On the Distortion of Cirrus Logic CS431xx-Based Devices: A Comparative Review

[CedarX]

HiBy W3 II has also been tested in a Bluetooth mode. Detailed measurements of all Bluetooth codecs included in it are published here, as well as the first published measurement of the UAT codec developed by HiBy.

OT, but the UAT codec measurements are very interesting: it’s not that good in the end compared to LDAC. Personally, I’ve never been able to make that codec work reliably (using HiBy-only devices).

 

[reallystate]

OT, but the UAT codec measurements are very interesting: it’s not that good in the end compared to LDAC. Personally, I’ve never been able to make that codec work reliably (using HiBy-only devices).

It's especially interesting what happens to the frequency response depending on the signal level.

 

[Jalan]

Hi, I'm sorry but i don't really understand all HiBy W3 II measures. If someone can globally explain how good/bad they are comparing other CS43131 devices (assuming LDAC and just one CS43131 power limitations). Thx !

 

[reallystate]

Hi, I'm sorry but i don't really understand all HiBy W3 II measures. If someone can globally explain how good/bad they are comparing other CS43131 devices (assuming LDAC and just one CS43131 power limitations). Thx !

It has a sound defect called Currus Hump, although you might not hear it. Despite this, when using the LDAC codec, it is quite good in terms of sound quality. It does have the ability to be used via USB, but the quality in this case will be worse than the cheapest dongle on the CX31993. Well, the main problem with using it is that, regardless of the connection method, it has an extremely uneven volume control scale. And it doesn't have a parametric equalizer. Therefore, even in the case of using an exclusively unbalanced connection, I recommend considering purchasing at least Snowsky Retro Nano / FiiO BTR13.

 

[Jalan]

Thank you for your answer !

Why this difference between LDAC and USB on the HiBy ?

Snowsky Retro Nano looks good, and Iove the removable battery but I can't find any measurment but with two CS43131 can we assume ir could be near the Melody ?

Moreover it is more expensive than BR13 (desktop), no measurment neither but we could assume Fiio try to have more quality in desktop product...

 

[confucius_zero]

it would probably interest Hiby owners.

did someone contact officially Hiby to get an answer,Maybe its a hidden future or update ?

^you were the curious one

 

[reallystate]

Snowsky Retro Nano looks good, and Iove the removable battery but I can't find any measurment...

I know that soon one person will start taking measurements of this device and perhaps publish them on this forum.

 

[confucius_zero]

Snowsky Retro Nano looks good

and DRE can be disabled on the device's menus it seems. I have one on the way but I'm not in a position to provide measurements unfortunately. I've been using a melody (dre can be turned off on that dongle).

I've asked shanling to see if the H0 and H2 has DRE-switch-off options and the brand doesn't seem to be concerned of this issue, so it doesn't consider a firmware update to add this feature.

 

[pioterd]

"DRE off" should not be considered as a feature, but as a lazy-shortcut instead of properly programming the DAC (lowering DRE activation threshold to around -45 or even -50dBFS). But people get what they get, and sometimes are even grateful for that..

 

[IsaacOscar]

"DRE off" should not be considered as a feature, but as a lazy-shortcut instead of properly programming the DAC (lowering DRE activation threshold to around -45 or even -50dBFS). But people get what they get, and sometimes are even grateful for that..

I don't know about the Retro Nano, but I suspect it's the same as the Melody where Fiio did what you suggested as well (i.e. with the firmware updated, the DRE activation threshold is low, but you can also switch it off completely).

 

[Jalan]

Tha'ts a good news...but there always could be other "surprises" than DRE so measurments are great

 

[pioterd]

However, we see one suspicious register appear on the CS43131 datasheet:

https://statics.cirrus.com/pubs/proDatasheet/CS43131_DS1155F2.pdf

This DRE is in the Table 3-6 (continuation of table 3-5), which partly explains how "wideband flatness mode" works for specific condition:

Dynamic range
(MCLK = 19.2 MHz,
MCLK_SRC_SEL = 01,
MCLK_INT = 1).

I insist on reading datasheets more carefully. Not only looking for "DRE".

"
4.2.2 Wideband Flatness Mode
The CS43131 specifically optimizes for wideband flatness playback, which is enabled by PCM_WBF_EN bit. This mode
should only be enabled when PCM plays back at 192 kHz. When operating in other speeds or modes, this bit should be
properly disabled. If wideband flatness mode enabled, the filter spec complies to Table 3-10. There is no option for filter
roll-off or phase response in this mode. NOS filter mode should be disabled for proper operation. Note that wideband
flatness mode can only be enabled or disabled when PDN_HP is set.
"
7.5.2, page 128 od CS43131 datasheet (DS1155F2):
PCM_WBF_
EN
Wideband flatness mode enable. This should only be used in PCM playback when xSP sample rate is at 192 kHz. This
bit must be changed while PDN_HP is set.
0 (Default) Wideband flatness mode disabled
1 Wideband flatness mode enabled

"

Now, maybe readers shall be suspicious if you, @jkim , are the one who wants for any reason discredit CS431XX ? The problem is not in this Dac/Amp chip, but in the way some audio companies implement and program it. Maybe datasheet is too long for some? Or too many additional appendix files?

