Euphony OS w/Stylus player setup and issues thread

[aangen]

17 hours ago, flkin said:

 

There are many technologists that claim this is the case and are happy using the most basic setup but I'm sure most of us here disagree. Somehow there's an art to delivering bit perfect info to the DAC and the way it's done greatly influences the sound. That Ethos must be something special to sound better than your MU1. Are you still using that streamer? Did you manage to get Euphony into the Grimm for the comparison?

 

Kin

Hi Kin!

I am still enjoying my Grimm MU1 a great deal. It is a wonderful device! I am not shopping for an upgrade. 
 

The Ethos just seems to dig more out of 16.44.1 than I imagined possible. It’s really quite special!

 

Cheers! 
Al

[flkin]

duplicate

[flkin]

12 hours ago, al2813 said:

..

3. The point my friend raised to me was the difference in sound quality that can be observed when ripping the same CD. As discussed earlier SW has a big part in this (bit perfect etc.), but HW seems to matter as well (where in theory it shouldn't since the 1s and 0s are identical). Of course hearing the difference is subject to many factors some more or less objective (the level of your setup) and some totally subjective ("analogue sound"). The type of music, the quality of the mastering and the recording will also play in this scenario

It’s not the 1s and 0s that are different, it’s the process of taking that data and delivering it to the DAC that is different. 
 

One starts from an optical interface and the other starts from a storage medium like a SSD. And the path to the DAC is totally different. Somehow along the way what arrives is electrically different while the data remains the same. Data is in a binary digital format only in theory. In practice the information is delivered in an analogue form. I’m not surprised that CD playback sounds different compared to computer playback.

 

The full buffering of tracks before playback function of Euphony is an attempt to reduce this effect. Electrically, local files and streamed files all start from the same location (RAM) after buffering and undergo the same processes and path to the DAC. The offline function further reduces the effect of the connected LAN. Thereby reducing upstream component effects too. 

[Exocer]

Hi All -

Has anyone attempted to use Intel i350 or i210 based fiber NICs with Euphony? Were you successful in this endeavor?

 

If so, did you need to use a specific kernel?

 

I can't get either of these NICs to pick up an IP on my network. It's as if the kernel does not properly detect either NIC. Of course, diagnostics are difficult with Euphony being so locked down. Interested in hearing the experiences of others. I figure i'll ask here and also open a ticket with Euphony support.

 

Worth noting:
1. The SFPs I'm using are indeed compatible with the fiber nic.

2. The NIC was tested and confirmed to be functional prior to sending it to me (and the source is trustworthy).

 

Cheers and TGIF all.

 

-Rob

[dc-audiogeek]

The pace of updates for Euphony Stylus has slowed down considerably. We're paying subscription prices now, so we should be getting regular fixes in return. Please, when you encounter bugs or missing features with the software, create a ticket and report them to Euphony Audio. I just reported the lack of support for multi-disc albums (they're shown as discrete multiple albums, not as a single album with a number indicating how many discs) in the way the library is set up. That's been an issue for years. Robert is answering support tickets now, so you're making requests of the boss directly when you create one.

[Exocer]

4 hours ago, dc-audiogeek said:

The pace of updates for Euphony Stylus has slowed down considerably. We're paying subscription prices now, so we should be getting regular fixes in return. Please, when you encounter bugs or missing features with the software, create a ticket and report them to Euphony Audio. I just reported the lack of support for multi-disc albums (they're shown as discrete multiple albums, not as a single album with a number indicating how many discs) in the way the library is set up. That's been an issue for years. Robert is answering support tickets now, so you're making requests of the boss directly when you create one.

I submitted a ticket. Will keep you all posted.

[flkin]

12 hours ago, dc-audiogeek said:

..

Robert is answering support tickets now, so you're making requests of the boss directly when you create one.

When I reported the fact that all v4 versions until the recent version was accessing the local storage ssd every second or so despite playing fully buffered local or streaming tracks Robert’s response was:

 

“ There is nothing to worry about. Everything is functioning normally nor effect on SQ .. “

 

On further insisting this should not be occurring and that the matter is significant, eventually Željko replied that it was being looked into. And this was resolved in the latest v0531. 
 

