Perfect, exactly what I was hoping. Thanks!
It sounds to me like there would be very little real-world, easily-noticeable difference between the Mamboberry LS DAC+ and the Boss DAC, in stock form. Plus they cost about the same.
--
Precisely. And in the world of RPi audio, not a lot of money for a sometimes bewildering choice. And a further choice as to whether you leave things as they are, or start tweaking
Precisely. And in the world of RPi audio, not a lot of money for a sometimes bewildering choice. And a further choice as to whether you leave things as they are, or start tweaking
This gentleman has made me $150 poorer - but i've been enjoying a RPi and a nice sounding DAC!
Back to topic...
Audiophonics i-sabre V3 gets my vote.
This gentleman has made me $150 poorer - but i've been enjoying a RPi and a nice sounding DAC!
Back to topic...
Audiophonics i-sabre V3 gets my vote.
150$? That's just the first installment
Welcome to the RPi merry-go-round, where it stops, no body knows...
So where does it stop?
I've been doing some interesting comparisons the past several weeks. Last night I did a preliminary head-to-head of my modified Mamboberry (original, not LS) and a 'twinned'-mode Piano 2.1 (1 channel off each DAC chip), both on a Kali on an Ian IsolatorPi (in beta) and then on Pi. In this setup the Pi, IsolatorPi, Kali, and tested DAC all were powered by separate supplies... a modified K&K Audio Low Voltage Power Supply (if you go for their 12 watt version, it will be very similar to my modified 4 watt version) for the RPi and 3 Uptone Audio LPS-1 supplies, 1 each on the IsolatorPi, Kali, and target DAC. The RPi ran PcP, same exact setup for both DACs, I just powered the stack down and swapped the DACs out, then powered back up.
All of the associated LMS server machine & networking gear are run off DIY'd linear supplies.
In addition to each other, I also compared them to my modified Sony HAP Z1-ES. This is a $2000 USD retail hard-drive player to which I've applied some serious mods that dampen the structure, beef up and optimize the power supplies (including converting the DC-DC converters used for the digital processing side of the unit to linear regulators) and upgrade signal path parts. I recently had the chance to compare it to a totally stock unit and it is much above that. AND I got a chance to see a 'professionally' modified unit with $2600 of mods performed. My mods were loosely inspired by what that modifier did while going further... so by a similar yardstick, my mods could be valued in the $3000-$3400 USD range for a total value of roughly $5000.
You can see some of my mods to the Sony in this thread:
http://www.diyaudio.com/forums/digital-source/252985-sony-hap-z1es-hi-res-source-new-2014-a-2.html
I've taken it a bit farther since my last posts there.
First, with all separate good power supplies, the Piano 2.1 Twinned clearly topped the modified Mamboberry... more real, more space. The MB was pretty good, but the Piano sounded more like real music.
Set the Mamboberry aside and compare the Piano 2.1 to the Sony... hmmm, close. The Sony bests the Piano, but not by much. I wouldn't feel like I downgraded much if I had to stick with the Piano. The Sony led on inner detailing and palpability. And it should... better DACs in true dual mono mode with current output so you can use a better I/V stage by themselves should have the Sony beating the ES9023 MB or PCM5122 Boss or 2x PCM5142 Piano.
Then in my other system, I have a Boss on a Kali on an RPi, also running PcP with a duplicate of the LMS server/networking setup with the linear supplies. While I haven't gone head-to-head Boss vs Piano, these systems are similar in sound and what I hear from the Boss in Kali-fed slave mode is a bit more transparency than the Piano 2.1 with sweeter highs and mids (those SMD film caps). I'll go head to head between the 2 soon AND compare the Boss in master vs slave modes... a quick check a couple of weekends ago when 1st setting up the IsolatorPi had me preferring the Kali-fed slave mode, but both were good.
My lessons out of all this?
1. You can get very good sound out of an R-Pi setup, but power supply and signal integrity is key. On the power supply side I consider a good linear (and likely also the iFi Soundcheck-modified... I haven't tried one myself, but have no reason to doubt his assessment) as a minimum.
2. Also as a minimum for good sound use separate supplies on the R-Pi and DAC (or Kali-DAC stack... the twinned Piano 2.1 still sounds very good when sharing a supply with the Kali). You can always start with a single shared supply and add more as budget allows for improvements in SQ.
