Life after Yggdrasil?
Jul 18, 2016 at 3:22 AM Post #526 of 1,366
The only tough part about @Torq
 assessment is that his own personal brain being used to the yggdrasil's sound may inhibit his ability to hear another sound signature as actually better, even for his tastes. Usually we know what we like right off the bat but there are exceptions.

I have tried gear myself and used what I had first as a reference. That made a lot of other gear sound off in balance...until I actually had it a few days and let my mind get used to the sound of the other gear. Then I was able to listen with less bias and later my opinion of the new comer changed. 
...


definitely. what wows you at first may not be something that will wow you over time.
...


What both of you are saying is true for me as well.

However, as I read Torq's work in this thread a couple of things come to mind that leave me unconcerned.

1) He seems to be spending a considerable amount of time listening to a considerable amount of gear.

2) He seems to have both: a discerning ear that's been around the block more than once, and a reasonably disciplined approach (yet without letting that suck all of the joy out of the process).

Also, having met him at a local meet, shared gear and thoughts probably adds a fair bit to this confidence. :)
 
Jul 18, 2016 at 1:53 PM Post #527 of 1,366

PS Audio

I’m going to cover a few points common to both the DirectStream and DirectStream Junior DACs first and then discuss differences and sound individually.
 
..
 
One of the key features of the PS Audio DirectStream DACs it that they can be upgraded, by the user, in a very simple process.  Essentially, you download the new software (which is defines the configuration or programming of the FPGA, copy it to an SD card (or USB stick for the Junior), and plug it in and reboot the unit.
 
Since the implementations for up/over-sampling, PCM-to-DSD conversion, and digital-to-analog conversion, as well as overall system operation, are implemented on the FPGA, these “software” upgrades have the potential to radically alter the performance of the unit.
 
This is a smart move.
 
Linn does something similar (although their FPGA doesn’t do the actual D-to-A part of the job – that’s done with a D/S chip, and the FPGA handles up/over-sampling, filtering, and room-correction).  And over many upgrades to my Linn DSM units, there have been significant improvements in sound quality – so the feature is potentially huge.
 
Chord also use an FPGA implementation for their DACs, however they’ve chosen not to make these user-upgradeable and if they ever released an update (which I don’t think is likely based on comments from Rob Watts, their chief designer), the units would have to be returned to the factory to be modified; so unless no updates are planned there (e.g. say if Rob didn’t think he can make further improvements), that’s a lost opportunity in my opinion.
 
In the case of the PS Audio units, does this upgradability translate into anything useful?
 
I have to say that’s a resounding “YES”!
 
I deliberately began my audition on an older OS (“Yale”).  Upgrading to the latest version, “Torreys” brought immediate, obvious and surprisingly wide ranging improvements (not just differences).
 
This is very exciting feature, with huge potential.  Upgradability is very nice to have and it being something that the user can do themselves, quickly and easily, makes it even better. 
 
[size=17.03px]DirectStream DAC [/size][size=17.03px](+)[/size]
 
When running on the “Yale” operating system, I felt the DirectStream was good, but not really as competitive as it was going to need to be to warrant going on to the next round of evaluation.  Detail was a little smoothed in the upper registers, there was a bit more “mass” in the mid-range than I believe was properly authentic, and dynamics were not quite on the level that I’d expect with this tier of product.
It was still an enjoyable listen but wasn’t quite good enough for me to want to take further.
 
Installing the “Torreys” release made a surprisingly large improvement, almost across the board.  A simple software update transformed the DAC and elevated it from “good but no cigar” to fully competitive, and even edging ahead in overall musicality and emotion.  The remainder of these comments are based on listening to the “Torreys” operating system release (which is the latest release as of this writing).
 
This doesn’t quite reach the detail levels of DAVE, which is only fractionally ahead of Yggdrasil there anyway, but the DirectStream is largely on par with Yggdrasil.  They do seem to trade places here and there with different recordings.  Detail and resolution are, nonetheless, very high and there’s no sense of artificiality or exaggeration to that detail, unlike, say, some of the Sabre based units (Auralic’s Vega in particular).
 
Tone is very pure, with excellent timbral rendering.  The nemesis for most DACs in these evaluations, rending of piano, was handled extremely well with the only discernible difference coming from a sense of very-slightly (and I do mean VERY slightly) slower attack from the PS Audio unit.
 
Imaging/sound-stage … is taken by Yggdrasil.  It yields a better sense of three dimensionality and  slightly better separated 2D projection, although the DirectStream certainly projects a solid, stable and 3D image – if a bit behind in the vertical domain.
I found the overall sound to be extremely well balanced and perhaps a bit richer than Yggdrasil, particularly in the mid-range and treble.  Articulation, space/air, instrumental separation were all fully competitive.
 
