Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter

According to whom? And what makes someone a "boutique audio shill"?
One who shills for boutique audio. The following quote is an example of things they post.
"That is why I reject the claim that $200 or $500 dacs that measure pretty well sound as good as any dac can ever sound, they don't."

Does that mean they are a boutique shill who deals in audio?
The following quotes are examples.
"there are ties, as audio development was and is part of my work."

"By the way: I am in the audio business and I must be biased because of this."

Please explain your understanding of neurology and perception because it isn't matching mine. What faculty of the psyche verifies "what is actually heard" in your model? How does it differentiate from an imagined sound?
My explanation was on the language. One of the ways to see if your perception is only in your head or actual sound wave variation is to measure it at your ear location.
Your argument is no stronger than it was first time I read it. You claim to speak for everyone and I find that, well, a bit arrogant.
You are entitled to your own opinion.
 
My explanation was on the language. One of the ways to see if your perception is only in your head or actual sound wave variation is to measure it at your ear location.

Ah, ok.

I'm going to go out on a very long limb here and guess you don't have a lot of experience in setting up microphones to record music. I do have some experience in that area.

Measuring Instrumentophiles in general: I appreciate what you are trying to do to rid the windswept, flowing countryside of cornfields of those scoundrel liars who pretend their equipment sounds different and better in ways that can't be measured so they can empty the wallets of unsuspecting rubes.

However...

1. I don't believe they are all lying. Some of them, perhaps.

2. My multiple DAC comparisons were carefully level-matched blind A-B testing with my wife switching between using a Schiit Sys. She didn't know which was which either. The results were 100% consistent and, frankly, it was quite easy to pick out the winner. 1 unit sounded very substantially better than the other, better-measuring units and the gulf was wide. My band's bass player came to the same conclusions. You don't have to believe me, of course. But don't discourage others from trying it out for themselves with similar testing that should rule out any conscious or subconscious prejudices.

Are you sure enough about your conclusions to be flexible enough to let in the possibility that sometimes the worse-measuring equipment sounds better?
 
However...

1. I don't believe they are all lying. Some of them, perhaps.
You are free to believe what you want. My assessments on them are based on years of observations.

2. My multiple DAC comparisons were carefully level-matched blind A-B testing with my wife switching between using a Schiit Sys. She didn't know which was which either. The results were 100% consistent and, frankly, it was quite easy to pick out the winner. 1 unit sounded very substantially better than the other, better-measuring units and the gulf was wide. My band's bass player came to the same conclusions. You don't have to believe me, of course. But don't discourage others from trying it out for themselves with similar testing that should rule out any conscious or subconscious prejudices.
That's interesting. What process did you use for matching levels and to what db?
Are you sure enough about your conclusions to be flexible enough to let in the possibility that sometimes the worse-measuring equipment sounds better?
What constitutes better sound in audio devices designed for replaying?
 
I think Andrea is a man on a mission, its a project to benefit everyone that is for him a greater good than just doing for himself. In his enthusiasm he has tried to educate others. Its frustrating when one is so engaged to understand many people don't want to be educated, they want to keep believing whatever they already believe. They have already found comfort in their own religion. For Andrea its a fight every step of the way, new knowledge is not allowed to come incrementally as it has throughout history. Now the only acceptable way to move forward is deemed by some to be with published academic studies before all else, either you have the resources to do it, or you have to shut up. Why? Because people will always find some way to spend money foolishly, and self-appointed protectors want to censor what your delicate ears might hear about small steps forward in audio reproduction that occur without meeting arbitrary publications standards set up by a few protectors. I promise you, ESS, ADI, TI, almost nobody will publish everything they know, and they will use what they know to try to make money off you. Are they all evil then? If you want to be protected from people making money off you using unpublished research then worry about those guys. Enthusiastic hobbyists are the least of your worries. An oh, by the way, Ian is an enthusiastic hobbyist who has turned his hobby into a way to make some extra money at your expense. Do you want him censored too?

The only difference between those other guys and Andrea and me is that we are not in it to make money at this time. That is true despite the prosecutorial cases the conspiracy theorists cook up for social media consumption in internet forums. If and when that ever changes for me, I will announce it here in this forum so everyone can take it into consideration as they see fit.
 
