|
|
Post by gld3gld3 on Nov 12, 2018 20:28:42 GMT -5
boomzilla when will you learn to STFU? Somehow your insistence on being right turns every one of your threads into a 200 page epic and you never listen to anyone. Plus, it is a violation of forum rules to run down other company's products. You know nothing about this topic yet you continue to talk out of your ass like you are some sort of expert. Please stop. Doh - I'm going to have to nominate this for the most hypocritical post of the year category. How many forum rules are being broken concerning courtesy, respect, insults and profanity? Methinks ol' DY needs a timeout to get his blood pressure down and reread the rules he apparently wants to selectively enforce. emotivalounge.proboards.com/thread/10157/forum-etiquette-readThe irony of "insisting being right" was not lost on me when seeing DY quote/repost himself within 3 hours. Gold, Jerry. Gold. We all have bad days. None of us is aware of what is going on in each other's lives. I try and be as respectful and positive as I can when posting online. If I start getting angry at something I give the forums a rest (edited for spelling mistake). Okay, time to go do some yoga 😉 |
|
klinemj
Emo VIPs 
Official Emofest Scribe
Posts: 15,095
|
Post by klinemj on Nov 12, 2018 21:36:48 GMT -5
That's kind of a fair summary... except for one thing. The USB interface itself is packet-based, so it inherently has a lot of jitter. It's up to the USB input circuitry on the receiving device to be immune to it. In other words, you can never "design a perfect USB source", because some of the issues involved are INHERENT in the USB interface.
The Eitr upgrades the input circuitry on the receiving device. Therefore, it has the opportunity to improve the input quality, by making it more immune to noise and jitter. This should allow the DAC to work very well with even less than perfect quality input signals. However, as I understand things, the UltraRendus are all sources, which places limitations on what they can even theoretically do. The source can do a lot to minimize noise, which can certainly help, especially if the receiving device is somewhat noise sensitive.
However, because USB is a packet-based interface, no matter how good your source is, it simply cannot eliminate all jitter. At best, it can make the receiver's job a little bit easier by minimizing the jitter as much as it can. (No matter how good your source is, a bad USB input on the DAC can still mung things up.)
From an engineering perspective, it seems easier, and more likely to succeed, to address these issues from the receiving end. However, I should note that the UltraRendu is a streaming device, and so it replaces A COMPUTER. Therefore, things like not having a fan, and simply being compact in size, and not requiring a keyboard and monitor, count in its favor.
The Eitr is a separate device, which you put between a source and a DAC you already have. The UltraRendu is a source device, and needs to be considered as such. (You're buying the Rendu INSTEAD of a computer and not IN ADDITION to it - and this would alter the value proposition considerably.)
(Of course, not all computers are big and expensive.... ie the Raspberry Pi is about $50...)
So, to put this in context of the Ultrarendu and other similar devices, the Eitr can fix known problems with bad USB implementations, while the Ultrarendus attempt to not create these USB problems in the first place. Some of the approaches taken to address these USB issues seem to be held in common between all devices being discussed here. Just. Go. Listen. To. One. (Really...) Mark |
|
|
|
Post by pedrocols on Nov 12, 2018 22:56:54 GMT -5
Kind of interesting how an Engineer conveys his knowledge which he acquired over years of school and practice and people still have the desire and feel compelled to argue just because they want to hear themselves talking. Most of the comments here including this one has zero value. So stop talking out of your ears just because you think you know something and want people to believe you know something...  |
|
|
|
Post by novisnick on Nov 12, 2018 23:17:20 GMT -5
The answer is, yes! One Roon ready device and it will work similarly to JRiver.
Done!
I’d say, for the good of the furum that this is the answer the OP was seeking. No beed to beat it any further.
👋
|
|
|
|
Post by pedrocols on Nov 12, 2018 23:34:09 GMT -5
The answer is, yes! One Roon ready device and it will work similarly to JRiver. Done! I’d say, for the good of the furum that this is the answer the OP was seeking. No beed to beat it any further.