Dynamics/transients just favored Yggdrasil.  There was a slightly better sense of things starting/stopping instantly with Schiit’s DAC.
Any sense of glare, harshness or artificiality was completely absent.  Music was portrayed in a very smooth and engaging manner and it was hard not to get lost in it and just sit listening.  Yggdrasil has sometimes been accused of being clinical, analytical or lean/bright, and while I do not share those opinions, they’re definitely not adjectives one would apply to the DirectStream unit.
 
Overall this was a fabulous listen that didn’t give up much, if anything, to Yggdrasil and where it did fall short, it wasn’t buy much or in more than a couple of places.  This absolutely had to go on the shortlist, not just because of how it sounds, but also because of the (realized) potential of it’s upgradability.
 
[size=17.03px]DirectStream Junior [/size][size=17.03px](+)[/size][size=17.03px](!)[/size]
 
PS Audio claim this is 85% of the performance of its progenitor, at two-thirds the price.  The deal is actually better than that, though, as the price of the Junior includes the “Network Bridge II” card.  This would be an $899 option for the bigger DAC.
 
Then there’s trade-in options, which are available direct from the manufacturer, and which can reduce the price of the unit substantially.  This slants things further in favor of the Junior if you want that Network interface (and I definitely would, see below).
 
Using the same trade-in option for both units, you get $4,999 + $899 for the DirectStream w/ Network Bridge II vs. $2,999 (also with the network card, since it’s included).  That means you’re paying just over half the price of the bigger DAC to get largely the same capabilities and most of the performance.
 
So, now on to the sound … which will be a VERY brief discussion as departures from the bigger DAC are just not that large.
 
This sounds uncannily like its big brother.  I doubt, without listening the way I do when I’m auditioning, that I’d be able to tell the difference at all.  Junior loses out a little on the detail side of things,  dynamics/transients aren’t quite at the level of big-brother, and it is a little behind on imaging, but tone/timbre, PRaT and involvement are all right there with the “full” version.
 
And in some situations I found that, musically, I preferred the Junior – this occurred primarily when feeding my WA5-LE in single-ended mode (it’s a single ended design, despite offering convenience balanced inputs and outputs).  I’m not sure why this would be the case, unless it’s down to the output stage on the Junior not being dual/differential (which it’s big brother is), but it was not a universal thing – mostly seeming to affect detail and transients.
 
The short version here is that I feel the Junior is a lot closer to its more expensive relative than the 15% differential that PS Audio cite.  Maybe that’s from measurements, I don’t know, but sitting and listening to them … I’d say it’s closer to 95%.  I would, for example, say the Junior is closer to the full-blown DirectStreamm than Gungnir Multi-bit is to Yggdrasil (and they’re pretty close, even if their signatures are different).
 
Given that this is so close to its big brother, is significantly cheaper, and that the full-version easily earned its place in my shortlist, this unit has to go on there as well.  I could, for example, hedge my bets and buy one of these, get the advantage of the included Network interface, and another Yggdrasil and still come in shy of the price of the full-blown DirectStream DAC.
 
[size=17.03px]Network Bridge II[/size]
 
For most DACs I was NOT comparing different inputs – I pretty much used AES/EBU if it was available, TOSLINK if it wasn’t, and only USB if it was the only option and/or I was experimenting with DSD.
 
With the PS Audio units, I did use both their AES/EBU inputs and the Network Bridge II.
 
There’s not a lot to say here … except that I generally preferred listening to both units via their Network interfaces.  Detail and timing seemed improved, as did instrumental separation and layering using the Bridge as an input.  And if I was buying the DirectStream I would definitely get it with the Network Bridge II option.
 
Also, since I’m a Roon user, the upcoming support for Roon would be something that’s very nice to have.
 
[size=17.03px]A footnote on DSD playback:[/size]
 
I did try DSD playback with these units and did NOT suffer any of the dropout issues that I was getting with the Chord units.  I didn’t find that feeding the unit native-DSD improved its performance at all (nor pre-converted PCM content – but that’s probably because it can do its own conversion internally), but I will say that, of the handful of DACs I did bother testing with DSD, this pair performed the best.
 
That applies in both function and sound.
 
Note that I can’t count the Chord units in the comparison there because, while they did sound good when they were playing, drop-outs above single-rate DSD (which I didn’t like in comparison to ANYTHING) were so frequent that there was no way one could listen to an entire track without multiple interruptions and being jarred out of the moment.
 