Last edited:
If one need zero latency he can design himself the device.
It is not forbidden.

This is the wrong thread, this device has latency since it's a FIFO buffer.

Of course, I'm one of those who has gotten a lot of enjoyment from it since I have bought about ten FIFO buffer from Ian (and a lot of other devices).

But if i needed zero latency I would design the device myself, of course with my very long developing time, that's not a problem for me.

And of course we design audio devices following our own approach, not the approach of someone who are looking for a zero latency device on a FIFO buffer thread.

Then we share our designs since it's not forbidden.
And of course we care less than zero about those who are looking for zero latency devices since we do not follow market expectations.
We build this devices for ourselves.

IMHO, as I said several times in this thread, Ian's devices are good devices, and for how much they cost they are a bargain.

Are they useful?
I believe so, although I cannot proof anything with the J-Test.
Of course, I care less than zero about the J-Test performed by the Lord of the Blue Screen.

If the required resource is the APx555 I prefer to save the money and shut up.
 
Last edited:
Hi,

coming from the DDDAC thread, I need some help please regarding FIFO Pi Q3.

I have it in my system for quite some time now. I'm very pleased soundwise when a track is running, no pop / clicks, very clean and detailed sound, great soundstage.

But there is one issue I'm struggling with: When I start playing a new track or jumping to a new position within the same song, there is often one short (loud) click / pop. After that the track plays without issues with great sound. The click volume differs, sometimes it is quiet I can almost not hear it, but most of the times it is really loud and I'm afraid that my speaker get broken!?! Also the 'position' of the click moves: Most of the time on the left side, sometimes the click moves to the middle and right side of the soundstage. Really strange.

I have done some tests but to summarize:
RPI I2S -> FIFO PI Q3 -> DDDAC: Clicks and Pops at track start
RPI I2S -> DDDAC: no issues

I'm running volumio, but I have also tested with moode. Same issue.

If have done some tests with different settings without success:

I've set the I2S DAC Model to HifiBerry and test wise to Generic I2S.
I have switched on audio respampling to force 24 Bit output. And I've disabled resampling and used the fifo pi 32 bit transformation.
If have exchaned the Accusilicon clocks against the default clocks for one test.

Because I thouht this comes from EMI issues I have ordered a GPIO extension cable to get some space between RPI and FIFO PI. But also not luck with this.

In the DDDAC thread there are a few other people with the same effect, but they have not done tests to isolate the fifo pi as the cause of the issue, so maybe I'm still the only person with this issue!? Just doing something wrong, or my copy is broken!?

Do you have an idea? Someone with the same issues? I would really love to keep the FIFO PI in the system because sound is great when track is running, but the clicks in the beginning are driving me crazy...
Don't know if it is sample rate related. This weekend I can do some test, also of the click is not that within the same album.

I have Dynaudio Esotec tweeters in my system. Maybe a bit less sensitive. Could be[emoji4]

Let's see if the click is gone with another fifopi. Is it Q2 or Q3 model?

Ad.

@Ad: I have a Q3 model.

I asked my dealer for exchange and got now a reponse. He offers me to check the fifo pi in his workshop. He will charge me 30€ if the device is not defective. This means for me in worst case, that my fifo pi will be away for some weeks plus the risk with shipping to another country and at the end I need to pay for it because my dealer tells me everything is fine in his workshop. Great :(

Actually I'm not sure if should send it back. Means, if my fifo pi is really defective or if it is a design issue of the fifi pi or some kind of incompatibility with DDDAC setup. I guess not, because there is not much negative feedback in this thread. But there are some people with similar issue who learned to live with it and maybe there are more out there I do not know!?

I would appreciate any advice? :confused:
 
@michl2604
Probably it would take some test equipment to figure out what's going on between FIFO_Pi and dddac. If nobody using FIFO_Pi and dddac has the equipment then the problem may not get solved. If FIFO_Pi works fine with other synchronous mode dacs then it could be a timing or glitch problem with an I2S signal dddac uses for clocking the output, say possibly, LRCK, instead of the MCLK that some other dacs use. Something like that. If the problem is from inside FIFO_Pi then it might still be possible to fix with a modification to FIFO_Pi. Depends on exactly what was found. OTOH, if the problem were found to be with the interface between FIFO_Pi output and dddac input, say, perhaps, reflections on one or more of the I2S lines causing occasional timing violations going into dddac, then that might be possible to fix with a modification to dddac, or perhaps with some custom I2S interface circuity between FIFO_Pi and dddac.

