actually have 14-bits of audio resolution. Please re-read my above posts and also about SINAD and ENOB.
Basically, the DAC itself can resolve 21-bits or even 32-bits audio, but the final resolution is also affected by the analog parts (transistors, opamps etc.) and power supply noise as well.

As for the 16-bit vs. 24-bit vs. 32-bit resolution for audio files, I guess there are many dedicated thread about this matter. From my understanding, 16-bits files are perfect in a perfect world, but if you want to have some more volume headroom, then 24-bit will ensure you enough bits for that. I personally vote for 24-bit audio resolution, especially if I want to post-process some files, but also to care not if a file was recorded to -6dB instead of -1...0dB.

As for the 32-bits DACs, there's no other way to have built-in digital volume control, but to build 32-bits DACs. Otherwise, there' not enough bits to lower the volume and not add background noise. Essentially, it's all in here: http://www.esstech.com/files/3014/4095/4308/digital-vs-analog-volume-control.pdf.

 

[Joong]

I think you are missing the point: a 21-bit DAC can actually have 14-bits of audio resolution. Please re-read my above posts and also about SINAD and ENOB.
Basically, the DAC itself can resolve 21-bits or even 32-bits audio, but the final resolution is also affected by the analog parts (transistors, opamps etc.) and power supply noise as well.

As for the 16-bit vs. 24-bit vs. 32-bit resolution for audio files, I guess there are many dedicated thread about this matter. From my understanding, 16-bits files are perfect in a perfect world, but if you want to have some more volume headroom, then 24-bit will ensure you enough bits for that. I personally vote for 24-bit audio resolution, especially if I want to post-process some files, but also to care not if a file was recorded to -6dB instead of -1...0dB.

As for the 32-bits DACs, there's no other way to have built-in digital volume control, but to build 32-bits DACs. Otherwise, there' not enough bits to lower the volume and not add background noise. Essentially, it's all in here: http://www.esstech.com/files/3014/4095/4308/digital-vs-analog-volume-control.pdf.

I might learn from you. Please teach me if I am wrong.
Thanks for your revert.
I am sensor engineer who always care for enough resolution in high phase speed so that we can track the signal without missing part.
Any ADC or DAC spec in sensor engineering world should be defined in terms of the speed or signal bandwidth.
For instance 1khz signal band, we need 2khz sampling in order to construct 1khz signal by 2 points construction, which means that at the 1khz two points is enough to recover the signal.
For me 16 bits in 1khz is great value if spatial resolution is 2^16 points for one entire cycle.
It is easy to resolve any static signal to great resolution, but when signal speed is high it is entirely different story and it requires high speed of sampling clock.

 

[raoultrifan]

We all learn something from each other, this is the beauty of such a forum, thank you!

In audio field things are a bit different; basically, if your analog stage (after the DAC chip) has background noise introduced into the signal path, then your THD+N will be lower.

Example:
Let's have a DAC with a theoretical THD of 0.0001%, which is 120 dB of SINAD (you can use http://www.sengpielaudio.com/calculator-thd.htm for conversion); that represents an ENOB of 120/6=20 bits of audio resolution.
After adding the THD and the noise of the I/V stage + Low Pass stage + the output buffer stage we might have a THD + N of let's say 0.005% which is 86 dB of SINAD; that represents and ENOB 86/6=14 bits of audio resolution.
Basically, because of the added noise and THD of the analog parts the final resolution is way lower than DAC chip's resolution.

However, more than 20-bits of audio makes no sense at this moment. This is about maximum resolution that can be measured by existing equipment (e.g.: Audio Precision).

 

[Joong]

With PLAY v1.6 there might be some differences, but the new and improved power design from PLAY v2.2, PLAYMATE and SWING will, most likely, not be influenced by the input power supply. Worth mentioning that included SMPS power brick from BURSON is really low noise, so spending money on an external power supply might not be so worthy.

PLAY v2.2 with default SMPS
PLAY v2.2 with my DIY Linear PSU​

The only difference is on the 3rd harmonic that decreases from -102dB to about -120dB (which is a, good thing actually), but the THD is barely improved. Of course, this improved will not be audible in any way.


Around 400uV of AC RMS & noise on my DIY Linear PSU on 12V/3A resistive load. PLAY/PLAYMATE/SWING will only take about 0.6A when playing, so LPSU's noise will even be lower, barely measurable by my scope.

I personally do not recommend swapping the default SMPS power brick provided by BURSON, my main concern being related to loss of the warranty, combined with zero audible improvements (although, slight improvements might appear on dedicated measurements).

Please, do understand that inside PLAY/PLAYMATE/SWING/FUN there are dedicated power regulators that care little to none about the main power supply used! Worth telling that these regulators might not like so much the Linear PSU used by yourselves and the possibility of oscillations are raising some concerns (it happened to me with FUN where inside regulators didn't liked my DIY linear power supply).

L.E.: Actually, Alex from BURSON told me that swapping the power brick will not void the warranty, so feel free to knock yourself out, just not expect any miracles.

My LPS is coming soon...
We will see how LPS can add value.
SMPS has endless harmonics and some of them are nonlinear transfer to audible band, so that when Anti-alias filter is not properly implemented they are still influential even though their band is way beyond audio-band.