Robert might be the boss but he might not be prioritizing details in the same way as some of us. Might have to gently push if you feel it necessary 

[di-fi]

On 7/1/2023 at 2:20 AM, flkin said:

And this was resolved in the latest v0531. 

Hi Kin,

Do you know what the current access rate is?
Like at the change of tracks, possibly including a few seconds before?

Is there a way to prolong this? How does the “sit and relax” option handle this access rate in the latest version?

 

Thanks for any insight.

[di-fi]

On 7/1/2023 at 2:20 AM, flkin said:

Robert might be the boss but he might not be prioritizing details in the same way as some of us. Might have to gently push if you feel it necessary 

After spending 6 to 12 months with the new subscription model, I still perceive a certain level of deception by some users. The way Euphony communicates and follows up has changed, requiring users to adjust their expectations. While the sound quality remains excellent, in my opinion, there is since vagueness and a reduction in certain aspects of the follow-up.

 

Based on my analysis, it seems unlikely that Euphony will return to its previous business model. Željko's hands-on approach, constant communication with forum members (especially on AS), and his exceptional problem-solving skills were truly remarkable. However, this approach, although admirable, can lead to burnout, as it entails countless (probably unpaid) hours driven by passion and ingenuity, which greatly impressed those who visited this thread.

 

Unfortunately, I understand that this leaves a bitter taste of not being heard, similar to seeking rewards for requests on Roon forums. Perhaps running a profit-oriented business in this manner is the only way to cater to the majority of users who are less interested in fine-tuning to their precise needs?

[flkin]

3 hours ago, di-fi said:

Hi Kin,

Do you know what the current access rate is?
Like at the change of tracks, possibly including a few seconds before?

Is there a way to prolong this? How does the “sit and relax” option handle this access rate in the latest version?

 

Thanks for any insight.

 

Hi Paul,

 

How I know there is local disc access is because my Pink Faun 2.16x has an LED light on the front panel which blinks each time the disc is accessed. Ever since v4 came out, I noticed that there was a blink around once per second despite play back from RAM or even during Ramroot. This didn't occur during v3. I just didn't get around to writing to Robert/Željko about it until recently. Turned out to be some redundant log entries.

 

With the latest v0531, the blinking stopped. Using this as a guide, I see buffered files in RAM working as they should be, without local disc access apart from the flash right before the track loading it into RAM. For normal playback, buffering is the current and next track. So on the first track, 2 tracks are loaded into RAM and it's pretty fast - like maybe 2-3 seconds or so (dependent on network, whether disc is local or networked like in a NAS etc I guess). So the disc access affects only the first few seconds of playback which is too short to be significant. At the end of the first track, the third track is loaded into RAM while the second track starts playing. Again it's a very short disc access and again not significant. 

 

There is that blue indicator showing that the track is loaded into RAM. I can confirm that in v0531 there is no further disc access once that blue indicator shows. 

 

The sit and relax mode is simply a way to pre-load all tracks into RAM before disconnecting from the network for a set time. It's like a turntable playback, lower the needle and sit back to listen. I noticed that this sit and relax mode works for both local files as well as streamed ones. You can buffer entire streamed albums into RAM and then use the sit and relax mode so that's a great function which was not available in v3 before. Mixed playlists of local and streamed tracks seems to work fine also. Tidal seems to load files faster than Qobuz which takes many more seconds to preload. With v0531, playlist buffered to RAM and using sit and relax mode, Euphony will theoretically offer the most electrically 'quiet' playback possible today in any playback software I am aware of. 

 

All of the above is related to local internal disc access not NAS disc access which is outside the streamer. For that, I believe there is a cache to local function you can use which copies frequently used albums into the local cache drive. Then the tracks are treated as local from there. I don't use this cache function this but I guess the buffering function will take a lot longer especially for those using 100Mbps ethernet speeds (eg. passing data through A to B side of the existing EtherRegen, although there are rumours the next ER version will be faster).