👋I think his question was answered already 20 pages ago... |
|
|
|
Post by gld3gld3 on Nov 12, 2018 23:37:31 GMT -5
Kind of interesting how an Engineer conveys his knowledge which he acquired over years of school and practice and people still have the desire and feel compelled to argue just because they want to hear themselves talking. Most of the comments here including this one has zero value. So stop talking out of your ears just because you think you know something and want people to believe you know something... I mostly participate in this forum to learn. When I feel I have something to add or share I do so. It's great to share in our ideas, interests and experiences. Unfortunately, online forums are not perfect and can be very inefficient ways of communicating at times. For example, I am guilty of filling this thread with too many posts, but I felt like I was being misunderstood and wanted to make sure my ideas were coming across clearly. Not sure that I succeeded entirely. It turns out that I was not being as clear as I could have been. I am glad some people like Keith are able to step in and clarify some things in a rather graceful manner. |
|
|
|
Post by novisnick on Nov 12, 2018 23:52:55 GMT -5
The answer is, yes! One Roon ready device and it will work similarly to JRiver. Done! I’d say, for the good of the furum that this is the answer the OP was seeking. No beed to beat it any further.
👋I think his question was answered already 20 pages ago... Yep! And yet it won’t DIE! LOL |
|
|
|
Post by wilburthegoose on Nov 13, 2018 9:37:15 GMT -5
No - you can use Tidal without Roon. But IMHO, Roon significantly adds to the experience (great metadata, ability to integrate with your own collection of music). Tidal has a fine iOS app. Works great - I use it in my car all the time. PS - I'm definitely not a fan of JRiver for anything. I don't like their business model/leadership. This was post #2. Other than the fact that I forgot about Tidal's PC app, I think it answered Boom's question. |
|
|
|
Post by chicagorspec on Nov 13, 2018 10:57:49 GMT -5
Wow, a thread that evolved to discuss something tangential to the initial inquiry. I guess there's a first for everything...
|
|
|
|
Post by gld3gld3 on Nov 13, 2018 11:15:51 GMT -5
A timely release from Darko on the potential the Raspberry Pi has to provide good quality sound as a streamer (with attachments, but still at a lower cost than competitors): Click here for link to Youtube videoReally, watch it all the way through and there are some really interesting results/conclusions towards the end (subjective, I know)! |
|
|
|
Post by garbulky on Nov 13, 2018 12:27:20 GMT -5
Well, the GOAL of recordings is to be accurate to the source. The succeed or fail to varying degrees. The goal of these devices, however, is to make all recordings "sound good," depending on what sound the manufacturers thought would sell... So should your sound system be as accurate as possible, or homogenize everything to someone else's idea of "sounds good?" The “Goal” of recordings is to create art! And not necessarily to be accurate to the source. The “Goal” of these devices is to replicate that “art” . You make one heck of an assumption of what the final goal is. Art or copy? I’m Not trying to be argumentative, but we all see things differently and need to see through other people’s eyes. (Ears) Art is the producer's goal - the one that makes the recording happen. You make a good point though, sometimes the two go hand in hand. For instance Red Hot Chilli ppepper's albums are famous for horrible compression. In this case the goal is clearly not accuracy to source but towards the producer's vision - which in this case bafflingly made the recording sound quite a bit worse than it could have. Though it would be nice to say that the producer still wanted that s**ty sound to be recorded as accurately as possible, it's probably more accurate to say that he was out for a beer during this thought process.  But the reproduction of recordings - its purpose is just as confusing. Some people want exactly what's on the tape, while others prefer a certain sound. |
|
|
|
Post by Casey Leedom on Nov 13, 2018 12:53:12 GMT -5
A timely release from Darko on the potential the Raspberry Pi has to provide good quality sound as a streamer (with attachments, but still at a lower cost than competitors): Click here for link to Youtube videoReally, watch it all the way through and there are some really interesting results/conclusions towards the end (subjective, I know)! Very nice. Thanks for the reference. I do wish that he'd added a tiny bit of "interaction" description. For instance, early on when he described the Raspberry Pi's native USB output as being electrically noisy he simply said that could/would make the DAC sound bad, and a bit later he mentions that one of the "HAT"s would reduce that and "jitter". It would have been nice if he'd added that this could affect DACs which have poorly designed USB Inputs sections which are susceptible to electrical noise on the USB Input and/or don't regenerate their own clock for the DAC circuitry. But, overall, a really nice video. Casey |
|
|
|
Post by novisnick on Nov 13, 2018 14:10:37 GMT -5
The “Goal” of recordings is to create art! And not necessarily to be accurate to the source. The “Goal” of these devices is to replicate that “art” . You make one heck of an assumption of what the final goal is. Art or copy? I’m Not trying to be argumentative, but we all see things differently and need to see through other people’s eyes. (Ears) Art is the producer's goal - the one that makes the recording happen. You make a good point though, sometimes the two go hand in hand. For instance Red Hot Chilli ppepper's albums are famous for horrible compression. In this case the goal is clearly not accuracy to source but towards the producer's vision - which in this case bafflingly made the recording sound quite a bit worse than it could have. Though it would be nice to say that the producer still wanted that s**ty sound to be recorded as accurately as possible, it's probably more accurate to say that he was out for a beer during this thought process. But the reproduction of recordings - its purpose is just as confusing. Some people want exactly what's on the tape, while others prefer a certain sound. Yep! |
|
KeithL
Administrator 
Posts: 10,273
|
Post by KeithL on Nov 13, 2018 14:28:29 GMT -5
You've also got to remember that the Raspberry Pi is more than just a single product... it's a sort of ecosystem. (Darko is talking about one particular group of Pi accessories.... there are lots of others...)