3. If you don't want to start with a Kali, go for a Boss in master mode. If you can start with a Kali, then it's a matter of preference for your ears and system... both the Boss and Piano 2.1 are good on a Kali. To my ears and in my setups I think the Boss is better, but others might decide differently.
4. Reclocking and isolation are great in this setups. An upgrade path might be a RPi / Boss combo with 1 supply (the Boss has best-in-breed noise filtering), later add a separate supply for the Boss and again later a Kali. And top it off someday with a couple more power supplies and either Ian's or the up-coming Allo isolator.
Soundcheck, Yatsushiro, and DJ have some great advice here and I'd follow all if starting out.
But the final lesson here is that a well-powered and setup RPi DAC stack in an optimized setup (see Soundchecks blog on the LMS server & networking optimization) can hold its own with some of the higher price spread... even if you discount the improvements and cost of the my mods to the Sony, that my RPi setup could come that close even to a stock HAPZ would be a feat.
RPi setups, done right, aren't just for low-fi anymore!
Greg in Mississippi
I've been doing some interesting comparisons the past several weeks. Last night I did a preliminary head-to-head of my modified Mamboberry (original, not LS) and a 'twinned'-mode Piano 2.1 (1 channel off each DAC chip), both on a Kali on an Ian IsolatorPi (in beta) and then on Pi. In this setup the Pi, IsolatorPi, Kali, and tested DAC all were powered by separate supplies... a modified K&K Audio Low Voltage Power Supply (if you go for their 12 watt version, it will be very similar to my modified 4 watt version) for the RPi and 3 Uptone Audio LPS-1 supplies, 1 each on the IsolatorPi, Kali, and target DAC. The RPi ran PcP, same exact setup for both DACs, I just powered the stack down and swapped the DACs out, then powered back up.
All of the associated LMS server machine & networking gear are run off DIY'd linear supplies.
In addition to each other, I also compared them to my modified Sony HAP Z1-ES. This is a $2000 USD retail hard-drive player to which I've applied some serious mods that dampen the structure, beef up and optimize the power supplies (including converting the DC-DC converters used for the digital processing side of the unit to linear regulators) and upgrade signal path parts. I recently had the chance to compare it to a totally stock unit and it is much above that. AND I got a chance to see a 'professionally' modified unit with $2600 of mods performed. My mods were loosely inspired by what that modifier did while going further... so by a similar yardstick, my mods could be valued in the $3000-$3400 USD range for a total value of roughly $5000.
You can see some of my mods to the Sony in this thread:
http://www.diyaudio.com/forums/digital-source/252985-sony-hap-z1es-hi-res-source-new-2014-a-2.html
I've taken it a bit farther since my last posts there.
First, with all separate good power supplies, the Piano 2.1 Twinned clearly topped the modified Mamboberry... more real, more space. The MB was pretty good, but the Piano sounded more like real music.
Set the Mamboberry aside and compare the Piano 2.1 to the Sony... hmmm, close. The Sony bests the Piano, but not by much. I wouldn't feel like I downgraded much if I had to stick with the Piano. The Sony led on inner detailing and palpability. And it should... better DACs in true dual mono mode with current output so you can use a better I/V stage by themselves should have the Sony beating the ES9023 MB or PCM5122 Boss or 2x PCM5142 Piano.
Then in my other system, I have a Boss on a Kali on an RPi, also running PcP with a duplicate of the LMS server/networking setup with the linear supplies. While I haven't gone head-to-head Boss vs Piano, these systems are similar in sound and what I hear from the Boss in Kali-fed slave mode is a bit more transparency than the Piano 2.1 with sweeter highs and mids (those SMD film caps). I'll go head to head between the 2 soon AND compare the Boss in master vs slave modes... a quick check a couple of weekends ago when 1st setting up the IsolatorPi had me preferring the Kali-fed slave mode, but both were good.
My lessons out of all this?
1. You can get very good sound out of an R-Pi setup, but power supply and signal integrity is key. On the power supply side I consider a good linear (and likely also the iFi Soundcheck-modified... I haven't tried one myself, but have no reason to doubt his assessment) as a minimum.
