PI2AES Digital Audio Shield for the Raspberry PI

[odessamarin]

@amirm
Could you please also include eventually measurement of the PI2AES powered by 12V battery. Similar way I use it.
I am so curious to see anything that correlate with what I hear. And may help with skepticism around here.
Thank you.

 

[Soniclife]

Yes, I can take identical BNC cables and randomize them at the DAC input part. This way I should not know who is who.. do listening and write down preference for ch1 and ch2. then see. And repeat. Should work.

Sounds like a great way to do it, as long as your brain does not work out if there is a tell as to which is which.

 

[Michael Kelly]

Update on PI2AES final revision testing. I had four boards built up and have tested all of them. Results are excellent, though to be honest I personally can’t tell if they have improved dramatically from the previous revision. Three of these units will be sent out to reviewers and hopefully they will agree or at least tell me that the results are as good as or better than the previous rev.

Sometime this week we will be turning on pre-ordering for the unit. We first have to work out a delivery schedule just so we know what to tell people for delivery time. My guess at this point is 3 to 4 weeks max from tomorrow we will start delivering. It might be a bit sooner, but some of the parts we have to order will take a week or two to get in hand

 

[zzz]

@Michael Kelly Will an acrylic case be made that is compatible with the ports in the new Raspberry Pi 4?

 

[Michael Kelly]

Yes. It only requires a new side to accommodate the micro hdmi ports and the bigger usb type-c port. We will have this option ready when we start shipping in few weeks.

EDIT - This does not mean the PI2AES will work with the PI 4. We bought a couple for initial testing, but the player SW (Volumio, DietPi, RUne, etc) we will need may not be ready for some time. Until that happens we can't be sure it will work.

 

[Michael Kelly]

The PI2AES is open for pre-orders now! Delivery time is 1-2 week for the first batch. When these are sold out we will put another batch in, but with a 3-4 week lead time.

Cheers,
Michael

 

[odessamarin]

PI2AES vs. Allo Digione Signature based on blind test and something more…

Before I will go with results, I want to briefly describe how I get PI2AES for test.

After I get Chord DAVE in my system, I understand how nicely it benefits when good source provided. With Sennheiser hd800s directly connected to DAVE even changes in USB cable will effect sound. Therefor I was looking for good source and I liked the idea to have RPi based streamer for my audio collection hosted in local network NAS. First, I get Allo DigiOne and after a while Signature version from Allo. Later I also found many good reviews for 502DAC (from PI 2 Design), therefore I decide to get one as well. After I place my order, I found Michael’s posts regarding new PI2AES development. So, I wrote him and ask if he would be interested for test and compare his new device (PI2AES) with other boards I already have. He kindly agreed and send me PI2AES together with my 502DAC.

After small discussion and advices here, blind test approach was performed to compare Allo Signature with PI2AES. For this purpose two identical BNC coaxial cables was used to randomly reconnect them to DAVE inputs (BNC1 and BNC2). Then the same tracks was played simultaneously on two boards and evaluated blindly, selecting there and back different BNC inputs with Dave remote control. The same RPi 3+ and Volumio versions was used in each board. Digione Signature (DS) and PI2AES boards was powered separately from RPi. For Digione SIgnature a Li-ion battery pack from Allo was used as clear power source. Compact 12v lead battery was used to power PI2AES. Standard but good 5V power adapters used to power RPi in both cases.