For example, the power supply is just a standard wall wart, so you have a choice of which one to use (you could even use a fancy linear supply if you like). Likewise, there are several companies making "hats" for the Raspberry P, which are just standardized add-on boards. And there are, quite literally, thousands of cute little cases, cabinets, and mounting brackets available for them. (Everything from rainbow colored neon glow plastic, to machined metal, to a wooden case that looks like an old table radio, to brackets for mounting it on the back of your monitor.)
Also, unlike most audiophile gadgetry, the Raspberry Pi itself, and most of the Raspberry Pi accessories, are extremely affordable. A Raspberry Pi, including the computer itself, a case, and a power supply, will cost you less than $60 total.
(Incidentally.... HDMI... and multiple USB ports... and Ethernet... and WiFi... all built in the standard model... )
And, if your DAC is sensitive to USB noise, so you want to avoid using the direct USB output connection, you might want to try a S/PDIF output, with full galvanic isolation...
Here's an output hat, with a transformer isolated Coax S/PDIF output, and a Toslink output..... To get full galvanic isolation, you'll have to spend extra for the fancy premium version..... the premium version costs $35.
HiFiBerry makes a bunch of different Raspberry Pi audio accessories.
And, if you want to make a free standing music player, they even make an amplifier hat.
And, for software, you've also got a wide range of choices... Many people like Volumio for just audio... And Kodi (XMBC) is popular for those who want video too... Both are quite economical (actually they're both free).
The best part is that you can do a lot of experimentation and customization.... without breaking the bank.
A timely release from Darko on the potential the Raspberry Pi has to provide good quality sound as a streamer (with attachments, but still at a lower cost than competitors): Click here for link to Youtube videoReally, watch it all the way through and there are some really interesting results/conclusions towards the end (subjective, I know)! Very nice. Thanks for the reference. I do wish that he'd added a tiny bit of "interaction" description. For instance, early on when he described the Raspberry Pi's native USB output as being electrically noisy he simply said that could/would make the DAC sound bad, and a bit later he mentions that one of the "HAT"s would reduce that and "jitter". It would have been nice if he'd added that this could affect DACs which have poorly designed USB Inputs sections which are susceptible to electrical noise on the USB Input and/or don't regenerate their own clock for the DAC circuitry. But, overall, a really nice video. Casey |
|
|
|
Post by vortecjr on Nov 13, 2018 17:29:12 GMT -5
You've also got to remember that the Raspberry Pi is more than just a single product... it's a sort of ecosystem. (Darko is talking about one particular group of Pi accessories.... there are lots of others...)
For example, the power supply is just a standard wall wart, so you have a choice of which one to use (you could even use a fancy linear supply if you like). Likewise, there are several companies making "hats" for the Raspberry P, which are just standardized add-on boards. And there are, quite literally, thousands of cute little cases, cabinets, and mounting brackets available for them. (Everything from rainbow colored neon glow plastic, to machined metal, to a wooden case that looks like an old table radio, to brackets for mounting it on the back of your monitor.)
Also, unlike most audiophile gadgetry, the Raspberry Pi itself, and most of the Raspberry Pi accessories, are extremely affordable. A Raspberry Pi, including the computer itself, a case, and a power supply, will cost you less than $60 total.
(Incidentally.... HDMI... and multiple USB ports... and Ethernet... and WiFi... all built in the standard model... )
And, if your DAC is sensitive to USB noise, so you want to avoid using the direct USB output connection, you might want to try a S/PDIF output, with full galvanic isolation...