Now they will add "DRE switch" to their devices menus, but in fact, the problem is not in DRE. So what does this switch do actually?

I have Fosi DS2, new Hiby FC4 CS43131, Shanling UA mini - NOT a single click or anything unwanted playing both Song of the Sisters and Original_Cmaj files. All filters checked (if possible; in DS2 not possible to change filter). Volume attenuated using only hardware buttons (Fosi and Hiby) or through EddictPlayer app.

Look at HV_EN parameter - I think that problem with this "hump" may be related to voltage/gain settings somehow. Analog Output Characteristics / PCM and DSD Processor Mode Parameter / Wideband Flatness Mode Parameter

 

[MaxwellsEq]

The problem is not in this Dac/Amp chip, but in the way some audio companies implement and program it

That's a good perspective. If DRE is disabled and the correct filter used, these generations of DACs still have a performance that is superior to human hearing. Some manufacturers did a poor implementation - the poor implementation measures extremely well on a test bench but sounds bad on some music.

 

[zany.ninja]

Now, maybe readers shall be suspicious if you, @jkim , are the one who wants for any reason discredit CS431XX ? The problem is not in this Dac/Amp chip, but in the way some audio companies implement and program it. Maybe datasheet is too long for some? Or too many additional appendix files?

That's an unfair statement to throw at @jkim . He has repeatedly said that the CS431XX is an excellent chip, and that DRE is not inherently flawed, it's just that it's possible at worst Cirrus may have set the default DRE values too aggressively, and many audio companies perhaps didn't bother to change them. Or perhaps it could be another implementation issue. But he hasn't tried to discredit Cirrus in any shape or form, but rather he has put a lot of time and effort into testing, measuring, and exploring the source of the distortion problem, and arriving at possible causes and solutions. He has been extremely fair, balanced, and measured in his assessments and statements. Your attack on him is uncalled for.

The datasheet that you think holds all the answers absolutely does not, there is a lot more documentation under strict NDAs which contain trade secrets that Cirrus obviously wouldn't want to be made public.

Now they will add "DRE switch" to their devices menus, but in fact, the problem is not in DRE. So what does this switch do actually?

I have Fosi DS2, new Hiby FC4 CS43131, Shanling UA mini - NOT a single click or anything unwanted playing both Song of the Sisters and Original_Cmaj files. All filters checked (if possible; in DS2 not possible to change filter). Volume attenuated using only hardware buttons (Fosi and Hiby) or through EddictPlayer app.

Look at HV_EN parameter - I think that problem with this "hump" may be related to voltage/gain settings somehow. Analog Output Characteristics / PCM and DSD Processor Mode Parameter / Wideband Flatness Mode Parameter

You're just speculating. Do some actual science to back it up. @jkim did actual tests and measurements - very thoroughly, with proper testing equipment and software, before and after firmware updates, with the DRE switched on and off, of numerous devices - that show measurable differences/improvements of the DRE updates. I know who I'd believe.

Whatever values they're changing under the hood with this "DRE switch", whether DRE_EN or HV_EN or whatever, device makers have implemented this DRE fix largely, if not entirely, thanks to the hard work of @jkim and really we should all be grateful to him for it.

 

[fws]

That's an unfair statement to throw at @jkim .

I second that. The issue been around for years, first reported by users as they heard audible distortion, but as not everyone did, it was assigned to their wider setup or a flawed unit. The simple fact that we now know how to pinpoint and test for the issue all due to amazing work jkim put into it. I don't see any displaced blame there, again CS431xx got plenty of praise including confirming they are still performing superb even without DRE, but one of the goals of this forum, at least as I see it, is to hold designers/manufacturers/brands accountable and not let them run with whatever corner cutting they apply just because customers at large won't notice it. It is because of this approach, while DRE switch was added to some products, at the same time DRE was adjusted to work fine in some of them. If not for this topic, an odd listener would still hear distortion and people would dismiss him. Worse yet, if left unchecked it might have compelled manufacturers to run DRE on other chips where it is turned off by default or use even more aggressive settings adding even more distortion.

Also, before throwing such accusations, it has to be considered that jkim's measurements likely boosted TRN Black Pearl sales through the roof. Likely other CS431xx products which were shown to work well from the get go or received firmware update, also gained some sales from this (even more so since TRN did bait and switch), so like Melody, DS2, KA15, JM20 Max - especially since simultaneously some of those were confirmed to perform superbly.

 

[pioterd]

That's an unfair statement to throw at @jkim .

I did not make any statement, just hipothesis, I totally disagree it may be unfair in any way. But sure, we have a right to share our thoughts. Fortunately science is more often based on hipothesis and relativity, rather than completely proved thesis, which are very rare.

before throwing such accusations, it has to be considered that jkim's measurements likely boosted TRN Black Pearl sales through

I did not make any accusation; and even if I did, why shall I consider boosting sales of TRN devices as....as what?