2. Also as a minimum for good sound use separate supplies on the R-Pi and DAC (or Kali-DAC stack... the twinned Piano 2.1 still sounds very good when sharing a supply with the Kali). You can always start with a single shared supply and add more as budget allows for improvements in SQ.
3. If you don't want to start with a Kali, go for a Boss in master mode. If you can start with a Kali, then it's a matter of preference for your ears and system... both the Boss and Piano 2.1 are good on a Kali. To my ears and in my setups I think the Boss is better, but others might decide differently.
4. Reclocking and isolation are great in this setups. An upgrade path might be a RPi / Boss combo with 1 supply (the Boss has best-in-breed noise filtering), later add a separate supply for the Boss and again later a Kali. And top it off someday with a couple more power supplies and either Ian's or the up-coming Allo isolator.
Soundcheck, Yatsushiro, and DJ have some great advice here and I'd follow all if starting out.
But the final lesson here is that a well-powered and setup RPi DAC stack in an optimized setup (see Soundchecks blog on the LMS server & networking optimization) can hold its own with some of the higher price spread... even if you discount the improvements and cost of the my mods to the Sony, that my RPi setup could come that close even to a stock HAPZ would be a feat.
RPi setups, done right, aren't just for low-fi anymore!
Greg in Mississippi
Last edited:
Greg.
Interesting writeup. I'm wondering if your setup complies to the OPs idea of a "cheap HAT".
Would be nice if you could also tell us what software and what setup, including alsamixer settings for PCM51xx dacs you're using. A single (wrong) setting can make "the difference".
How does your "twinned" Piano2.1 setup look like? You might have seen over at the Piano21
thread that I finally got stuck after putting some hours into it.
On the other side, by looking at your conclusion "Piano21 vs. Boss", my motivation to put any more work into this is kind of fading away. You basically confirm what I was expecting. And since I still prefer my Sabre9023 over these (my) PCM51xx dacs putting any more effort would be just for the fun of it.
And... ...not a single word about your Soekris - to put all what's being said into perspective.
Interesting writeup. I'm wondering if your setup complies to the OPs idea of a "cheap HAT".
Would be nice if you could also tell us what software and what setup, including alsamixer settings for PCM51xx dacs you're using. A single (wrong) setting can make "the difference".
How does your "twinned" Piano2.1 setup look like? You might have seen over at the Piano21
thread that I finally got stuck after putting some hours into it.
On the other side, by looking at your conclusion "Piano21 vs. Boss", my motivation to put any more work into this is kind of fading away. You basically confirm what I was expecting. And since I still prefer my Sabre9023 over these (my) PCM51xx dacs putting any more effort would be just for the fun of it.
And... ...not a single word about your Soekris - to put all what's being said into perspective.
I also don`t know this guy, and just read this one article from him. His measurement method is sound, though. And the differences are that small, that I can really say that anyone thinking one sounds better over the other do it in a purely subjective manner, and as such should redefine better as "better for my particular system, room, and MY ears"...
Despite what the "golden ears" think about themselves here, their brains still won`t be able to distinguish a 0.055% difference. Reminds me of the discussion where audiophiles were bragging about the warms and transparency of coat hangers used as speaker cables, and how they sounded better than 1000 dollar cables.
Despite what the "golden ears" think about themselves here, their brains still won`t be able to distinguish a 0.055% difference. Reminds me of the discussion where audiophiles were bragging about the warms and transparency of coat hangers used as speaker cables, and how they sounded better than 1000 dollar cables.
Last edited:
Greg.
Interesting writeup. I'm wondering if your setup complies to the OPs idea of a "cheap HAT".
The overall setup may be above the "cheap" level, but both the Boss & Piano are not what I'd call expensive. AND they are my Favs AND are so with no performance-enhancing mods.
Adding a Kali does up the cost, but that can be done incrementally OR just go with the Boss in master mode and call it a day.
Would be nice if you could also tell us what software and what setup, including alsamixer settings for PCM51xx dacs you're using. A single (wrong) setting can make "the difference".
As I posted 2x over in this thread http://www.diyaudio.com/forums/vend...pi-subwoofer-out-2-1-2-2-a-3.html#post5077022 with the Piano I am using PcP (which means Squeezelite) with the Generic I2S driver (which means no alsamixer settings available).