Jul 18, 2016 at 2:14 PM Post #528 of 1,366
I am sufficiently impressed with the DirectStream units, the Junior in particular (once you take value into account), that I am very seriously considering pulling the trigger on a DSD Junior regardless of what else I may decide to purchase when this process is over.
 
That would give me an additional DAC now, which I sorely need. And it would be easy to re-purpose the DAC as another Roon endpoint in my speaker system once I made a final choice (if that was something different).
 
The price is hard to argue with, since I have more than one suitable, unused, component that I can use to get the trade-in price ... so it would be $2,999.  And the upgradability aspect is extremely compelling on top of that. 
 
It's only the desire to test my Yggdrasil with the RedNet 3 network interface, which I still haven't gotten around to setting up, and my pending travel, which has me holding back on this.  Though, rest assured, even if I do pull the trigger on an interim DAC, I will still be completing my evaluations and making a final decision at that point.
 
Jul 18, 2016 at 4:16 PM Post #529 of 1,366

Thank you @Torq
regular_smile .gif

 
I am very much looking forward to your opinion on Auralic Aries vs. RedNet 3 vs. Network Bridge II vs. Sonicorbiter SE (optical output) or anything else you would consider to be of interest, please. 
wink_face.gif

Just starting this very exciting journey into the digital world, I am not sure why would there be any significant differences between the units… 
confused_face(1).gif


 
Jul 18, 2016 at 4:26 PM Post #530 of 1,366
 
Thank you @Torq
regular_smile .gif

 
I am very much looking forward to your opinion on Auralic Aries vs. RedNet 3 vs. Network Bridge II vs. Sonicorbiter SE (optical output) or anything else you would consider to be of interest, please. 
wink_face.gif

Just starting this very exciting journey into the digital world, I am not sure why would there be any significant differences between the units… 
confused_face(1).gif

 
Oh, there are many possibilities for why those units might perform differently.  That doesn't mean that they will, just that there are ample avenues for variation.  For example, on the hardware side, the quality of their power-supplies/regulation, how much noise they generate and how it affects their internals, the quality/accuracy/stability of their clocking implementations, presence of proper galvanic isolation for non-optical connection).
 
On the software/protocol side of things, there are many possible variables as to how the software does what it does, for example, RAAT (Roon's transport protocol), Dante, AES67 ("Ravenna") and OpenHome/DLNA/UPnP protocols are all very different in terms of how they manage the data stream and the clock.
 
In addition to the units you've mentioned, I'll be including the Linn DSM units as Ethernet interfaces.  While they do a LOT more than that, and are very expensive compared to dedicated network interfaces, they are perfectly capable of acting as a network audio interface, and I've used mine with my Yggdrasil in that fashion quite frequently.
 
Jul 18, 2016 at 4:35 PM Post #531 of 1,366
 
 
Oh, there are many possibilities for why those units might perform differently.  That doesn't mean that they will, just that there are ample avenues for variation.  For example, on the hardware side, the quality of their power-supplies/regulation, how much noise they generate and how it affects their internals, the quality/accuracy/stability of their clocking implementations, presence of proper galvanic isolation for non-optical connection).
 

I was pondering just this these days... Does the power supply have any effect on optical transports? It is clear that for both USB and Coax using an LPS can help things, but I can't figure out if this is supposed to help in any way the streaming of bits via optical SPDIF...
 
Jul 18, 2016 at 5:07 PM Post #532 of 1,366
  I was pondering just this these days... Does the power supply have any effect on optical transports? It is clear that for both USB and Coax using an LPS can help things, but I can't figure out if this is supposed to help in any way the streaming of bits via optical SPDIF...

Before the bits are sent by optical, they have to be clocked. If the clock is getting a dirty power supply then it can introduce jitter into the signal. The advantage of optical over the others is that it has intrinsic galvanic isolation, but the power supply is still important.
 
Jul 18, 2016 at 5:43 PM Post #533 of 1,366
Thank you @Torq
normal_smile .gif

 
 
Are there any beginner level tutorials to learn the software/protocol side of things you could please recommend? 
 
For example, I am trying to understand what it would take to setup, under Mac OS, a system with the main objective to play music stored on the local network with possible integration with streaming services (e.g. Spotify, Qobuz, Tidal…). 
confused_face(1).gif

 
So far, I see it unfolding like this:
OpenHome Player
Linn Kazoo control app
Linn Kazoo Server
 
O, perhaps:
JRiver
 
To understand if I am on target here, and what configuration is actually better, and why, I definitely need some education… 
normal_smile .gif

 
Will it be possible for you to briefly describe the software/protocol side of the system employed in your tests, please? 
wink_face.gif

 
 
 

 
Jul 18, 2016 at 5:46 PM Post #534 of 1,366
  I was pondering just this these days... Does the power supply have any effect on optical transports? It is clear that for both USB and Coax using an LPS can help things, but I can't figure out if this is supposed to help in any way the streaming of bits via optical SPDIF...

