Breeze Audio DAC

[JonMarsh]

This test will cover the basic capabilities of a very inexpensive AK4490 DAC from Breeze Audio. The test sample was provided by Renron, and he's posted about this little guy in the Digital Audio section of the forum.







Let's start off with the 20Hz to 20kHz frequency response at -6 dBFS:



The result is within a quarter of a dB from 20 Hz to about 12kHz, but rolls off within the top octave substantially, being down almost 4 dB at 20kHz relative to the passband level.


Let's get a bit of an idea of what's going on in the low level area looking at an averaged FFT with signal at 1000Hz and -60 dBFS. (one wouldn't just want to look at the output with no input signal, as many DACs have muting circuits that kill the output completely to reduce the measured noise)




Knowing this is a Delta-Sigma DAC, it's pretty likely that the spuriae below -110dB are idle tone artifacts. That's at a very low level, so quantifying the audibility is difficult.

Next, we'll look at 50Hz with an input level of - 6dBFS -




Probably a true 2nd order harmonic at the -96 dB level. There's also a tiny bit of line frequency related noise at 120Hz- very low in level at -120 dB. But there is a fair amount of repeating stuff at 50Hz intervals that is not random noise.

What about the low level -90 dBFS 1kHz test? (toggling the LSB)




It's quite recognizable, though with some DC offset from the zero position. Typical Gibbs ringing is not visible readily visible.

Here's a somewhat better example in a $1K DAC, the Cambridge Audio 851D.






What about high frequency intermodulation and image rejection? The 19+20kHz CCIM test will sort that out, with 19 and 20kHz driven each at -6dBFS.


Things are not really right with the AudioBreeze DAC. This is what the output from the DAC looks like in scope mode:





This is what it should look like, with the proper levels and beat summing:




This is what the FFT looks like, and this tells the tale...




There is virtually no image rejection, and very high levels for the aliased image frequency set. Yes, the CCIM distortion side bands and 1kHz difference frequency are essentially non-existent (very good), but the image behavior is not acceptable.

OTOH, otherwise, it still may be one of the world's best $82 DACs... :W

[BobEllis]

Similar question as your other test, what audible defect would the image frequencies being so high in level cause? Trying to learn how to relate measurements to what I can expect to hear.

[JonMarsh]

When the image rejection is this low, what it means is that the entire band of audio signal is reflected up around 1/2 the sampling frequency, appearing at a lower, but in this case, substantial level. That is a lot of supersonic content, and tweeters and amplifiers don't really like that. It wold certainly stress out a Class D or tube amplifier. With more time, I could setup a test to quantify the image output level, but since spare time is something I'm short of, and this is an $82 dollar DAC, I didn't mess with it for now. If Ron wants to chime in that he'd like to see more analysis I'll take a look at it next weekend.

I was planning on shipping both units to Ron no later than Wednesday, unless something else comes up at work! (been a lot of that lately).

There was a Japanese boutique brand of modified CD players whose name I ' recall this morning, that had this issue also, as part of their Schtick was to "get rid of the time distortions" of the conventional filter. Well, that they did, but introduced significant other problems.

This sort of problem will increase the high frequency stress on the system, and depending on the performance of downstream components, may produce quite audible intermodulation distortion in the audio range. (audible compared with a well designed player, maybe not compared with an iPod or phone or DVD player).

[JonMarsh]

OK, I'm beating a dead horse, but once again, here's the Cambridge 851D, my personal reference for "inexpensive serious entry level but good" DAC's. Same test.




The way the test was done, it did capture the image region, where there was no spurious response; I just adjusted the X-axis in preparing the published image for clarity and more detail in the audio range.


BTW, I expect there are other very good DAC's in this price range (around $1K out the door), I just haven't tested them- and most of them probably don't have a very high grade analog volume control and multiple inputs configured as a digital preamp. For someone with an all digital setup, this 851D is a very good performer.

[Renron]

OTOH, otherwise, it still may be one of the world's best $82 DACs... :rofl:

The wife unit and I are still laughing out loud, looking at the waveforms is shocking, but not unexpected.
For the price paid, it works, about all that could be expected.
I wish I understood 25% of what was posted here and in the Modi MB thread. That said, I can only thank you for sharing your valuable time testing this polished turd of a DAC. Sometimes you really do get what you paid for.
Ron

[BobEllis]

I'm looking forward to your comparison of our DACs, Ron. That and your impressions of the difference Jon's high end stuff makes when you visit him with your Ardents.

[Renron]

Me too Bob, I'd really like to hear just how good the "mids" are on excellent equipment. We've set a tentative date sometime in December.
I'll get your DAC back to you very quickly once it arrives here. Thanks Bob!
In English (not Marsh speak) what did Jon think of the Modi MB ? I read the thread but................. Duh?
Ron

[BobEllis]

My take was that he didn't get a chance to listen to either and had technical concerns about both. Without having tested other entry level DACs there's not a way to get a feel for whether they'd sound better or worse than others in class.

Subjectively, we both like them better than the DACs in our sources. Whether that's euphonic or accurate remains to be seen. I've read a number of reviews that suggest that the Momby is a bit on the euphonic side -a bit warm. That's not how I'd describe the difference between it and my Oppo, though.

Take your time with the Momby. I'm in no rush. More interested in hearing your comparison.

[JonMarsh]

Me too Bob, I'd really like to hear just how good the "mids" are on excellent equipment. We've set a tentative date sometime in December.
I'll get your DAC back to you very quickly once it arrives here. Thanks Bob!
In English (not Marsh speak) what did Jon think of the Modi MB ? I read the thread but................. Duh?
Ron

Yeah, sometimes I'm not exactly the great communicator, as for me, the test data is what mainly tells the story.