All the above having been said, when people report pops at the beginning of music tracks or when moving between track positions, often it turns out to be a problem with the player software and or OS causing a glitch. Another possibility might be a problem in the interface between the I2S source output and FIFO_Pi I2S input. For example, when an USB to I2S board is used as the I2S source it may assert a mute signal on one of its non-I2S pins that could clue in that a dac should be muted. Since FIFO_Pi does not read and react to non-I2S signals there is some chance that a glitch could be passed through while a mute signal was asserted. Some of possible causes related to software can be checked by changing player software, OS, and or hardware I2S sources feeding FIFO_Pi. Substituting other dacs instead of dddac might also provide clues.

Hopefully, its starting to become clear why nobody has given you much advice since the symptoms you describe can have many possible causes, some of which can be hard to find.
 
Last edited:
If one need zero latency he can design himself the device.
It is not forbidden.

No you cannot buffer something synchronously and have zero latency.

This is the wrong thread, this device has latency since it's a FIFO buffer.
wrong thread for what? as I said, i'm doing DSP, there is no way to have zero latency unless all audio is processed with the FIR, saved to disk and replayed, without any kind of buffer. I dont need zero latency, that would be stupid, as DSP also brings latency and wasnt what I searched for, since it does not exist.

It could be useable, if delay was lower and known at all times, then that value passed to the video player software, allowing video to be delayed the same amount, but this process is too convoluted and would only be doable with some media. It is especially problematic for things like streaming and netflix etc, which have only become more prevalent and important in my entertainment diet after covid19, so I will do without a buffer and simply drive the source with a clock that is synchronous to the dac. Delay with the fifo could be minimised by upsampling to the max sample rate; filling the buffer quicker, but cannot be eliminated.

Of course, I'm one of those who has gotten a lot of enjoyment from it since I have bought about ten FIFO buffer from Ian (and a lot of other devices).
oh yeah, you bought them for your own enjoyment .... riiiiight.

But if i needed zero latency I would design the device myself, of course with my very long developing time, that's not a problem for me.
indeed, I am, but as I mentioned, I dont need zero latency, but I also decided I dont really need a buffer. it adds unwanted complexity to an already complex system design.


And of course we design audio devices following our own approach, not the approach of someone who are looking for a zero latency device on a FIFO buffer thread.
once more for those in the back ... or hard of hearing ...

Then we share our designs since it's not forbidden.
And of course we care less than zero about those who are looking for zero latency devices since we do not follow market expectations.
We build this devices for ourselves.
bahahahahahahahahaha. again... I dont need ... oh nevermind :p


IMHO, as I said several times in this thread, Ian's devices are good devices, and for how much they cost they are a bargain.
you really expect me to believe that? i've read this thread. weve all read your posts on his products and your tone, masquerading as a search for knowledge.

Are they useful?
I believe so, although I cannot proof anything with the J-Test.
Of course, I care less than zero about the J-Test performed by the Lord of the Blue Screen.
agreed, J-test is fairly useless as a single metric, which is often the way its used. its started to change with THD at 1khz vs multitone testing; I suppose it will start trickling down somewhat with jitter testing too. the usefulness of each metric (THD+N @ 1kHz and J-test) have a lot of parallels, as do the more useful replacements: phase noise vs Frequency and multitone distortion/distortion vs frequency

If the required resource is the APx555 I prefer to save the money and shut up.
not likely.

Do go on with your obvious red herrings. you really doubled down on trying to paint me as silly for wanting a fifo buffer with no latency, when I did not mention such a requirement. So much so, that at least half of your post and all of the points you tried to score on me missed the mark... oh well. better luck next time.

I have covered this before
 
Last edited:
It looks like when I say to design a device I mean a FIFO buffer.
Wrong, If I meant FIFO buffer I would write "FIFO buffer" and not "device".