 

What I would like see happen is that the sit and relax mode uses the time duration for the remainder of the playlist rather than the entire playlist duration. In the same way that the "Buffer to RAM" button buffers the remaining playlist into RAM rather than the entire playlist.  Sometime back I ask Željko about adjusting for this but his reply was I wasn't using his software the right way. ☺️ It was designed for entire albums/playlists, click sit and relax button and listen to the whole album/playlist. Not start the sit and relax function half way through a playlist! Looks like I have to drop the turntable Stylus at the beginning of the record and not in the middle of the record! 😆 

 

Kin

[flkin]

3 hours ago, di-fi said:

After spending 6 to 12 months with the new subscription model, I still perceive a certain level of deception by some users. The way Euphony communicates and follows up has changed, requiring users to adjust their expectations. While the sound quality remains excellent, in my opinion, there is since vagueness and a reduction in certain aspects of the follow-up.

 

Based on my analysis, it seems unlikely that Euphony will return to its previous business model. Željko's hands-on approach, constant communication with forum members (especially on AS), and his exceptional problem-solving skills were truly remarkable. However, this approach, although admirable, can lead to burnout, as it entails countless (probably unpaid) hours driven by passion and ingenuity, which greatly impressed those who visited this thread.

 

Unfortunately, I understand that this leaves a bitter taste of not being heard, similar to seeking rewards for requests on Roon forums. Perhaps running a profit-oriented business in this manner is the only way to cater to the majority of users who are less interested in fine-tuning to their precise needs?

 

The pay once use forever business model isn't sustainable for esoteric software like Euphony. The market size is just too small. Getting renewed income is the only way the business can continue. Just have to pay the annual fee if we want to keep Euphony alive. They will loose customers from the subscription method for sure -  from their point of view they probably decided to only keep those that makes them some profit and let the rest go. Not going for market share anymore.

 

For sure they have slowed down in updates, more effort to improve the software would be appreciated and expected actually. Optimised USB drivers and kernels would be nice. So would being able to edit track names and other tags. A better integrated search system for local and streamed files together would be really nice. 

 

I know Željko reads these message from time to time. Perhaps everyone can list out their wish lists for him and team to read?

[di-fi]

8 hours ago, flkin said:

 

Hi Paul,

 

How I know there is local disc access is because my Pink Faun 2.16x has an LED light on the front panel which blinks each time the disc is accessed. Ever since v4 came out, I noticed that there was a blink around once per second despite play back from RAM or even during Ramroot. This didn't occur during v3. I just didn't get around to writing to Robert/Željko about it until recently. Turned out to be some redundant log entries.

 

With the latest v0531, the blinking stopped. Using this as a guide, I see buffered files in RAM working as they should be, without local disc access apart from the flash right before the track loading it into RAM. For normal playback, buffering is the current and next track. So on the first track, 2 tracks are loaded into RAM and it's pretty fast - like maybe 2-3 seconds or so (dependent on network, whether disc is local or networked like in a NAS etc I guess). So the disc access affects only the first few seconds of playback which is too short to be significant. At the end of the first track, the third track is loaded into RAM while the second track starts playing. Again it's a very short disc access and again not significant. 

 

There is that blue indicator showing that the track is loaded into RAM. I can confirm that in v0531 there is no further disc access once that blue indicator shows. 

 

The sit and relax mode is simply a way to pre-load all tracks into RAM before disconnecting from the network for a set time. It's like a turntable playback, lower the needle and sit back to listen. I noticed that this sit and relax mode works for both local files as well as streamed ones. You can buffer entire streamed albums into RAM and then use the sit and relax mode so that's a great function which was not available in v3 before. Mixed playlists of local and streamed tracks seems to work fine also. Tidal seems to load files faster than Qobuz which takes many more seconds to preload. With v0531, playlist buffered to RAM and using sit and relax mode, Euphony will theoretically offer the most electrically 'quiet' playback possible today in any playback software I am aware of. 