Blind listening confirmed Allo DS and PI2AES sound different. First, I decide to go through different tracks, switching BNC inputs to find out which sound I prefer and why. Both inputs sounded very clear. The main difference I found was how echo extensions from instruments and voices was reproduced, how far and deep they propagate and reverberate. How the volume of instruments sound was represented and how well instruments separated from each other. After I clearly decide, which BNC input sounds better to me, I just unplug it from board of my choice to predict whether I right (true positive - sound disappears) or I am wrong (false positive – sound continuous to play). Then I repeat this test over and over… randomize BNC at the Dave inputs, listening both inputs, stop on the input I like, do my prediction which board is it, disconnecting BNC form this board and see if I right or wrong. And the winner of this tests is… PI2AES! Period. Blindly, after few iterations I was able to reliably identify sound of PI2AES many times in a row in the different tracks and albums. Sound of PI2AES was always better to reproduce soundstage scene depth, instruments and voice volumes and separation. PI2AES has sound that hypnotize you somehow, bringing you inside of the scene, very holographic and realistic. I think part of the credits can go to Chord Dave, let’s beck to this later. As for Allo Digione Signature, yes.. I know its measurements is very good, and it sound good indeed. It’s very clear and transparent. Black background, the soundstage is wide but for some reason it feels like echo can’t propagate in depth well, like some wall behind the stage truncate it and block this propagation. It give impression of restricted depth space. Otherwise, it sound very good. Most probably without PI2AES reference in my case I would not say it flat or not dimensional. But when I blindly compare two inputs, one of them was continuously sound better from this aspect. And it was PI2AES. More natural, more spatial, more holographic and realistic. Even more credits go to PI2AES for more output options (full size AES, Buffered I2S, optic). Which is important. BNC was used during blind test to have the same interface from two boards, but AES from PI2AES sounds even better than its BNC, jaw dropping experience as I mention in my first impression posts before.

To finalize this report I decide to include one more thing



I was wonder and ask myself whether this nice holographic 3D representation may come from digital “manipulations” within digital pathway/gears (Dave, PI2AES, Volumie..) or it could be signature of the well reproduced recording. Therefore, I decide to compare one of my favorite album “Autour de Chet” that I have in Hi-Res and in LP vinyl. For this Digital vs Analog test, I connected LP Turntable and Dave DAC (PI2AES and Allo Digione SIgnature) to the Brystone BHA-1 amplifier. The album play was simultaneously started on LP and both boards, the sound was evaluated by switching amp inputs (analog vs digital), and DAC inputs (PI2AES vs Allo DS). It was very interesting listening session. To make long story short... PI2AES connected to Dave was really close to the analog reproduction in a way of sound stage image representation, depth extension and instruments layering. Of course, PI2AES was brighter and less low frequency extended sound in compare to the LP analog reproduction (very warm, deep, dark and flued). But they were very close to reproduce sound volume and space, I actually was surprised how similar they are. For me, it’s just confirm that correct special representation can indicate more precise digital source and its conversion to analog.I want thanks’ again Michael for development of this product for us.

RPi DDC Borads: Allo Digione Signature, PI2AES (v.P3)

DAC: Chord Dave

Headphones: Sennheiser hd800s, Grado RS1 (pink driver) HP1000, PS1000.

 

[Music1969]

Hi @odessamarin , did you try the TOSlink output of PI2AES? With all other DAC inputs plugged...

 

[kaka89]

If my DAC already supports USB, how much difference does such device makes?

 

[johan]

e to compare one of my favorite album “Autour de Chet” that I have in Hi-Res and in LP vinyl. For this Digital vs Analog test, I connected LP Turntable and Dave DAC (PI2AES and Allo Digione SIgnature) to the Brystone BHA-1 amplifier. The album play was simultaneously started on LP and both boards, the sound was evaluated by switching amp inputs (analog vs digital), and DAC inputs (PI2AES vs Allo DS). It was very interesting listening session. To make long story short... PI2AES connected to Dave was really close to the analog reproduction in a way of sound stage image representation, depth extension and instruments layering. Of course, PI2AES was brighter and less low frequency extended sound in compare to the LP analog reproduction (very warm, deep, dark and flued). But they were very close to reproduce sound volume and space, I actually was surprised how similar they are. For me, it’s just confirm that correct special representation can indicate more precise digital source and its conversion to analog.I want thanks’ again Michael for development of this product for us.

Thx you Odessa for your subjective review of the Digione Signature vs pi2Aes .

As you know this is Audio Science Review and measurements trump personal and subjective reviews .How a DAC will sound with a specific transport has a lot to do with jitter attenuation on your DAC...and personal preference .

I have no doubt that you prefer Pi2Aes , aparently even revisions that were corrected , (and it seems the latest revision its well engineered , congrats @Michael Kelly ) but stuff like "sound stage image representation, depth extension and instruments layering" has no place on ASR
One of your findings was that using a lead battery on PI2AES opened the stage and made the AES better sounding that Digione Sig . Do you still prefer the lead battery sound ?

I do not want to demean your opinion ,but understand that opinionated people always believe they know better. This used to be old audiophile way . Its changing now thx to people like Amir .


