ifi Zen DAC and Headphone Amp Review

[raif71]

 

[audiofool]

I've been waiting for this review, thanks Amir!
Looks like results are similar to other ifi products using the same dac chip. Miska, developer of HQPlayer has measured some of the iFi products - they all do substantially better in DSD mode. Any chance you could run the measurements using HQPlayer with DSD 256, the trial version would work. Also, the version of firmware is critical I believe it is 5.2b that needs to be run to get the best performance from DSD. Definitely a fidley product but the cheapest I have found that runs native DSD.

 

[Thomas savage]

Nothing wrong with the 50s and 60s for style and performance...

Grace Kelly
View attachment 38858
Honor Blackman
View attachment 38859
Tania Mallet
View attachment 38861

Congratulations on getting PussyGalore into the discussion.

 

[ishouldbeking]

I have an SMSL SU-8 which has RCA Out, are there any amps with rca in to amplify

There are plenty of great headphone amps with RCA inputs.

And even if there weren't, you just use an RCA to 3.5mm cable, which are cheap and readily available. If you just need a good amp at a good price to go with your SMSL, I'd get a JDS Atom and be done with it.

 

[amirm]

Miska, developer of HQPlayer has measured some of the iFi products - they all do substantially better in DSD mode. Any chance you could run the measurements using HQPlayer with DSD 256, the trial version would work.

That would be for a future project to evaluate such differences. For now, by the time you pay for US $240 for HQPlayer, you might as well buy a much more performant DAC and not need a powerful PC for HQPlayer to boot.

 

[AnikBanic]

The firmware you mean? 5.3 is the default one yeah so that employs the default TI filter which Amir measured here.

The 5.3C version uses the GTO filter instead.

It is quite awkward to need firmware updates to change the filter, but I understand its because they are using the XMOS chip to implement the GTO filter so no way to get around it.

They do also offer custom XMOS drivers on their website which I would recommend over the default windows ones too. They give you a control panel that allows you to chose the USB streaming latency. For some reason it defaults to safe rather than standard. Not sure if that impacts the measurements or not, but as I understand it the lower latency should in theory give less jitter though you have to be careful not to go too low or it causes drop outs. I set it for low latency and havent had any issues with dropouts so just left it on that.

i got firmware update failed wtf feelsbad

 

[AnikBanic]

PEOPLE ON WINDOWS 10 IF YOU WANT TO FLASH FIRMWARE TO 5.3C MAKE SURE YOU RIGHT CLICK THE FIRMWARE FILE AND PRESS UNBLOCK OTHERWISE IT WON'T FLASH

It flashed successfully now

 

[Zek]

Did you lose the color change for various files after flashing?

 

[AnikBanic]

I'm still geting pop when i'm pausing a song and resuming it idk why

 

[AnikBanic]

Did you lose the color change for various files after flashing?

haven't tried

MQA is magenta

 

[tential]

I'm quite sad you bought this product just last week considering there is other gear sent to you that hasn't been tested yet.

It's your own time though so my opinion hopefully doesn't hurt you.

 

[amirm]

I'm quite sad you bought this product just last week considering there is other gear sent to you that hasn't been tested yet.

It's your own time though so my opinion hopefully doesn't hurt you.

I make a judgement call on sequence of gear that I test. New gear often gets priority due to the fact that it impacts many people who are waiting to purchase them or not.

Other intervening factors are when someone is within a return period for some product so I have to test gear quickly.

I usually promise about 4 to 5 weeks to test equipment so that I have the leeway to switch things around for optimal execution.

 

[audiofool]

That would be for a future project to evaluate such differences. For now, by the time you pay for US $240 for HQPlayer, you might as well buy a much more performant DAC and not need a powerful PC for HQPlayer to boot.

HQPlayer is just an example, there is free software such as mansr's port of sox that can output dsd. Even the RME ADI-2DAC measures better using native DSD so I look forward to seeing your results someday. Thanks again.

 

[Raimei Templar]

Did you lose the color change for various files after flashing?