 

All of the above is related to local internal disc access not NAS disc access which is outside the streamer. For that, I believe there is a cache to local function you can use which copies frequently used albums into the local cache drive. Then the tracks are treated as local from there. I don't use this cache function this but I guess the buffering function will take a lot longer especially for those using 100Mbps ethernet speeds (eg. passing data through A to B side of the existing EtherRegen, although there are rumours the next ER version will be faster).

 

What I would like see happen is that the sit and relax mode uses the time duration for the remainder of the playlist rather than the entire playlist duration. In the same way that the "Buffer to RAM" button buffers the remaining playlist into RAM rather than the entire playlist.  Sometime back I ask Željko about adjusting for this but his reply was I wasn't using his software the right way. ☺️ It was designed for entire albums/playlists, click sit and relax button and listen to the whole album/playlist. Not start the sit and relax function half way through a playlist! Looks like I have to drop the turntable Stylus at the beginning of the record and not in the middle of the record! 😆 

 

Kin

 

[flkin]

The importance of the RAM buffer strikes again!

 

https://alpha-audio.net/review/zzyzx-revelation-femto-ssd-bit-by-bit/

 

If everything is buffered and played from RAM, an audiophile RAM powered externally is what is needed

[Smaragdhk]

51 minutes ago, flkin said:

The importance of the RAM buffer strikes again!

 

https://alpha-audio.net/review/zzyzx-revelation-femto-ssd-bit-by-bit/

 

If everything is buffered and played from RAM, an audiophile RAM powered externally is what is needed

 

Exactly the reason why I DON'T USE the "buffer to RAM", this way all the audio is streamed from the femto SSD.

 

[flkin]

3 hours ago, Smaragdhk said:

 

Exactly the reason why I DON'T USE the "buffer to RAM", this way all the audio is streamed from the femto SSD.

 

 

Can a computer use data from SSD into CPU without going through the RAM? Not sure this is possible..

[Smaragdhk]

55 minutes ago, flkin said:

 

Can a computer use data from SSD into CPU without going through the RAM? Not sure this is possible..

 

Well, there is a clear difference if one ticks the "buffer to RAM' or not... not entirely sure why, but yes, I agree... it somehow should use the RAM modules when playing back, for that reason I also chose the Apacer RAM. 

[StreamFidelity]

31 minutes ago, Smaragdhk said:

Well, there is a clear difference if one ticks the "buffer to RAM' or not...

 

Processing in Audio PC passes through several memory layers. First in the CPU itself with L1-L3 caches. The L3 caches and L2 caches are memory in the CPU. L2 for the respective core and L3 for multi-core operation. The higher the memory, the lower the latencies, since less has to be moved to the slower working RAM. However, the CPU memories are still very small. For an Intel i9-13900K L3=36MB and L2=32MB.

 

A DRAM of today usually has 16GB or more. The RAM has much higher latencies than L2/L3 Caches, but much lower latencies than an SSD. Here is a good overview:

 

Source: PC Perspective INTEL’S OPTANE DC PERSISTENT MEMORY DIMMS PUSH LATENCY CLOSER TO DRAM

 

The data from an SSD (or from streaming) is therefore always loaded into the RAM beforehand. Without RAM the computer does not work. 

What distinguishes normal data processing from buffers to RAM, when data always goes through RAM? 

 

With buffer to RAM, a complete piece of music or an entire album is loaded into RAM. The route from the SSD to RAM is no longer needed. Advantages: there is less noise and interrupts due to lower acitivities. The music can be played more interference-free. 

 

But it should not be forgotten that other memories are nevertheless unavoidable. For example the buffer at USB or Ethernet. It is all a complex interaction.
 

[di-fi]

9 hours ago, Smaragdhk said:

I also chose the Apacer RAM

Agreed, and in particular the (industrial) wide temperature type. Select the better performing Apacer type memory and do not just buy any type because of the brand name.