As of right now , Digione Signature has the best jitter measurements of all transports and thats not subjective.



Enjoy your music and enjoy Audio Science Review .

 

[Michael Kelly]

Thx you Odessa for your subjective review of the Digione Signature vs pi2Aes .

As you know this is Audio Science Review and measurements trump personal and subjective reviews .How a DAC will sound with a specific transport has a lot to do with jitter attenuation on your DAC...and personal preference .

I have no doubt that you prefer Pi2Aes , aparently even revisions that were corrected , (and it seems the latest revision its well engineered , congrats @Michael Kelly ) but stuff like "sound stage image representation, depth extension and instruments layering" has no place on ASR
One of your findings was that using a lead battery on PI2AES opened the stage and made the AES better sounding that Digione Sig . Do you still prefer the lead battery sound ?

I do not want to demean your opinion ,but understand that opinionated people always believe they know better. This used to be old audiophile way . Its changing now thx to people like Amir .


As of right now , Digione Signature has the best jitter measurements of all transports and thats not subjective.



Enjoy your music and enjoy Audio Science Review .

Full disclosure - I am the designer and manufacturer of the PI2AES and 502DAC boards that Odessamarin reviewed.

Johan - I see you are a manufacturer. May I ask what it is you make? As a manufacturer myself, and as an engineer, I am a big believer in measurements. But I have always held the position that we can not measure everything that affects sound to our human ears. Is it not reasonable to use the blind testing to determine that this one set of human ears prefers one sound chain over another? I agree that ASR is a valuable place for measurement based discussions. But just as I do not like when other, more subjective oriented sites seem to dismiss measurements (or fail to take them at all), I believe we should not dismiss nor lump odessamarin's results with those who you feel are just "opinionated people". He seems to have done what others here on ASR have asked for: blind testing over multiple times and conditions. Do we not want this here? Amir certainly includes subjective listening tests to most of his reviews. And as far as I can tell they are not blind tests, simply his observations with each unit individually.

Cheers,
Michael

P.S. I am sending Amir a system with the PI2AES and a PI3 for his review. Obviously he has to put it into his ever growing queue of devices to test, so no idea when that will happen, if at all.

 

[Nango]

@amirm knows, first things first!!

 

[johan]

Hi Michael

I am the designer of Digione Sig , Digione from allo

I have read Odessa reviews . My favorite is his opinion on how the lead battery (vs li ion battery )transformed the sound ..profoundly . As en engineer I know thats not possible .

Now , in my experience , audiophile world is full of opinionated people like Odessa (and since he thinks that lead battery sounds better...) , but things are changing .

Now you say that we should take Odessa blind testing into measurements . I am sorry , but I don't think thats possible . A blind test is a serious test and to be scientifically valid certain rules must to be fallowed . Are you positive Odessa fallowed those rules ?

If he didn't , its just an opinion disguised as a real test better left for other forums .

Amir includes subjective listening tests yes , after he takes the test data and makes it public. Note the difference between Odessa and Amir.

 

[Veri]

And the winner of this tests is… PI2AES! Period. Blindly, after few iterations I was able to reliably identify sound of PI2AES many times in a row in the different tracks and albums. Sound of PI2AES was always better to reproduce soundstage scene depth, instruments and voice volumes and separation. PI2AES has sound that hypnotize you somehow, bringing you inside of the scene, very holographic and realistic. I think part of the credits can go to Chord Dave, let’s beck to this later. As for Allo Digione Signature, yes.. I know its measurements is very good, and it sound good indeed. It’s very clear and transparent. Black background, the soundstage is wide but for some reason it feels like echo can’t propagate in depth well, like some wall behind the stage truncate it and block this propagation. It give impression of restricted depth space. Otherwise, it sound very good. Most probably without PI2AES reference in my case I would not say it flat or not dimensional. But when I blindly compare two inputs, one of them was continuously sound better from this aspect. And it was PI2AES. More natural, more spatial, more holographic and realistic. Even more credits go to PI2AES for more output options (full size AES, Buffered I2S, optic). Which is important. BNC was used during blind test to have the same interface from two boards, but AES from PI2AES sounds even better than its BNC, jaw dropping experience as I mention in my first impression posts before.

