I think I agree with your diagnosis on variables to observe...but trust me, the system I am listening to is beyond 98% of the stuff you could listen to on the Highend-fare in Munich where I was at in May...independantly of price point (most of the chains where clearly above xxx.xxx$).
So, I strongly disagree with you on "you will never hear what's really on a recording as it should sound, never know how good a proper system can sound." And I guees you did not mean it unfriendly...but actually before stating such stuff...come to my house, listen...and than know what the standard is we talk about before coming up with statements like that.
I dont want to go into this discussion, but I think this is the type of mindset where many experienced guys stopped at some point to share their findings as it is not fun anymore.
So, I strongly disagree with you on "you will never hear what's really on a recording as it should sound, never know how good a proper system can sound." And I guees you did not mean it unfriendly...but actually before stating such stuff...come to my house, listen...and than know what the standard is we talk about before coming up with statements like that.
I dont want to go into this discussion, but I think this is the type of mindset where many experienced guys stopped at some point to share their findings as it is not fun anymore.
Well, it was intended to be a strong statement. What concerns me is not what you are doing for fun, but how other people could be influenced to believe playing around with stuff that is obviously broken, rather than fixing it, would be a good model for them to follow.
Very interesting tests and results, so many unexpected influences...
I'm usually sick when I cross those posts, "c'mon, I haven't finished the building yet, and there's so many new questions... let's disconnect the internet... okay, keep reading"
Looking forward to seeing how your dust settles, Blitz.
I'm usually sick when I cross those posts, "c'mon, I haven't finished the building yet, and there's so many new questions... let's disconnect the internet... okay, keep reading"
Looking forward to seeing how your dust settles, Blitz.
And this is the next strong statement which is actually only reflecting your own believe...which seems to be strong like you have a religious mission to educate others.
we both have a different approach to achieve the same thing. Not sure why you want other people to educate what they want to try or what they want to follow or not.
Actually a hobby is about having fun. And i had fun to try this crazy SLC card thing on a crazy sdtrans384 player...and came to the same conclusion than my fellow diyaudio colleague...
These statement and discussions with you are no fun at all for me and i will not continue this communication...as it is not about actualy real observations but about your theories and believes...bye
we both have a different approach to achieve the same thing. Not sure why you want other people to educate what they want to try or what they want to follow or not.
Actually a hobby is about having fun. And i had fun to try this crazy SLC card thing on a crazy sdtrans384 player...and came to the same conclusion than my fellow diyaudio colleague...
These statement and discussions with you are no fun at all for me and i will not continue this communication...as it is not about actualy real observations but about your theories and believes...bye
Seems like Andrea designed the clocks, FIFO, and DAC_Lite with the goal that they should not be influenced by things like power supply noise from a computer. Then some of what we read in threads is to the effect that the FIFO isn't fully isolating because some input devices sound different from one another. A problem with the FIFO design? Maybe not. Maybe a problem with the system integration design? Again, I don't think it makes sense to buy carefully designed equipment intended to be as accurate as possible, then to defeat its intended purpose by by introducing computer power supply noise to tweak the sound. That's just common sense, nothing to do with religion.
Here is my return,
I owned a Odroid C2 during 2 years,Volumio 3 is no longer developed for the odroid and I want to use the next FifoPi Q7
So I replaced Odroid C2/Volumio 2 by one PI 4/Volumio 3,
for the moment, the PI4 is installed with a Isolator Pi (hat I already use with the Odroid C2)The latter is connected in I2S with UFL cables of 10 centimeters,
For me the passage to Raspi4/Volumio 3 to was very beneficial,I won in quality Especially with regard to details in the medium,
I don't know what made the difference, the Raspi or Volumio 3,But the result is there,I think Volumio is desperately developed for the Raspi.