[di-fi]

8 hours ago, StreamFidelity said:

 

Processing in Audio PC passes through several memory layers. First in the CPU itself with L1-L3 caches. The L3 caches and L2 caches are memory in the CPU. L2 for the respective core and L3 for multi-core operation. The higher the memory, the lower the latencies, since less has to be moved to the slower working RAM. However, the CPU memories are still very small. For an Intel i9-13900K L3=36MB and L2=32MB.

 

A DRAM of today usually has 16GB or more. The RAM has much higher latencies than L2/L3 Caches, but much lower latencies than an SSD. Here is a good overview:

 

Source: PC Perspective INTEL’S OPTANE DC PERSISTENT MEMORY DIMMS PUSH LATENCY CLOSER TO DRAM

 

The data from an SSD (or from streaming) is therefore always loaded into the RAM beforehand. Without RAM the computer does not work. 

What distinguishes normal data processing from buffers to RAM, when data always goes through RAM? 

 

With buffer to RAM, a complete piece of music or an entire album is loaded into RAM. The route from the SSD to RAM is no longer needed. Advantages: there is less noise and interrupts due to lower acitivities. The music can be played more interference-free. 

 

But it should not be forgotten that other memories are nevertheless unavoidable. For example the buffer at USB or Ethernet. It is all a complex interaction.
 

Interesting explanation. But besides the used memory types, there definitely is also an important audible relation with the chosen quality of power (supply) and how the “better” power is connected to RAM or SSD.

 

For example the M.2 Femto SSD being powered through external 5V by a very good LPS can still augment SQ of it’s music files in a considerable way according to most of it’s users even tough latency is below par in above example. 
Or SSD’s can be each mounted on a different power board so isolation and good, ideally exclusive 5V power, can elevate SQ in a very substantial way.
 

Both Innuos and Antipodes are exploiting the use of exclusively SSD for music files with excellent power with no attention for special quality RAM. Antipodes uses one specific type of Samsung SSD and Innuos is now using SLC ram for their OS in at least the most recent top server.

[brucew]

3 hours ago, di-fi said:

Interesting explanation. But besides the used memory types, there definitely is also an important audible relation with the chosen quality of power (supply) and how the “better” power is connected to RAM or SSD.

 

For example the M.2 Femto SSD being powered through external 5V by a very good LPS can still augment SQ of it’s music files in a considerable way according to most of it’s users even tough latency is below par in above example. 
Or SSD’s can be each mounted on a different power board so isolation and good, ideally exclusive 5V power, can elevate SQ in a very substantial way.
 

Both Innuos and Antipodes are exploiting the use of exclusively SSD for music files with excellent power with no attention for special quality RAM. Antipodes uses one specific type of Samsung SSD and Innuos is now using SLC ram for their OS in at least the most recent top server.

What is unclear to me is if I set music to buffer to RAM and get an industrial grade Apacer RAM, does Optane do me any good, or are they alternate approaches to SQ, practically speaking? Separately powering HDD/SDD or RAM is not an option for me.

[di-fi]

59 minutes ago, brucew said:

What is unclear to me is if I set music to buffer to RAM and get an industrial grade Apacer RAM, does Optane do me any good, or are they alternate approaches to SQ, practically speaking? Separately powering HDD/SDD or RAM is not an option for me.

Hi, 

It can be confusing, absolutely. Let me try:

 

Besides the usual configuration where operating system and files are on the same drive, the operating system can also be located on an high quality Optane drive (the smaller size like 16 or 32 Gb.) The Optane drive sits between the processor and the SSD or HDD drives with the music files (the latter being larger in size from 1 Gb and up usually)

 

The RAM is only temporary memory. It’s empty when it is not powered. But essential for a music computer to work. It will only temporarily hold the music files (possibly in buffer, like a whole song or  complete album) coming from :

* local NVMe or SSD

* from a NAS

* streaming services like Tidal.

 

Good quality RAM will benefit SQ. RAM with good power will sound even better.