It is good @jkim did measurements, but he also commented a piece of datasheet, without any proof that it is exactly source of the problem (that is speculation). I know that this datasheet is just a part of the whole documentation.

According to DS2, new FC4 and UA mini I have - I do not speculate. There is no audible distortion on my pc, notebook and smartphones with these devices and headphones I use.

 

[fws]

I've combed through FiiO's news about firmware updates listing DRE related changes, as well as line of products running Cirrus Logic chips. Most were reported, but just wanted to have a little summary.

FiiO products with DRE related firmware updates
FiiO KA15 (V1.0.5)
https://www.fiio.com/newsinfo/1072224.html
FiiO KA5 (V1.70)
https://www.fiio.com/newsinfo/1077813.html
FiiO M21 (V1.0.2)
https://forum.fiio.com/note/showNoteContent.do?id=202505140958060216556
Snowsky Retro Nano (V6.0)
https://www.fiio.com/newsinfo/1110321.html
Snowsky Melody (V1.1.0)
https://www.fiio.com/newsinfo/1074912.html
Snowsky Echo Mini (V2.4.0)
https://www.fiio.com/newsinfo/1077417.html
JadeAudio JM21 (V1.0.8)
https://www.fiio.com/newsinfo/1087737.html

Reported DRE issue, no fixes
FiiO BTR13 (CS43131)
FiiO K11 (CS43198)
https://www.audiosciencereview.com/...ble-dac-amp-review.68800/page-10#post-2518924

Needs clarification
FiiO KA13 (CS43131) - reported no DRE issue in jkim's list, but can't find the basis for it
https://docs.google.com/spreadsheets/d/1fYP8cqPMtZigR1VdPIHaax5nyGDaHZpKeS1w3AKd5RI/edit?gid=0#gid=0

Unknown (no fixes, no measurements) products released prior the first fixes
FiiO Q11 (CS43198)
FiiO DM13 (CS43198)
FiiO KA2 (CS43131)
FiiO KA11 (CS43131) - possibly ok due to similarity to JCAlly JM20 Max
FiiO KB3 (CS43131)

Unknown new products released after the first fixes (if someone has such, please report if DRE switch is there)
Snowsky DISC (CS43131)
Snowsky Echo (CS43198) - not to be mistaken with Echo Mini

 

[IsaacOscar]

I've combed through FiiO's news about firmware updates listing DRE related changes, as well as line of products running Cirrus Logic chips. Most were reported, but just wanted to have a little summary.

Thanks, I already posted a similar table here https://www.audiosciencereview.com/...vices-a-comparative-review.63038/post-2501644

although I didn't bother specifying which of the two problematic chips the devices have (I don't think anyones found a difference between the two?)

But it's been lost in the mass of other posts on this thread. Perhaps a mod (@RickS) can move my post to be at the start of this thread? Or @jkim can edit their first post to include the details?

Reported no DRE related issue despite no firmware updates
FiiO KA13 (CS43131)

Do you have a link I can add to my table?

FiiO M21 (CS43198)

This does have the DRE switch see here.

FiiO KA11 (CS43131) - possibly ok due to similarity to JCAlly JM20 Max

Nope, it does have the issue see here:

https://www.audiosciencereview.com/...11-portable-dac-amp-review.68800/post-2518924

Unknown new products released after the first fixes (if someone has such, please report if DRE switch is there)
Snowsky DISC (CS43131)
Snowsky Echo (CS43198) - not to be mistaken with Echo Mini

Thanks, I've added these to my table.

 

[fws]

Thanks, I already posted a similar table here https://www.audiosciencereview.com/...vices-a-comparative-review.63038/post-2501644

Oh, my bad then, somehow missed it or perhaps I saw it and forgot (just me), you did a more thorough job - kudos!

Do you have a link I can add to my table?

Nope, I actually started the write down because I went through the news to check if KA13 got update, and only noticed it was on the excel list just before posting without checking for the actual posts about it. My bad, could be it is not fixed, and I just went on to "confirm" a mistake. The only mention about KA13 seems to be here, and rather indicates the issue is there

On the Distortion of Cirrus Logic CS431xx-Based Devices: A Comparative Review

1: Using the Shanling app, I can set the dongle to High Gain or Low Gain. Which should I use for best quality output? i.e. lowest noise and distortion. I get adequate volume with either option for my use, and I use the dongle with easy to drive 28 and 60 Ohm IEMs and headphones, and also...

www.audiosciencereview.com

Maybe @jkim can shed some light?

Thanks, I've added these to my table.

I just did some fixes based on yours as well, maybe I should remove mine since I wont be able to edit later on.

Oh and the practical difference between the chips for the end user is just the impedance sensing in CS43131.

 

[IsaacOscar]

I just did some fixes based on yours as well, maybe I should remove mine since I wont be able to edit later on.

Yeah you can only edit within 24-hours of your post, unless you're a Forum Donor (or you ask a mod to edit it for you?)