Here's an output hat, with a transformer isolated Coax S/PDIF output, and a Toslink output..... To get full galvanic isolation, you'll have to spend extra for the fancy premium version..... the premium version costs $35.
HiFiBerry makes a bunch of different Raspberry Pi audio accessories.
And, if you want to make a free standing music player, they even make an amplifier hat.
And, for software, you've also got a wide range of choices... Many people like Volumio for just audio... And Kodi (XMBC) is popular for those who want video too... Both are quite economical (actually they're both free).
The best part is that you can do a lot of experimentation and customization.... without breaking the bank.
Here is some more information you may find interesting: Darko alluded to some of the issues with a Raspberry Pi with regards to network and USB traffic. This is reported to occur with high sample rates when using Ethernet and USB for USB Audio. This issue can be worked around by using WiFi for network and or using i2s to an add on hat, but it comes with added work on your part. Audiophile gadgetry can be expensive especially in low quantities which is what you have in a niche market. Lets take one part, the oscillator, on a typical $30 add on hat. Ask yourself how much this one costs the developer and what kind of quality you can expect from it. We make an external USB to SPDiF converter and each oscillator costs us $10 in quantities of 1000 units. FYI the design takes 2 units. The oscillator we use have very low phase noise. Speaking of oscillators we use 2 units with different frequencies for a reason. When you see a board with only one oscillator it needs to do some math to able to handle the family of audio rates which are 44.1kHz based or 48kHz based. That math which calculates the frequency needed from the incorrect frequency adds jitter...I'll spare you the math. In regards to power you really want separate power for the clean side of things (DAC and audio oscillators) and the dirty side of things (processor and network). If you do this then you want galvanic isolation between the clean side and dirty side of things. I can go on, but you get the idea... I also agree that you will be doing a lot of experimentation and customization. No worries...been there, done that, and still doing it. |
|
|
|
Post by Casey Leedom on Nov 13, 2018 17:55:48 GMT -5
Hhmmm, WiFi "solves" an issue with the Raspberry Pi? Odd. The on-board RJ45 100Mb/s Ethernet and the USB 2.0 ports are all on the same USB 2.0 Hub, a LAN9514 chip hung off the BCM2837 USB interface (according to the documentation I can find). The WiFi/BlueTooth is implemented via a BCM43438 hung off the BCM2837's SPI/SDIO/UART pins ... So conceivably, if you were attempting to stream stupidly high-resolution audio one might worry about the bandwidth available through the USB 2.0 fabric ... but that should be capable of 480Mb/s and even DSD512 is "only" ~25Mb/s ...
Casey
|
|
KeithL
Administrator
Posts: 10,273
|
Post by KeithL on Nov 13, 2018 18:24:50 GMT -5
It all depends on the context in which you take things.
Some people use the Raspberry Pi for VIDEO..... Since it has its own HDMI output (HD but not 4k) you might conceivably be taking in streaming video over Ethernet... Or you might be reading video files via Ethernet from a NAS, or via USB from a local hard drive, in which case bandwidth would definitely be an issue. I wouldn't bet on its having enough processing power to decode h.265 files either (even at just HD resolution) - although I only use mine for audio.
Of course, if you're playing audio over USB, the same holds true (but, as you say, bandwidth shouldn't be an issue). (Of course, if you're playing USB audio from a USB hard drive, you're using that same hub to read the file and play the audio stream...)
Likewise, the Raspberry Pi is a "headless player", so you will be controlling it over the network - but using a web browser - which uses very little bandwidth.
The WiFi components in the Pi are also able to act as a WiFi hot-spot....
Therefore, you can connect to it directly, ad hoc, using WiFi, if there is no other network available.....
The Raspberry Pi is an interesting combination of cool and useful features... and some limitations...
Many people overestimate the amount of computing power necessary to play even high-res audio.
(I have an old laptop, with a slow single-core 1.7 gHz Celeron processor, that can play 24/192k audio flawlessly using Windows 7 or Windows 10.) However, with less capable hardware, you do have to be more careful to optimize things.