 

Also good quality SSD or NVMe (where sound files are located) will benefit SQ, and again with good power they will sound better. With independent external power they sound best. But it is not always obvious to power them independently. 
 

Hope this helps

[davide256]

5 hours ago, brucew said:

What is unclear to me is if I set music to buffer to RAM and get an industrial grade Apacer RAM, does Optane do me any good, or are they alternate approaches to SQ, practically speaking? Separately powering HDD/SDD or RAM is not an option for me.

I think you are confusing RAM boot with buffer to RAM. RAM boot and Optane drive boot are functionally similar, which sounds better will

depend on the HW design of the machine, ie which contaminates the power supply bus with less noise during use.  RAM buffer is a convenience

for album/playlist playback. and can be used to minimize network activity during local playback. Since Optane as a drive has become an obsolete feature,

Apacer memory with an Intel 670P SSD is a good option.

 

 

[flkin]

Internal drives are a necessary evil in streamers. At the very least we need one for the OS and more (or partitions) if we store music, playlists etc in them. If their power comes from the ATX or the motherboard, like the case of a NVMe  storages, the drive is sharing some of fits electrical disturbances to the system or is being affected by the pretty noise PC environment. Externally power drives like the Relevation Femto addresses this with external power. Some companies, like Pink Faun, uses their own power boards for their drives to isolate from this disturbance. 
 

There’s a lot discussed about the noise and sound character of different drive brands and models but less so about how they are powered or even less the method used for formatting the drive. However the electrical signature of a drive is the sum total of the drive’s inherent noise, the noise added from its power supply during use and the way the OS takes the info from the drive. We can’t really hear drives in isolation.

 

After surviving this evil, we just move what we can into the ram and work from there. Let the drives sleep and get out of the way electrically as much as possible. The ram has lowest latency, fastest connection to the CPU and is needed anyway for computer operations. 

[brucew]

8 hours ago, davide256 said:

I think you are confusing RAM boot with buffer to RAM. RAM boot and Optane drive boot are functionally similar, which sounds better will

depend on the HW design of the machine, ie which contaminates the power supply bus with less noise during use.  RAM buffer is a convenience

for album/playlist playback. and can be used to minimize network activity during local playback. Since Optane as a drive has become an obsolete feature,

Apacer memory with an Intel 670P SSD is a good option.

 

13 hours ago, di-fi said:

Besides the usual configuration where operating system and files are on the same drive, the operating system can also be located on an high quality Optane drive (the smaller size like 16 or 32 Gb.) The Optane drive sits between the processor and the SSD or HDD drives with the music files (the latter being larger in size from 1 Gb and up usually)

 

Good quality RAM will benefit SQ. RAM with good power will sound even better.

 

Also good quality SSD or NVMe (where sound files are located) will benefit SQ, and again with good power they will sound better. With independent external power they sound best. But it is not always obvious to power them independently.

If booting to RAM, then in theory an Optane drive would have little further effect I guess. If booting to an Optane drive but buffering files to RAM, then the Optane relates to OS/software related overhead and noise and perhaps cache but not to the difference bewteen buffered & non-buffered files.

 

I didn't realise that Optane as a drive is obsolete. I've been seeing some for sale and thought I'd seen 11th gen NUC's spec'd to be compatible. (my budget lives closer to the NUC world than more configurable ITX, etc.)

[davide256]

34 minutes ago, brucew said:

 

If booting to RAM, then in theory an Optane drive would have little further effect I guess. If booting to an Optane drive but buffering files to RAM, then the Optane relates to OS/software related overhead and noise and perhaps cache but not to the difference bewteen buffered & non-buffered files.

 

I didn't realise that Optane as a drive is obsolete. I've been seeing some for sale and thought I'd seen 11th gen NUC's spec'd to be compatible. (my budget lives closer to the NUC world than more configurable ITX, etc.)

I couldn't get an 11i3 NUC to recognize 32gb of Optane memory that I'd used in an 8 series NUC as a drive. There are still some

Optane SSD's floating around, those should work but avoid the hybrid ones.