Blitz I would have a piece of advice to give you,drop the USB and connect your SBC in I2S,But you need an I2S insulator,Maybe we will have identical results
Anyone who would like to talk about how to make USB audibly equivalent to SBC or any other source should feel welcome to discuss further by way of PM. No particular need to interject unwanted input into discussion of other approaches. Seriously and hopefully, everyone will be able to find here in the forum whatever approach to audio seems most sensible to them. At least maybe that much will be agreeable.
Again, no personal disrespect intended even if certain technical details seem to find some amount of disagreement. The only concern from me is for newbies who may be influenced to believe one thing or another without opportunity of exposure to differing and reasonably-well-informed views.
Again, no personal disrespect intended even if certain technical details seem to find some amount of disagreement. The only concern from me is for newbies who may be influenced to believe one thing or another without opportunity of exposure to differing and reasonably-well-informed views.
Well, I stayed with 5V to get the higher output voltage (for free, no addiotinal stage etc)....but I am finializing the battery supply which runs the 3.3V and will report back how it sounds (besides lower output voltage).
As this DAC has such an ultra resolution and is so neutral like a piece of wire...I decided to work on the source is more important, I literally can listen to the sound difference of the latest Linux-Kernels ...no joke...they have change quiet a lot on the USB-Sound-interface in Linux...with lower latency as a target...which is not what we need for PCM-Playback...you pay a price when you lower latency typically with higher overhead and much more interrupts.
As this DAC has such an ultra resolution and is so neutral like a piece of wire...I decided to work on the source is more important, I literally can listen to the sound difference of the latest Linux-Kernels ...no joke...they have change quiet a lot on the USB-Sound-interface in Linux...with lower latency as a target...which is not what we need for PCM-Playback...you pay a price when you lower latency typically with higher overhead and much more interrupts.
Blitz,
your opinion please...
the FiFo from Andrea should make quality/jitter of all sources unimportant as long as I2S is coming out and able to run through the FiFo. After that super clean clocked I2S should go into your DAC.
So how come you hear differences? Is it because the kernel is changing data?
Well, I think this theory is only theory and simply not the case in reality.
I would rather see Andreas and Ian Fifo as kind of "sound improvement" mechanismns...I believe that Fifos can only reconstruct to a certain level. I dont want to say garbage in - garbage out...as the Fifo is doing a lot. But directionally it is correct that the better the input-signal, the better the output-signal of the fifo. When "data" (beyond bit-perfect..) is lost in translation (jitter and Operation system noise in the source), its lost...how can the Fifo know that. And we talk here about OPerating system internals like timer frequency, FIFO-logistics, determanistic vs. throughput rules of RT-Kernels and chip-design and bus design of the source, clocksource of the architecture (ARM vs. x86), let alone physical layouts of motherboards, PSUs etc.
In many cases this level of reconstrunction of a FIFO is good enogh for an average system as the rest of the chain needs to be good enough to show differences. In many cases a Blusound NT130 is a perfect solution for anyone and we all should stop right there.
So, when people say, they cant hear the difference...I believe them. Their system are not capable enough than to show a difference. With their system, I wold not hear a difference either I guess.
But if you have an ultra-resolution chain, you will hear the difference. A bluesound nt130 with spdif output sounds different than a Raspberry with USB and a N5100-x86 machine sounds different from a AMD-8core Ryzenmachine similar to the Pink Faun 2.16. I got all of them here, all work on the same network, the same NAS. The DIY-sources have all the same Linux-Kernels with the same custom configurations in the Kernel-Build and in the run-setup (core/IRQ-isolation, PLaying from RAM etc).
Its the difference between very good and great.
and having a great source infront of your FIFO/DAC is not my idea or invention...look into what Ian is offering tons of tweaks for the PSU of Raspberry...or PinkFaun...or Sotm...any high-end company selling xxxxx$ Streamers, xxxx$ USB-cards etc. there are dozens of them...I hate their prohibitive price tags only.