After hearing how well that setup worked, I wonder if part of it is using the default settings and NOT having the driver mess up the settings.
The Boss is with PcP 3.20 (still Beta... I need to set it up with the released version). No Alsamixer settings done here EXCEPT choosing slave mode (for the Kali) and selecting the Ringing-less Low Latency FIR filter via the Squeezelite controls.
See above and my posts in this and the Piano 2.1 thread. If any other specifics not answered, please ask.
I (and I am sure all of the RPi audio community that knows what you do) appreciate all the work you put into this environment behind the scenes and your published info-guides. We are lucky and happy to have you here!
As for PCM51x2 DACs versus ES9023 ones, some like one, some the other. I find good elements in both, but so far I've not been able to get my ES9023-based setups to equal the best of my PCM51x2 (& other PCMxxxx) setups.
Late last summer I did a pretty gonzo, over-the-top VREF mod as required for the best sound on the Rev1 DAMs, using a combination of SMD PPS and Black Gate caps... at current Black Gate prices over $800 of them per each of my 2 Soekris DAM DAC (luckily I stocked up on BGs while they were still available, so didn't have to go out and source them at this cost!). Results were good after the requisite Black Gate break-in. BUT with the improvements to my R-Pi setups starting last fall (upgraded networking, improved power supplies, off-the-grid LPS-1 supplies, and Kali reclockers) and my modified Sony HAP Z1-ES (additional case damping, further beefed up power supplies, better power supply rectifier diodes), my Soekris are now loosing out in many aspects. They still have that RTR-DAC quality that the Delta-Sigma cannot match, but I really need to do the power supply upgrades I have planned for them to keep them competitive with my other setups.
Did I cover everything??
Greg in Mississippi
P.S. DJ, I'm still putting Ian's IsolatorPi through its paces and am not ready to say much on its impact. Should be soon, within the next couple of weeks. BUT notice I still have my name in for 3 more units on the Group Buy thread!
P.P.S. One main reason for doing all the tests / mods on on I2S-connected RPi setups over the last 1 1/2 years is to understand how high can we get their performance. As of now, I'd say 'pretty darned high', but of course GREAT power supplies come at a cost. OTOH, good power supplies give a very sizable fraction of GREAT with much less cost. And you also have to add the cost of signal integrity interfaces (reclockers & isolators). Still, compared to what I heard on a stock HFBD+P or original Mamboberry directly on an RPi with a single power supply for both, my current setups are in an entirely different realm.
Many thanks to you Soundcheck... and to HiFiBerry (1st master mode RPi DAC), to Collybia (Mamboberry DACs with VG regulators & power supply bypassing on board), to Allo (Piano and Boss DACs with best-in-class power supplies and of course the IMHO indispensable Kali reclocker) and to Ian (and soon Allo) for isolator boards. AND many others, there are great aspects to many of the other DACs I've seen, but not listened to such as IQAudio, Durio Audio, Dial Audio (up and running soon!), Dion, Sabreberry, and others.
Where the Pi setups butt heads against a ceiling (IMHO) is in NOT having the board real estate to implement some of the refinements seen in higher-end setups... extreme power supply techniques such as high-current drinking / hot-running shunt regulators, dual-mono'd differential DACs (like in the Sony HAPZ, but with the right drivers, the Piano 2.1 might get there), DACs with external I/V (such as the ESS90xx & PCM179x, among others) and associated great I/V & output stages. The Piano 2.1 in 'twinned' mode shows what is available with 2 DAC chips (and I suspect a Super-Piano with some of the tweaks from the Boss, but slave-mode only and using a Kali+isolator MIGHT be the peak of what can be done currently), different chips (AKM?) or architectures (Soekris's mini RPi DAM?) might move things further, but you still have the problems of sufficient board space to implement other refinements. Adding some of this via stacked boards has worked well, but when you get to 3-4 boards stacked on top of your R-Pi and that's not enough, you are beginning to hit the wall!
Oh, and another 'refinement' might be to move towards synchronous-reclocked hardware-implemented master mode (I've seen it done, though can't share details).
Last edited:
@Greg Stewart, thanks!
I am trying to do what you did, but no success. I choose generice/simple ESS9023. Squeezelite is running but no sound. I do not use the Kali (because I didn't want to dismantle my main sound system). Did you make a change in the Picoreplayer settings?