My problems with his impression in bold. There isn't really any abx log here, just "one was continually better". I mean this kind of test is difficult either way, because even if a detailed log is produced one can still question its validity, but at least then we have some numbers. Now it's just one big anecdote. After "a few iterations" the difference was obvious, really..?

I agre with johan that "natural, more spatial, more holographic and realistic" is a little much for two proper and highly competent interfaces. One would almost think one of the two is very defect if there really would be so many audible difference, from a digital interface none the less!

 

[Michael Kelly]

My problems with his impression in bold. There isn't really any abx log here, just "one was continually better". I mean this kind of test is difficult either way, because even if a detailed log is produced one can still question its validity, but at least then we have some numbers. Now it's just one big anecdote. After "a few iterations" the difference was obvious, really..?

I agre with johan that "natural, more spatial, more holographic and realistic" is a little much for two proper and highly competent interfaces. One would almost think one of the two is very defect if there really would be so many audible difference, from a digital interface none the less!

I understand that much of the language he uses is not too your liking. But, he does say clearly:

"when I blindly compare two inputs, one of them was continuously sound better from this aspect " (bold yours, underscore mine)

Help me understand how he has not met your definition of blind A/B testing. What else can he do that is not measurement based? Is it simply that any non-measurement based review is invalid for you? Which I think is fine, but maybe needs to be made clear. I don't think that is Amir's position, but maybe I am wrong.

Cheers,
Michael

P.S. You may not believe it, but I would have been just as happy if he found your devices to be superior. We designed the PI2AES to be competent and flexible with the multiple output formats. Performance was important, but we were not going for that at all costs.

 

[Music1969]

My favorite is his opinion on how the lead battery (vs li ion battery )transformed the sound ..profoundly . As en engineer I know thats not possible .

Can't car batteries provide very very low impedance and very very high current, power source?

 

[Veri]

I understand that much of the language he uses is not too your liking. But, he does say clearly:

"when I blindly compare two inputs, one of them was continuously sound better from this aspect " (bold yours, underscore mine)

Help me understand how he has not met your definition of blind A/B testing. What else can he do that is not measurement based? Is it simply that any non-measurement based review is invalid for you? Which I think is fine, but maybe needs to be made clear. I don't think that is Amir's position, but maybe I am wrong.

Cheers,
Michael

P.S. You may not believe it, but I would have been just as happy if he found your devices to be superior. We designed the PI2AES to be competent and flexible with the multiple output formats. Performance was important, but we were not going for that at all costs.

-a proper log with some more information on procedure rather than going by words like 'obvious' 'immediate' 'continuous'
-less hyperbole and audiophile words to describe seemingly big differences where there should be none, this is a DDC..

I don't mean to disrespect or discredit, I just don't think that kind of 'review' is consistent with the whole ASR way of thinking. Sure it is nice to read such positive words about your product, but I and others will question certain things that are stated especially in this way. I think that can be expected. Maybe other members can give more pointers.

 

[Michael Kelly]

-a proper log with some more information on procedure rather than going by words like 'obvious' 'immediate' 'continuous'
-less hyperbole and audiophile words to describe seemingly big differences where there should be none, this is a DDC..

I don't mean to disrespect or discredit, I just don't think that kind of 'review' is consistent with the whole ASR way of thinking. Sure it is nice to read such positive words about your product, but I and others will question certain things that are stated especially in this way. I think that can be expected. Maybe other members can give more pointers.

I can say that based on my own measurements using an AP SYS-2522 (posted earlier in this thread) the Allo DS has a lower jitter noise floor. It reaches a jitter noise floor of around 0.5psec out to 20K, while the PI2AES gets to 0.6psec out to 12K, but then rises to around 0.8 psec after that. So by our own measurement, the Allo DS is superior and that's OK with me! It is a great product and I doubt anyone who has one would be unhappy.

Cheers,
Michael

 

[johan]

Hi Michael

I am trying to arrange for a double blind test for one of my products and from my understanding of what a blind test is , Odessa post... is anything but a blind test .

Further , Odessa hears a difference in battery chemistry .

Thats why he likes PI2AES better , he has bionic hearing .

 

[johan]

Can't car batteries provide very very low impedance and very very high current, power source?

Yeap thats possible on a AMP that needs low impedance high current source... here we are talking about devices that take 50ma and have lots of decoupling caps . Furthermore the output is digital , a simple hi and low. Battery impedance has no influnce on digital output