Hhmmm, WiFi "solves" an issue with the Raspberry Pi? Odd. The on-board RJ45 100Mb/s Ethernet and the USB 2.0 ports are all on the same USB 2.0 Hub, a LAN9514 chip hung off the BCM2837 USB interface (according to the documentation I can find). The WiFi/BlueTooth is implemented via a BCM43438 hung off the BCM2837's SPI/SDIO/UART pins ... So conceivably, if you were attempting to stream stupidly high-resolution audio one might worry about the bandwidth available through the USB 2.0 fabric ... but that should be capable of 480Mb/s and even DSD512 is "only" ~25Mb/s ... Casey |
|
KeithL
Administrator
Posts: 10,273
|
Post by KeithL on Nov 13, 2018 18:49:46 GMT -5
I haven't watched the video yet (I may have a chance tonight)... However, as I understand it, he's talking about a SPECIFIC set of Raspberry Pi "accessories"... being offered as a "system".
Therefore, he's presumably addressing a lot of stuff in that context...
For example, in my experience, USB powered DACs tend to be especially sensitive to a noisy USB power supply (because it is providing their power source). However, DACs that have their own AC power source, and so don't run from the USB power supply, tend to be much less so. (So that noisy USB power might be an issue if you connect the Pi to an Ego DAC, but probably not if you connect it to an XMC-1.)
Some of his comments seems to simply be about what I would call "good practices"....
The standard samples rates used by USB audio fall into two groups... multiples of 44.1k and multiples of 48k. In general, it's easier to generate much cleaner clock signals at all the multiples if you start with separate "masters" - one for each family of multiples (44/88/176 and 48/96/192).
This is standard practice for good quality asynchronous USB inputs (there are "super quality multipliers" that produce excellent performance with a single master clock - but they are also expensive). It's always a good idea to use good quality clocks - with good low phase noise specs. (However, many DAC manufacturers like to brag about their clock specs, when things like bad board layout can easily degrade the performance of even the best clock chip... so, again, it's the end result that counts.)
Likewise, it's always good practice to use isolated, and separately filtered power for analog and digital circuitry. However, this can be implemented by running the entire DAC from a separate power source, or by running the individual parts of circuitry in the DAC from separately filtered and regulated supplies. (Remember that the DAC itself has both digital and analog components... so should really have multiple separately regulated supplies anyway.)
Likewise, galvanic isolation will never hurt... but sometimes it helps and sometimes it doesn't make much difference.
I would take all of that discussion under the general heading of "if you want it to work well then there are lots of best practices you should be aware of and follow".
(The Raspberry Pi is essentially a little computer - and most of what you would consider to apply to ANY computer audio source also applies to it.) They're basically "showing off" that they understand the best practices involved and followed them when designing their device.... which is a good thing.
The reality is that good performance is some areas actually contributes to better OVERALL performance; but, in some areas, it is just a waste... and raises costs. For example, using a really high quality power supply to supply reference voltage to a DAC is a good idea... But powering an Ethernet switch with anything other than the cheapest possible SMPS that can do the job and meet spec is simply a waste of money better spent elsewhere...
I should note, for the record, that I've always used my Raspberry Pi as a DIGITAL AUDIO source, connected to a separate DAC, via USB. I haven't used any of the "dedicated DAC hats" or "dedicated S/PDIF output hats" for it... so I don't really know how one or the other of them performs.
(Although I see no reason to believe they would fail to perform as advertised.)
You've also got to remember that the Raspberry Pi is more than just a single product... it's a sort of ecosystem. (Darko is talking about one particular group of Pi accessories.... there are lots of others...)
For example, the power supply is just a standard wall wart, so you have a choice of which one to use (you could even use a fancy linear supply if you like). Likewise, there are several companies making "hats" for the Raspberry P, which are just standardized add-on boards. And there are, quite literally, thousands of cute little cases, cabinets, and mounting brackets available for them. (Everything from rainbow colored neon glow plastic, to machined metal, to a wooden case that looks like an old table radio, to brackets for mounting it on the back of your monitor.)
Also, unlike most audiophile gadgetry, the Raspberry Pi itself, and most of the Raspberry Pi accessories, are extremely affordable. A Raspberry Pi, including the computer itself, a case, and a power supply, will cost you less than $60 total.
(Incidentally.... HDMI... and multiple USB ports... and Ethernet... and WiFi... all built in the standard model... )
And, if your DAC is sensitive to USB noise, so you want to avoid using the direct USB output connection, you might want to try a S/PDIF output, with full galvanic isolation...
Here's an output hat, with a transformer isolated Coax S/PDIF output, and a Toslink output..... To get full galvanic isolation, you'll have to spend extra for the fancy premium version..... the premium version costs $35.
HiFiBerry makes a bunch of different Raspberry Pi audio accessories.