To put things in perspective, I'd say that I have considerable respect for Mike Moffat's concept about closed form filters and multi-bit converter-

• The LF response profile shows it is for all practical purposes DC coupled, which means low group delay and possibliy audibly better impact on transients, especially complex low frequency transients
• Image rejection and overall CCIM distortion are very good, which means two major sources of high frequency issues are not there.
• Being an industriral mutli-bit DAC the low level precision is quite good- right down to the last bit.
• OTOH, unfortunately, that last bit isi only 16 bits. Now, WHENEVER you have 16 bit data, all you're going to be doing is processing it with 16 bits, popular myths aside about how upsamping works. (Otherwise, those -90 dBFS test signals would look much different if they were up sampled and bit interpolated to 24 bits- I've yet to see a DAC do that, including the fairly large library of Stereophile test reports)
• So, the framework is there to do a pretty good job with 16 bit data.
• But then, there is that mid scale glitch. Let's use a different metaphor, and say it's a fly in the punch bowl.
• In the end, the Modi is a 250 dollar DAC, and the only way to evaluate how it sounds is to listen to it...
• Test data can give good hints as to some possible issues, and can also uncover clear excellence. When it comes to budget level devices, it's very much the skill of the designer in making the right trade-offs to produce the best sonic results for a relatively meager outlay. (OK, $250 is a bunch of money in some parts of the world- in other parts, it's the price of a dinner out for two... depends on what worlds you travel in. I travel in a lot of different worlds... )
• A more thorough technial assessment is posible, with testing, but unless other devices are tested the same way, it might be even more confusing, so I tend to stick with the basics as published by reviewers like Stereophile.
• The driving source unit and resistance to issues with source jitter are a big factor also in sonics, and that can't really be evaluated well with an AP, other than using an optical interface. (they have a lot of jitter). And if you use optical as your usual connection medium, other than being forced to because you need to break a ground loop that you haven't fixed the right way, then shame on you! :W

[JonMarsh]

Reviewing specs, I also have to say I'm conflicted about whether to recommend the Bifrost Multi-Bit or the Delta-Sigma (AK4490). The Multi-bit version also uses the AD5547C, which is the more accurate version, and the DAC will accept up to 192/24 data, BUT, it's only processing output at 16 bits. That's what you get when you upgrade to a Yggy, a multi-bit converter that will process at 21 or 22 bits. BTW, the best Delta-Sigma 24 or 32 bit parts struggle to get 21 bit accuracy.

BTW, checking online, one can find the Cambridge 851D several places for $899. AES/EBU inputs, balanced and unbalanced outputs, a multiplying DAC volume control with very high precision and low distortion (it IS called a Digital Preamp, for good reason, though compared with a conventional preamp it's a bit more limited in signal output level as the gain control is just an attenuator, so the typical 2VRMS for unbalanced, and 4VRMS for balanced applies.

I'm not a shill for Cambridge, but I am a shill for value. One might recall Stereophile's summary:

"The Cambridge Audio Azur 851D has bang for the buck all over the place. It approached the sound of DACs costing four times"

And that quote is relative to the original MSRP of about $1600.

This is why it's hard for the little guys to compete with the cost structure of the big guys. I think the price has been dropped on it to dealers, because perhaps the concept of digital preamp just didn't catch on, and wasn't "marketed" well.

[BobEllis]

You're doing a good job selling the Cambridge, Jon. The only issue I have is a low gain F5T amp that won't reach full power on 2V, but I need to have a bpass for the HT mode meaning another box anyway. Wish I had the budget to compare it to Yggy in person.

[JonMarsh]

You're doing a good job selling the Cambridge, Jon. The only issue I have is a low gain F5T amp that won't reach full power on 2V, but I need to have a bpass for the HT mode meaning another box anyway. Wish I had the budget to compare it to Yggy in person.

One can use it as just a DAC, with the attenuators set to 0 dB attenuation. And use it with a conventional preamp. I know someone doing that, as they have analog sources, too. I can't personally think of a pure DAC at $900 I'd recommend over that, though there certainly may be competitive units out there I haven't heard about.

closest priced recently reviewed DAC I can think of is the PS Audio NuWave processor, for $1299, and I can point out numerous measurements that don't compare well to the 851D.

[Renron]

Doc,
Thank you for the well spoken English translation of the graphs.
I'm sure to you they tell a "Lord of the Rings" Trilogy in 3D Kodachrome with Google-O-phonic surround sound.
To my uneducated eyes they are just graphs. (sad) Your posts above make complete sense to me and I thank you for that.
You sir, are a gift to us everyday audio people, and are appreciated beyond mere accolades.
Simply, a BIG Thank You Jon.
Ron

[Bertram]

I'm using one of these Breeze Audio DACs with the Xmos usb upgrade and it sounds dandy. Its fed from a QNAP 453Pro using a decent USB lead. Van den Hul interconnect, Amplifier is a Arcam A85, speakers are Epos 12.2 on their dedicated stands and cables are Chord Odyssey single wired. Sound is very clear. I'd recommend one in a heartbeat. It replaced a Schiit Bifrost, not because i was unduly unhappy with it but mainly because the QNAP had ceased to recognize it after a firmware update, rolling back the firmware didn't help. The build quality of the Breeze is obviously not in the same class as the Schiit but sonically it betters it to my ears (which is what counts) I'm currently going through that phase of listening to all my old albums with renewed joy. Buying things like the Breeze is always a bit of a punt but i'd be surprised if anyone was disappointed with it. Remember, some of the best kit can measure poorly.

[wkhanna]

welcome to the forum, Bertram....!

remember, all that matters is if YOU like it in your system...:W