 
 
  Before the bits are sent by optical, they have to be clocked. If the clock is getting a dirty power supply then it can introduce jitter into the signal. The advantage of optical over the others is that it has intrinsic galvanic isolation, but the power supply is still important.

 
Yep, that's the fundamental cause of potential issues where the PSU and clock-quality matter for optical connections.  With optical, transmitted noise isn't a factor, but timing could be.
 
The actual affect the PSU, or downstream power-regulation, can have depends on the precise implementation of the interface.  For example, using a VCO (or VCXO), if you don't have a stable voltage you'll get timing errors right there.  The effects would be extremely small, maybe non-audible, but it's hard to say.
 
However, more importantly in this case, and the principal reason I mentioned PSUs and clocks, was that when I do the comparisons for the various network interfaces I won't be limiting my evaluations to optical connections.  My preference is currently AES/EBU (provided it's done properly) as the interface between the Ethernet interface and the DAC.  So I'll likely evaluate all of the output options for each interface just to see how they stack up.
 
Jul 18, 2016 at 5:57 PM Post #535 of 1,366
  Thank you @Torq
normal_smile .gif

 
 
Are there any beginner level tutorials to learn the software/protocol side of things you could please recommend? 
 
For example, I am trying to understand what it would take to setup, under Mac OS, a system with the main objective to play music stored on the local network with possible integration with streaming services (e.g. Spotify, Qobuz, Tidal…). 
confused_face(1).gif

 
So far, I see it unfolding like this:
OpenHome Player
Linn Kazoo control app
Linn Kazoo Server
 
O, perhaps:
JRiver
 
To understand if I am on target here, and what configuration is actually better, and why, I definitely need some education… 
normal_smile .gif

 
Will it be possible for you to briefly describe the software/protocol side of the system employed in your tests, please? 
wink_face.gif

 
 
 

 
I'll come back a bit later today and talk about the possible options/configurations to achieve what you're asking (mostly because I have to run in a minute or three), but ahead of that I do have just enough time to describe what I was using, at least where possible.
 
In general, (and unless stated otherwise) my tests were performed using an Auralic Aries as an ethernet (or WiFi) to AES/EBU interface.  That was fed via Roon, which allows me to feed the Aries bit-perfect data either from my local CD rips (the vast majority of my collection) and high-res downloads, as well as seamlessly include TIDAL content, since Roon access both your local library and TIDAL at once.
 
In some cases, when things were very close, I fiddled around with using a Lifatec optical cable from my Mac Pro direct into the DAC.  In that case I would use Roon by default and then either the latest Audirvana or HQPlayer if I wanted to fiddle around.
 
If a native network interface was available (which will need some more discussion) then I used that.  That would be an OpenHome or DLNA setup for my Linn gear (Kazoo or Asset UPnP) and anything else compatible with those standards (e.g. the Hegel HD 30) or the native interfaces for the PS Audio and AES67 protocol for the Merging units.
 
All posted impressions can, however, be assumed to be using Roon->Aries->AES/EBU unless specifically stated otherwise.
 
I'll be back later to talk about how you set this stuff up in a few different scenarios.
 
Jul 18, 2016 at 6:06 PM Post #536 of 1,366
Being an owner of PSA's PWD I–>II and having gone thru the s/w and f/w update process, a few times, let me say that this ability to upgrade these units can be a wonderful thing, depending.
 
And granted this was a few years back and hopefully PSA has improved their programming and testing cycles, but sometimes the upgrades can be a mixed bag.
As in the upgrades were better in some ways and worse in others which was typical for some of their upgrades.
 
As for their bridge II, I hope it is all sorted out by now, and is a 'better' solution, but in some situations (probably not yours what with Roon involvement) the ethernet via the bridge can be somewhat limiting in that you need an ethernet capable player.  The Jriver ethernet player has limited functionality and there don't seem to be many 'full featured' ethernet players available.
 
The RedNet system, by using the ASIO drivers solves this 'problem', but then the need for the Bridge card becomes mute because the RedNet boxes convert straight into a suitable digital audio format, except that I2S isn't one of them.
 