There aren't too complex design, it only depends on the ability of the designer.

Music is a bit more complex than a multitone.
Of course, everyone is free to buy a 25000 USD tool in order to measure as in the 70s, I save the money.
 
@Ad: I have a Q3 model.

I asked my dealer for exchange and got now a reponse. He offers me to check the fifo pi in his workshop. He will charge me 30€ if the device is not defective. This means for me in worst case, that my fifo pi will be away for some weeks plus the risk with shipping to another country and at the end I need to pay for it because my dealer tells me everything is fine in his workshop. Great :(

Actually I'm not sure if should send it back. Means, if my fifo pi is really defective or if it is a design issue of the fifi pi or some kind of incompatibility with DDDAC setup. I guess not, because there is not much negative feedback in this thread. But there are some people with similar issue who learned to live with it and maybe there are more out there I do not know!?

I would appreciate any advice? :confused:

@michl2604

Different from other DACs, DDDAC has it's own logic to convert standard I2S format into left justified NOS mono block timings. Some times multiple DDDACs can be in parallel configuration to increase the output current and lower the noise floor.

Can you please post pictures of your system and configuration to see if I can figure out what's the issue?

Thanks,
Ian
 
2x AS338 received

Just received the 2 AS338 clocks from the Taobao Focus order

I'm installing in the LifePo4 + Ultracaps powered FifoPi .... hopefully I'll notice a positive difference from the AS318b and NDK clocks they are replacing.

Please forgive the rats-nest wiring- this unit is a work in progress.

Listening reports coming soon. It would be cool if the FifoPi allowed manually switching between 2 clocks from the same family for better A-B testing... just a thought.
 

Attachments

  • AS338-3-23-21.jpg
    AS338-3-23-21.jpg
    559 KB · Views: 360
Disabled Account
Joined 2020
^^

Thanks @JCMcNeil for your efforts to clock-roll no matter the outcome. At least you received those beauty's safely (Most important). Looking forward to your subjective conclusions.

I'm distracted on sources ATM, otherwise I would be hanging by a thread (pun intended) with news of new clocks to roll.

OAN, since I'm making sure my 12V source has eMMC, has anyone tried eMMC on the RPi to lower noise (clicks & pops)? It's not soldered in like on a 12V motherboard, but interesting to experiment.

RasPiKey: Plug and Play eMMC Module for Raspberry Pi | UUGear

On my sources, I want eMMC, the ability to load the whole OS + RAM buffer exclusively in RAM and network drive going forward. No USB hard drives or even internal SATA, etc. if I can avoid, just eMMC internally.

EDIT: I'm not an Allo fan, but Allo also offers this option:

eMMC
 
Last edited:
Does Ian's FifoPi keep both clocks powered in this manner? Or, if not, can it be modified to do so?

You can keep the DC power to the FifoPi clean side always on without shut down, so both of XOs will be powered continuously.

FifoPi Q3 also supports both XOs running at same time. In my system, I use two Pulsar OCXOs. They don't have any OE control. So both of them keep outputting even at time that is not selected as current MCLK. You can see my YouTube video for more information.

The Best Sound Quality Audiophile DIY DAC I Built in 2020 - YouTube

For XOs with the OE control, you can still keep them outputting all the time by disconnecting the OE pin from the XO sockets (they have internal pull-up).

Ian
 
@michl2604

Different from other DACs, DDDAC has it's own logic to convert standard I2S format into left justified NOS mono block timings. Some times multiple DDDACs can be in parallel configuration to increase the output current and lower the noise floor.

Can you please post pictures of your system and configuration to see if I can figure out what's the issue?

Thanks,
Ian

Thank you very much Ian for taking care of this case :)

Here are two pictures of my system. Please excuse the messy setup. My plan is to put the whole thing into a case once I'm finished / satified with it (That can't be long because the click thing is my only left issue).

If I can give more information or pictures please let me know.

Thank you in advance!!
 

Attachments

  • 2021_0324_17575300.jpg
    2021_0324_17575300.jpg
    727.2 KB · Views: 271
  • 2021_0324_17575700.jpg
    2021_0324_17575700.jpg
    680.5 KB · Views: 265