[991 Carrera 4S]

Hm, there was another thread about FiiO KA5 that had very interesting information from owners of KA5.
Unfortunately it seems to be deleted now (instead of merged into this thread here)?

While some third parties website publish measurement in much stringent condition, measurement done under THD+N < 1% is kind of an industrial wide practice. Besides, you can also find more detailed measurement data on FiiO official website as they are one of the first in the audio industry to actual publish measurement.

Hm, ok regarding „THD+N<1% is kind of an industrial wide practice”.

On the other hand, all values that iBasso claims for e.g. DC06 (320 mW @ 32 Ohm BAL, 125 mW @ 32 Ohm SE) seem to be measured at THD+N<0.002%.
FiiO seems to provide such measurements – which are of much more practical relevance than THD+N<1% – neither for KA5, nor for KA3.

I don't know if this is how FiiO has done in KA5, but you can use interleaving to combine the effect of multiple DAC chips to overcome what a single DAC chip can't do. iFi Audio does this on many of their DAC models (where per DAC chip can only do 192kHz, but can scale to 768kHz when interleaving multiple DAC chips).

Interleaving would lose the benefits of using two DACs in mono-mode? Or how does this work?

For your second question, the FiiO Control iOS version is not compatible with the KA5. But you could adjust all the menu options in the KA5 itselfs when connecting to iPhone.

But why is the iOS app not updated for KA5, whereas the Android app is updated?
Will the iOS app be updated to support KA5 in the future, or never?

The engineers did the special technical processing in KA5 so it could output 768k sampling rate stably.

But how exactly? By just downsampling the input? By some type of interleaving between the two DACs? By some type of magic?
And what are the drawbacks?

I've had the KA5 for 24 hours now. I'm not one to give reviews but from a purely functional point of view, all functions available on the android app are also physically accessable from the hardware itself. It's not even complicated.

Thats great!

 

[ClieOS]

Hm, ok regarding „THD+N<1% is kind of an industrial wide practice”.

On the other hand, all values that iBasso claims for e.g. DC06 (320 mW @ 32 Ohm BAL, 125 mW @ 32 Ohm SE) seem to be measured at THD+N<0.002%.
FiiO seems to provide such measurements – which are of much more practical relevance than THD+N<1% – neither for KA5, nor for KA3.

1% is not some trivial pick. The audio industry settles on 1% because studies shows that 1% is where distortion start to become audible / detectable to most people. Therefore it has a significant meaning in telling the user just how much maximum power they can get before distortion becoming an issue and thus why it becomes an industral wide practice to publish this figure. There is another power measurement - the cleanest power (which usually has lower THD+N than 1%) - which tell us the output power when the distortion is at its lowest. This number tells us the optimum power the amp is able to make under a certain load. It has much less real world usefulness because user don't really follow what is theoretically the most optimum setting (by the amp PoV) but rather set the volume (and thus power) to match their own listening habit. But because cleanest power is often much easier to achieve a nicer looking number, some manufacturer will opt to publish it instead. This is where you need to look at a complete measurement report and understand how to read measurement correctly to tell whether a manufacturer is playing trick with number or not.

Interleaving would lose the benefits of using two DACs in mono-mode? Or how does this work?

Interleaving is far too complex for a non-expect like me to explain in detail, but Google can provide a lot of white papers. However, for a very simplified way to explain this, say one CS43198 has two separate DAC cores inside (*one intended for left and one for right channel) with a maximum sampling rate of 384kHz each. You can set it up so that one of the core is oversampling from 0Hz to 384kHz, then the other core is oversampling from 384.1kHz to 768kHz. The results are then recombined back at the end and you can get the full 0Hz to 768kHz in one channel, then the process is repeated on another CS43198 to get the other channel (*remember we have two CS43198 in K5, with 4 DAC cores all together). Of course the actual interleaving will be much more complicated than what I have said, but it is kind of a basic 1+1=2 in the heart of it all.

 

[FiiO]

I've had the KA5 for 24 hours now. I'm not one to give reviews but from a purely functional point of view, all functions available on the android app are also physically accessable from the hardware itself. It's not even complicated.

It also does 768k.

Dear friend,

Yes. Thanks for sharing.

Best regards

 

[cleg]

I must admit that the last few pages are super fun to read. Anyway, here is my definitely not full-fledged review of KA5

 

[FiiO Kang]

Could you please quote power consumption in mW.

Work in normal listening. It's about 500-600mW.

 

[FiiO Kang]

550 mW in standby mode ???

What about when in use, balanced ?

At normal volume. The balanced is similar

 

[FiiO Kang]

Just take the number FiiO posted and multiply by 5?

P=V*A

 

[ClieOS]

All the dual CS43198 glory.

 

[jjxaker]

here is my definitely not full-fledged review of KA5

naturally, since this is just a chatter but most of this content is enough, there is no dispute.

 

[amosamayo]

does anyone have a comparison of the KA5 with the rest of the competition, like the L&P W2, Questyle M15, Cayin RU6?

 

[marekm]

...and iBasso DC04 Pro !

 

[ClieOS]

For those of you who care about measurement: RMAA (32 ohm load via E1DA) result is super clean. Output impedance on both 3.5mm and 4.4mm are around 0.2 ohm or so, with 4.4mm offering roughly double the power of 3.5mm on the same volume setting. Power consumption is around the same ballpark as FiiO's numbers. As far as measurements go, KA5 seems quite excellent..

Filters all look pretty flat and similar, except for NOS (*which is as expected).

 

[TommyGeenexus]

For those of you who care about measurement: RMAA (32 ohm load via E1DA) result is super clean. Output impedance on both 3.5mm and 4.4mm are around 0.2 ohm or so, with 4.4mm offering roughly double the power of 3.5mm on the same volume setting. Power consumption is around the same ballpark as FiiO's numbers. As far as measurements go, KA5 seems quite excellent..


Filters all look pretty flat and similar, except for NOS (*which is as expected).

Thanks, finally some more measurements (seem to be in line with what FiiO has released).
I'd be very much interested in clipping measurements (THD+N vs Power), is that something you can measure with your setup?

 

[rafaelo]

Hi, can someone please comment how is the behaviour of the volume control when is connected to an iPhone?

For example, my dc03 pro and xDSD (with usb) when connected to my iPhone I can only control the volume from the buttons (or volume control) of the device(dac) and I cannot change it from the iPhone at all. (Personally, i vastly prefer this case)

Of course with KA3 that does not have buttons I can change volume with the iPhone volume control buttons.

What is the case with KA5?

 

[ClieOS]

(THD+N vs Power)

That will need a software audio analyzer that are super expensive, so no.