And, if you want to make a free standing music player, they even make an amplifier hat.
And, for software, you've also got a wide range of choices... Many people like Volumio for just audio... And Kodi (XMBC) is popular for those who want video too... Both are quite economical (actually they're both free). The best part is that you can do a lot of experimentation and customization.... without breaking the bank.
Here is some more information you may find interesting: Darko alluded to some of the issues with a Raspberry Pi with regards to network and USB traffic. This is reported to occur with high sample rates when using Ethernet and USB for USB Audio. This issue can be worked around by using WiFi for network and or using i2s to an add on hat, but it comes with added work on your part. Audiophile gadgetry can be expensive especially in low quantities which is what you have in a niche market. Lets take one part, the oscillator, on a typical $30 add on hat. Ask yourself how much this one costs the developer and what kind of quality you can expect from it. We make an external USB to SPDiF converter and each oscillator costs us $10 in quantities of 1000 units. FYI the design takes 2 units. The oscillator we use have very low phase noise. Speaking of oscillators we use 2 units with different frequencies for a reason. When you see a board with only one oscillator it needs to do some math to able to handle the family of audio rates which are 44.1kHz based or 48kHz based. That math which calculates the frequency needed from the incorrect frequency adds jitter...I'll spare you the math. In regards to power you really want separate power for the clean side of things (DAC and audio oscillators) and the dirty side of things (processor and network). If you do this then you want galvanic isolation between the clean side and dirty side of things. I can go on, but you get the idea... I also agree that you will be doing a lot of experimentation and customization. No worries...been there, done that, and still doing it. |
|
|
|
Post by vortecjr on Nov 13, 2018 20:04:43 GMT -5
I haven't watched the video yet (I may have a chance tonight)... However, as I understand it, he's talking about a SPECIFIC set of Raspberry Pi "accessories"... being offered as a "system".
Therefore, he's presumably addressing a lot of stuff in that context...
For example, in my experience, USB powered DACs tend to be especially sensitive to a noisy USB power supply (because it is providing their power source). However, DACs that have their own AC power source, and so don't run from the USB power supply, tend to be much less so. (So that noisy USB power might be an issue if you connect the Pi to an Ego DAC, but probably not if you connect it to an XMC-1.)
Some of his comments seems to simply be about what I would call "good practices"....
The standard samples rates used by USB audio fall into two groups... multiples of 44.1k and multiples of 48k. In general, it's easier to generate much cleaner clock signals at all the multiples if you start with separate "masters" - one for each family of multiples (44/88/176 and 48/96/192).
This is standard practice for good quality asynchronous USB inputs (there are "super quality multipliers" that produce excellent performance with a single master clock - but they are also expensive). It's always a good idea to use good quality clocks - with good low phase noise specs. (However, many DAC manufacturers like to brag about their clock specs, when things like bad board layout can easily degrade the performance of even the best clock chip... so, again, it's the end result that counts.)
Likewise, it's always good practice to use isolated, and separately filtered power for analog and digital circuitry. However, this can be implemented by running the entire DAC from a separate power source, or by running the individual parts of circuitry in the DAC from separately filtered and regulated supplies. (Remember that the DAC itself has both digital and analog components... so should really have multiple separately regulated supplies anyway.)
Likewise, galvanic isolation will never hurt... but sometimes it helps and sometimes it doesn't make much difference.
I would take all of that discussion under the general heading of "if you want it to work well then there are lots of best practices you should be aware of and follow".
(The Raspberry Pi is essentially a little computer - and most of what you would consider to apply to ANY computer audio source also applies to it.) They're basically "showing off" that they understand the best practices involved and followed them when designing their device.... which is a good thing.
The reality is that good performance is some areas actually contributes to better OVERALL performance; but, in some areas, it is just a waste... and raises costs. For example, using a really high quality power supply to supply reference voltage to a DAC is a good idea... But powering an Ethernet switch with anything other than the cheapest possible SMPS that can do the job and meet spec is simply a waste of money better spent elsewhere...
I should note, for the record, that I've always used my Raspberry Pi as a DIGITAL AUDIO source, connected to a separate DAC, via USB. I haven't used any of the "dedicated DAC hats" or "dedicated S/PDIF output hats" for it... so I don't really know how one or the other of them performs.