I am trying to do what you did, but no success. I choose generice/simple ESS9023. Squeezelite is running but no sound. I do not use the Kali (because I didn't want to dismantle my main sound system). Did you make a change in the Picoreplayer settings?
DJ,
This setup is running an older version of pCP, 1.22.
In the pCP SqueezeLite configuration thread, I did use Soundcheck's preferred ALSA settings "65536:8:32:1", Buffer Size "50000:600000", Restrict Codec to "pcm,flac", & Priority Setting to 45. I'd try these first on your current pCP setup before trying it on your Kali.
Good luck!
Greg in Mississippi
P.S. I did try it with both Piano 2.1's I have here and both worked ok.
This setup is running an older version of pCP, 1.22.
In the pCP SqueezeLite configuration thread, I did use Soundcheck's preferred ALSA settings "65536:8:32:1", Buffer Size "50000:600000", Restrict Codec to "pcm,flac", & Priority Setting to 45. I'd try these first on your current pCP setup before trying it on your Kali.
Good luck!
Greg in Mississippi
P.S. I did try it with both Piano 2.1's I have here and both worked ok.
Since the Boss gets so much love, I bought one. It's up and running in my RPi3 running Moode 3.6.
Last night I had the Pi-Boss-Moode playing FLACs from a CD, comparing its playback to the original CD playing in a Pioneer PD-D6-J SACD player (which I think sounds pretty good). I think the Pi-Boss-Moode sounds just ever-so-slightly clearer, like each instrument stands out as its own separate entity just a bit more. If I want to really pick nits, I'd say the Pi-Boss-Moode sounds just a touch brighter than the SACD player. Otherwise quite comparable.
I then played some 24-96 FLAC files of an album which I also have on SACD. Again, comparable -- high -- sound quality. (I need to go back and compare those two again, since it was late and I needed to get to sleep.)
For the little money and large convenience of that little box holding 400+ albums worth of music ready to play with such good sound quality, I'm extremely happy with it!
The default setup in Moode is to run the Boss in slave mode, and that's how it's set now. I just read your post again and see that you're advising to set it to master mode. I'll try that tonight, but I don't know if that's supported by Moode. I hope it is...
--
This has only just occurred to me, reading your post about slave mode being default... but The Boss is a master DAC, so what is going on here in Moode? The slave mode is for when you want to run it with a Kali, so does one have to change this setting in Moode if using it without the Kali, or is it not important? Surely this setting doesn't mean The Boss is using the clock from the Pi?!
Ignore me, I realised/remembered that Moode actually sets to 'slave disabled' as the default, rather than 'slave'. Panic averted.
Agh, I'm sorry. Yes, the default is 'slave disabled' which of course means the Boss DAC is in Master mode. My bad.
Greg wrote:
Now that I have a Boss up and running, I'd like to try to find the best settings in Moode. I hope there's a way in Moode to select 'Ringing-less Low Latencey FIR filter' as a setting with SoX disabled. I'll have to see when I get home this evening...
Right now, I have SoX set to oversample to 32bit/384kHz, with its filtering set to Medium Quality (100dB noise suppression).
Being that I don't understand the inner workings of these DACs' oversampling and noise filters very well, I'm trying to learn what all the various settings mean. I'd like to hear how the Boss DAC performs after turning off all the driver and OS software and letting the Boss DAC's internal oversampling and filtering do the work. I'm not sure which combination of Moode/MPD/ALSA settings accomplishes that.
--
Greg wrote:
Now that I have a Boss up and running, I'd like to try to find the best settings in Moode. I hope there's a way in Moode to select 'Ringing-less Low Latencey FIR filter' as a setting with SoX disabled. I'll have to see when I get home this evening...
Right now, I have SoX set to oversample to 32bit/384kHz, with its filtering set to Medium Quality (100dB noise suppression).
Being that I don't understand the inner workings of these DACs' oversampling and noise filters very well, I'm trying to learn what all the various settings mean. I'd like to hear how the Boss DAC performs after turning off all the driver and OS software and letting the Boss DAC's internal oversampling and filtering do the work. I'm not sure which combination of Moode/MPD/ALSA settings accomplishes that.
--
To answer my own question about which oversampling/filtering settings sound 'better' with the Boss DAC...