Andrea's Fifo and DAC made something else which has been eye-opening for me: They have a transparency which makes these effects ...welll...transparent. Like a piece of wire or a clear glass of water you can look through. Before the DAC it self had his own way of masking things, so the quality of the source (sofar an Alix 1D board with Mpdpup and Sotm-Usb.card, 32 bit) was good enough as the next stage puts its own masquerade ontop of anything anyhow. There is the Sabre-Sound, the 1704, the Soekris, the AKM...all sound great...but all have their signature. Not Andreas DAC.
Its comparable to the effect going from a EL34-PP-UL amplifier to a really well made DHT-SE with nanocrystaline IT/OPT in fixed bias with a Poly-Prop (CDE947D)-PSU. You dont know the masquerade of IDHT/PP/UL and elecrtrolytics you listenined to until you take the mask off...and than you find the next masquerade going on elsewhere in your system....but in the end you have a realismn which is absolutely stunnig.
I would rather see Andreas and Ian Fifo as kind of "sound improvement" mechanismns...I believe that Fifos can only reconstruct to a certain level. I dont want to say garbage in - garbage out...as the Fifo is doing a lot. But directionally it is correct that the better the input-signal, the better the output-signal of the fifo. When "data" (beyond bit-perfect..) is lost in translation (jitter and Operation system noise in the source), its lost...how can the Fifo know that. And we talk here about OPerating system internals like timer frequency, FIFO-logistics, determanistic vs. throughput rules of RT-Kernels and chip-design and bus design of the source, clocksource of the architecture (ARM vs. x86), let alone physical layouts of motherboards, PSUs etc.
In many cases this level of reconstrunction of a FIFO is good enogh for an average system as the rest of the chain needs to be good enough to show differences. In many cases a Blusound NT130 is a perfect solution for anyone and we all should stop right there.
So, when people say, they cant hear the difference...I believe them. Their system are not capable enough than to show a difference. With their system, I wold not hear a difference either I guess.
But if you have an ultra-resolution chain, you will hear the difference. A bluesound nt130 with spdif output sounds different than a Raspberry with USB and a N5100-x86 machine sounds different from a AMD-8core Ryzenmachine similar to the Pink Faun 2.16. I got all of them here, all work on the same network, the same NAS. The DIY-sources have all the same Linux-Kernels with the same custom configurations in the Kernel-Build and in the run-setup (core/IRQ-isolation, PLaying from RAM etc).
Its the difference between very good and great.
and having a great source infront of your FIFO/DAC is not my idea or invention...look into what Ian is offering tons of tweaks for the PSU of Raspberry...or PinkFaun...or Sotm...any high-end company selling xxxxx$ Streamers, xxxx$ USB-cards etc. there are dozens of them...I hate their prohibitive price tags only.
Andrea's Fifo and DAC made something else which has been eye-opening for me: They have a transparency which makes these effects ...welll...transparent. Like a piece of wire or a clear glass of water you can look through. Before the DAC it self had his own way of masking things, so the quality of the source (sofar an Alix 1D board with Mpdpup and Sotm-Usb.card, 32 bit) was good enough as the next stage puts its own masquerade ontop of anything anyhow. There is the Sabre-Sound, the 1704, the Soekris, the AKM...all sound great...but all have their signature. Not Andreas DAC.
Its comparable to the effect going from a EL34-PP-UL amplifier to a really well made DHT-SE with nanocrystaline IT/OPT in fixed bias with a Poly-Prop (CDE947D)-PSU. You dont know the masquerade of IDHT/PP/UL and elecrtrolytics you listenined to until you take the mask off...and than you find the next masquerade going on elsewhere in your system....but in the end you have a realismn which is absolutely stunnig.
Last edited:
I dont know why this is the case. But this is like trying to explain the difference in sound of Caps. It ends up in exchanging believes and theories which is a waste of time. If you are happy with your source - stay with it. Like stay with your Wima MKP4. Absolutely fine. There have been times when Duelund or Miflex was not around and that was the industry-solution. In between we had Mundorf (you remember my visit to you, Doede ?)