(Although I see no reason to believe they would fail to perform as advertised.) You flipped the conversation on me:) You guys were talking about add on hats and not external USB DACs. No worries. It doesn't really matter how the DAC is powered or how expensive the DAC. For example DACs with their own power supply, with USB inputs, and unbalanced analog outputs can pass source noise very easily to their outputs. Not so much with balanced analog outputs on the same DAC. You said, "But powering an Ethernet switch with anything other than the cheapest possible SMPS that can do the job and meet spec is simply a waste of money better spent elsewhere..." I don't disagree and recommend this a lot. However, there is some research from John S. to suggest that there is noise from the power supply of a switch or router making its way over the network. He has some rules about how to avoid it with a grounding trick. This is very low level noise, but noise none the least. Again I don't disagree and feel there are more important fish to fry. |
|
klinemj
Emo VIPs
Official Emofest Scribe
Posts: 15,095
|
Post by klinemj on Nov 14, 2018 23:54:19 GMT -5
I haven't watched the video yet (I may have a chance tonight)... However, as I understand it, he's talking about a SPECIFIC set of Raspberry Pi "accessories"... being offered as a "system".
Therefore, he's presumably addressing a lot of stuff in that context... For example, in my experience, USB powered DACs tend to be especially sensitive to a noisy USB power supply (because it is providing their power source). However, DACs that have their own AC power source, and so don't run from the USB power supply, tend to be much less so. (So that noisy USB power might be an issue if you connect the Pi to an Ego DAC, but probably not if you connect it to an XMC-1.) Some of his comments seems to simply be about what I would call "good practices".... The standard samples rates used by USB audio fall into two groups... multiples of 44.1k and multiples of 48k. In general, it's easier to generate much cleaner clock signals at all the multiples if you start with separate "masters" - one for each family of multiples (44/88/176 and 48/96/192).
This is standard practice for good quality asynchronous USB inputs (there are "super quality multipliers" that produce excellent performance with a single master clock - but they are also expensive). It's always a good idea to use good quality clocks - with good low phase noise specs. (However, many DAC manufacturers like to brag about their clock specs, when things like bad board layout can easily degrade the performance of even the best clock chip... so, again, it's the end result that counts.)
Likewise, it's always good practice to use isolated, and separately filtered power for analog and digital circuitry. However, this can be implemented by running the entire DAC from a separate power source, or by running the individual parts of circuitry in the DAC from separately filtered and regulated supplies. (Remember that the DAC itself has both digital and analog components... so should really have multiple separately regulated supplies anyway.)
Likewise, galvanic isolation will never hurt... but sometimes it helps and sometimes it doesn't make much difference.
I would take all of that discussion under the general heading of "if you want it to work well then there are lots of best practices you should be aware of and follow".
(The Raspberry Pi is essentially a little computer - and most of what you would consider to apply to ANY computer audio source also applies to it.) They're basically "showing off" that they understand the best practices involved and followed them when designing their device.... which is a good thing.
The reality is that good performance is some areas actually contributes to better OVERALL performance; but, in some areas, it is just a waste... and raises costs. For example, using a really high quality power supply to supply reference voltage to a DAC is a good idea... But powering an Ethernet switch with anything other than the cheapest possible SMPS that can do the job and meet spec is simply a waste of money better spent elsewhere...
I should note, for the record, that I've always used my Raspberry Pi as a DIGITAL AUDIO source, connected to a separate DAC, via USB. I haven't used any of the "dedicated DAC hats" or "dedicated S/PDIF output hats" for it... so I don't really know how one or the other of them performs.
(Although I see no reason to believe they would fail to perform as advertised.) You flipped the conversation on me:) You guys were talking about add on hats and not external USB DACs. No worries. It doesn't really matter how the DAC is powered or how expensive the DAC. For example DACs with their own power supply, with USB inputs, and unbalanced analog outputs can pass source noise very easily to their outputs. Not so much with balanced analog outputs on the same DAC. You said, "But powering an Ethernet switch with anything other than the cheapest possible SMPS that can do the job and meet spec is simply a waste of money better spent elsewhere..." I don't disagree and recommend this a lot. However, there is some research from John S. to suggest that there is noise from the power supply of a switch or router making its way over the network. He has some rules about how to avoid it with a grounding trick. This is very low level noise, but noise none the least. Again I don't disagree and feel there are more important fish to fry. He really needs to actually listen to some options. As a great person once said "all models are wrong and some are useful". He's operating in a wrong model that is no longer useful. A listen to some of the last options would confirm that. Mark |
|