Yeah if your using GTO you will lose that as well as the "bit perfect" aspect of the dac. It always shows as MQA because its basically doing some sort of strange conversion of all your music into MQA. One of the reasons I am a bit suspicious of it lol.

The whole logic behind the use of the Burr-Brown chipset IFI uses is that it is a hybrid of delta-sigma and multibit that is able to process both DSD and PCM without any kind of conversion. idea being you will get both the "multibit magic" as well as the higher end delta-sigma performance. And as measurements show it does indeed outperform a straight multibit dac.

If your just going to convert anything seems like you might as well use a higher end Sabre/AKM chip at that point.

 

[Veri]

Yeah if your using GTO you will lose that as well as the "bit perfect" aspect of the dac. It always shows as MQA because its basically doing some sort of strange conversion of all your music into MQA. One of the reasons I am a bit suspicious of it lol.

The whole logic behind the use of the Burr-Brown chipset IFI uses is that it is a hybrid of delta-sigma and multibit that is able to process both DSD and PCM without any kind of conversion. idea being you will get both the "multibit magic" as well as the higher end delta-sigma performance. And as measurements show it does indeed outperform a straight multibit dac.

Holo Cyan/Spring still measure best as R2Rs. iFi's choice of DSD1793 chip does not seem to be capable of better than 96dB SINAD best-case. Not exactly spectacular...

 

[Kane1972]

Holo Cyan/Spring still measure best as R2Rs. iFi's choice of DSD1793 chip does not seem to be capable of better than 96dB SINAD best-case. Not exactly spectacular...

Ifi apparently tested as many chips as they could before choosing the DSD1793. They say they believe it sounds equally as good with both PCM and DSD, where other chips do sound better than the 1793 with one of those formats but worse with the other, so it made sense.

Now I know their opinion on what sounds good may not mean anything to those who rely solely on measurements, but some designers do design with their ears as much as with scopes. The chips they use are not cheap, and they have pushed those chips way beyond what the data sheets say they are capable of (higher sample rates etc) so I’m sure it would be easier for them to use more modern chips that would instantly measure a lot better without much effort.

I’m not trying to say that I don’t care about measurements, I do, not an I saying Ifi’s methodology correct or not, but I do believe they do have some good tech know how, so I’m sure when they do decide to make a DAC with a 9038 or something, they will do a job that will impress with measurements.

 

[Veri]

Well DSD1793 are like $6 apiece, probably much cheaper with the kind of bulk amounts they'd order. They could indeed use a more modern chip at that price.

 

[Stellabagpuss]

Ifi apparently tested as many chips as they could before choosing the DSD1793. They say they believe it sounds equally as good with both PCM and DSD, where other chips do sound better than the 1793 with one of those formats but worse with the other, so it made sense.

Now I know their opinion on what sounds good may not mean anything to those who rely solely on measurements, but some designers do design with their ears as much as with scopes. The chips they use are not cheap, and they have pushed those chips way beyond what the data sheets say they are capable of (higher sample rates etc) so I’m sure it would be easier for them to use more modern chips that would instantly measure a lot better without much effort.

I’m not trying to say that I don’t care about measurements, I do, not an I saying Ifi’s methodology correct or not, but I do believe they do have some good tech know how, so I’m sure when they do decide to make a DAC with a 9038 or something, they will do a job that will impress with measurements.

I have to agree, it's not all about measurements. And alot of the measurements on audioscience are hinting that one unit is better than another, and thats below human threshold,l understand that audioscience is mainly about measurements.

 

[BDWoody]

l understand that audioscience is mainly about measurements.

Not exactly...its about verifiable science where possible.

It's to help people understand that much of the 'high end' audio world revolves around nonsense that people very much want to believe, whether or not it is remotely true or verifiable. They count on the lack of accountability, and simply make ridiculous claims that few will bother to check or verify.

Fortunately, it really isn't that mysterious, and those with interest can bypass the bullshit brigade and look at what may actually matter.

It certainly isn't all about the measurements, but since what you hear can be measured, it can help identify glaring problems or weaknesses in the audible realm.

 

[Veri]

Well said, @BDWoody can't agree more.