I cant explain exactly what goes on, I could try to make up theories...but any explanation would be a best guess as there are a lot of effects which are noise and timing relating going on in a source computer....there are tons of white papers on OsNoise in Linux alone. There is since Kernel 5.16 even a new tool...to measure OSNoise and Latency effects of anything you do in a white box-approach (RTLA)...I used that and found a clear correlation between lower OsNoise = better sound...As well less wattage, lower power consuption, lower Mhz sound typlially tonal richer and more natural, I guess because of lower EMF/EMI etc...plus the TSC clocksource in x86 is CPU freq. dependant...so we change the basis for the clock-system...and we lerned from Andrea that not the fastes clock is always the best clocksource....there has been a HR-Timer for the Alsa-output INterface of the OS-System ben introduced, which needs to be enbaled at Kernel Build time, a fulltickless mode for multicpu-uninterrupted applications etcetc
I could make up many theories. But that is all guestimating. I dont want to educate anyone to follow my path. I develop this for me. I just wanted to give people who are interested a hint.
And yes, i have the 500$ japanese sdtrans SD-card-Reader her as well on a a battery PSU (directly I2S connected to Andrea's Fifo...)...and you ll find long posts here in DIYaudio that this is regarded by many as the best digital source...most analog sounding etc...well...I got bad news...there are better sounding solutions,much more convenient and much cheaper...
And that was my motivation...better sdtrans384 and get convenience of MPD.
(so, anyone in EU interested in sdtrans latest version ...sendme PM)
I cant explain exactly what goes on, I could try to make up theories...but any explanation would be a best guess as there are a lot of effects which are noise and timing relating going on in a source computer....there are tons of white papers on OsNoise in Linux alone. There is since Kernel 5.16 even a new tool...to measure OSNoise and Latency effects of anything you do in a white box-approach (RTLA)...I used that and found a clear correlation between lower OsNoise = better sound...As well less wattage, lower power consuption, lower Mhz sound typlially tonal richer and more natural, I guess because of lower EMF/EMI etc...plus the TSC clocksource in x86 is CPU freq. dependant...so we change the basis for the clock-system...and we lerned from Andrea that not the fastes clock is always the best clocksource....there has been a HR-Timer for the Alsa-output INterface of the OS-System ben introduced, which needs to be enbaled at Kernel Build time, a fulltickless mode for multicpu-uninterrupted applications etcetc
I could make up many theories. But that is all guestimating. I dont want to educate anyone to follow my path. I develop this for me. I just wanted to give people who are interested a hint.
And yes, i have the 500$ japanese sdtrans SD-card-Reader her as well on a a battery PSU (directly I2S connected to Andrea's Fifo...)...and you ll find long posts here in DIYaudio that this is regarded by many as the best digital source...most analog sounding etc...well...I got bad news...there are better sounding solutions,much more convenient and much cheaper...
And that was my motivation...better sdtrans384 and get convenience of MPD.
(so, anyone in EU interested in sdtrans latest version ...sendme PM)
Last edited:
Well, because this is what this thread is about.
I never will understand, why people not owning the equipment go into equipment they are not interested in just to argue around with others with their theories.
buy Andreas Fifo/DAC. build a nice system. Join the club and than we talk again about what you experienced. UNtil then---BYE
I never will understand, why people not owning the equipment go into equipment they are not interested in just to argue around with others with their theories.
buy Andreas Fifo/DAC. build a nice system. Join the club and than we talk again about what you experienced. UNtil then---BYE
Maybe you didn't understand what was being claimed? Conducted noise from a computer can work its way into dac and result in jitter and or signal correlated noise. That's not the same thing as a claim that the mechanism is the same as SPDIF timing jitter, which is the type of thing you seem to be talking about.