Using 16bit/44.1kHz FLAC files ripped from my own CD collection as source material, I compared the output from the Moode-RPi3-BossDAC to that from my Pioneer PD-D6-J SACD player (which I've enjoyed for some time now). The rest of the system is an Intact Audio inductive attenuator volume control into a Hafler P1000 power amp and Snell E/III speakers. I also listened in headphones through an Objective 2 amp, with Fostex T50P or Sony MDR-7506 headphones.
In all cases I had the ALSA volume control disabled, SoX multithreading = yes.
I experimented with SoX vs. no SoX. What I found was:
- With no SoX, just the Boss DAC set to do its own oversampling, I heard a sort of 'tinkly' sound, which made the electric guitar on Santana Abraxas sound a bit thin and drums sound bright and a bit lightweight. It still sounds 'good' but has a sort of 'electronic-ness' to it that I don't particularly like.
- With SoX set to 32bits/384kHz oversampling and Very High Quality noise reduction, the sound is subjectively very clean, but ever so slightly overly 'smooth'. I like this sound. I'm going to listen to it this way for a bit, to see if this is the setting I like best.
- With SoX set to 32bits/384kHz oversampling and Sox quality = High Quality, I find the sound is very clear, a bit bright compared to my Pioneer SACD player, but sounding more 'detailed.' The highs have more of that 'tinkly' sound to them than with SoX quality = Very High Quality (as opposed to only High Quality). The sound is just a little more 'exciting' than with SoX quality = Very High Quality. I like this sound too.
- Setting SoX to 32bits/*kHz and High Quality resulted in a more smoothed-over sound, reminding me of what CDs have sounded like for years (if that makes any sense). I don't like this sound as much as the previous two settings with 384kHz oversampling.
So, I seem to like SoX oversampling/filtering better than the PCM5122's internal oversampling/noise filtering. I'm surprised at the fairly large differences in subjective sound quality between the various oversampling/filtering settings.
This setup sounds every bit as good as any digital player I've had in my system. It's cheaper than my other favorite player, and holds over 400 albums too.
I'm looking forward to upgrading the psu to see what that does for the sound.
--
Using 16bit/44.1kHz FLAC files ripped from my own CD collection as source material, I compared the output from the Moode-RPi3-BossDAC to that from my Pioneer PD-D6-J SACD player (which I've enjoyed for some time now). The rest of the system is an Intact Audio inductive attenuator volume control into a Hafler P1000 power amp and Snell E/III speakers. I also listened in headphones through an Objective 2 amp, with Fostex T50P or Sony MDR-7506 headphones.
In all cases I had the ALSA volume control disabled, SoX multithreading = yes.
I experimented with SoX vs. no SoX. What I found was:
- With no SoX, just the Boss DAC set to do its own oversampling, I heard a sort of 'tinkly' sound, which made the electric guitar on Santana Abraxas sound a bit thin and drums sound bright and a bit lightweight. It still sounds 'good' but has a sort of 'electronic-ness' to it that I don't particularly like.
- With SoX set to 32bits/384kHz oversampling and Very High Quality noise reduction, the sound is subjectively very clean, but ever so slightly overly 'smooth'. I like this sound. I'm going to listen to it this way for a bit, to see if this is the setting I like best.
- With SoX set to 32bits/384kHz oversampling and Sox quality = High Quality, I find the sound is very clear, a bit bright compared to my Pioneer SACD player, but sounding more 'detailed.' The highs have more of that 'tinkly' sound to them than with SoX quality = Very High Quality (as opposed to only High Quality). The sound is just a little more 'exciting' than with SoX quality = Very High Quality. I like this sound too.
- Setting SoX to 32bits/*kHz and High Quality resulted in a more smoothed-over sound, reminding me of what CDs have sounded like for years (if that makes any sense). I don't like this sound as much as the previous two settings with 384kHz oversampling.
So, I seem to like SoX oversampling/filtering better than the PCM5122's internal oversampling/noise filtering. I'm surprised at the fairly large differences in subjective sound quality between the various oversampling/filtering settings.
This setup sounds every bit as good as any digital player I've had in my system. It's cheaper than my other favorite player, and holds over 400 albums too.
I'm looking forward to upgrading the psu to see what that does for the sound.
--
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.