Complications, iterations, combinations, abound…
atsmile.gif

 
JJ
 
Jul 18, 2016 at 6:53 PM Post #537 of 1,366
Quote:
  Being an owner of PSA's PWD I–>II and having gone thru the s/w and f/w update process, a few times, let me say that this ability to upgrade these units can be a wonderful thing, depending.
 
And granted this was a few years back and hopefully PSA has improved their programming and testing cycles, but sometimes the upgrades can be a mixed bag.
As in the upgrades were better in some ways and worse in others which was typical for some of their upgrades.
 
As for their bridge II, I hope it is all sorted out by now, and is a 'better' solution, but in some situations (probably not yours what with Roon involvement) the ethernet via the bridge can be somewhat limiting in that you need an ethernet capable player.  The Jriver ethernet player has limited functionality and there don't seem to be many 'full featured' ethernet players available.
 
The RedNet system, by using the ASIO drivers solves this 'problem', but then the need for the Bridge card becomes mute because the RedNet boxes convert straight into a suitable digital audio format, except that I2S isn't one of them.
 
Complications, iterations, combinations, abound…
atsmile.gif

 
JJ


Any examples of "full featured" Ethernet players you could please provide? 
confused_face(1).gif

Could you please share your experience working with RedNet 3 Ethernet to Digital Audio: how does it SOUND (referring to Focusrite’s philosophy that ‘sound is everything’)? 
regular_smile .gif

I must confess I miserably failed trying to understand even some basics about this Ethernet-networked audio interface: for example, where the hell AES/EBU inputs/outputs are, what ADAT inputs/outputs are and how many of them there are: 32 channels or 16 channels, etc. 
frown.gif


 
Jul 18, 2016 at 7:12 PM Post #538 of 1,366
  Being an owner of PSA's PWD I–>II and having gone thru the s/w and f/w update process, a few times, let me say that this ability to upgrade these units can be a wonderful thing, depending.
 
And granted this was a few years back and hopefully PSA has improved their programming and testing cycles, but sometimes the upgrades can be a mixed bag.
As in the upgrades were better in some ways and worse in others which was typical for some of their upgrades.
 
As for their bridge II, I hope it is all sorted out by now, and is a 'better' solution, but in some situations (probably not yours what with Roon involvement) the ethernet via the bridge can be somewhat limiting in that you need an ethernet capable player.  The Jriver ethernet player has limited functionality and there don't seem to be many 'full featured' ethernet players available.
 
The RedNet system, by using the ASIO drivers solves this 'problem', but then the need for the Bridge card becomes mute because the RedNet boxes convert straight into a suitable digital audio format, except that I2S isn't one of them.
 
Complications, iterations, combinations, abound…
atsmile.gif

 
JJ

 
I've done quite a bit of research on the DirectStream and it seems that for the latest versions, Yale and Torreys, that there is almost universal praise for the changes. There are some people who liked certain versions of the beta releases a bit better than the final version but I believe they generally still prefer the new final version over the old final version.
 
On the earlier releases, I think there were more negative comments such as some people finding the Pikes Peak upgrade slightly bright. Heavy tweakers can go ahead and try the older versions, which are all posted on their website, to suit their preferences.
 
I agree network bridge is limiting for anyone that wants to use the system for non-music purposes like watching movies in a home theater. I think Rednet will also provide superior sound quality based on the information in the XU208 and Rednet threads. Paul McGowan posted something to the effect of "The network bridge is similar in SQ to using USB with the LANRover", and the LANRover is based on the Startech gigabit LAN tweak. @rb2013 thought that USB stack even with the startech was far inferior to the Rednet.
 
Jul 18, 2016 at 8:51 PM Post #539 of 1,366
  Being an owner of PSA's PWD I–>II and having gone thru the s/w and f/w update process, a few times, let me say that this ability to upgrade these units can be a wonderful thing, depending.
 
And granted this was a few years back and hopefully PSA has improved their programming and testing cycles, but sometimes the upgrades can be a mixed bag.
As in the upgrades were better in some ways and worse in others which was typical for some of their upgrades.
 
As for their bridge II, I hope it is all sorted out by now, and is a 'better' solution, but in some situations (probably not yours what with Roon involvement) the ethernet via the bridge can be somewhat limiting in that you need an ethernet capable player.  The Jriver ethernet player has limited functionality and there don't seem to be many 'full featured' ethernet players available.
 
The RedNet system, by using the ASIO drivers solves this 'problem', but then the need for the Bridge card becomes mute because the RedNet boxes convert straight into a suitable digital audio format, except that I2S isn't one of them.
 
Complications, iterations, combinations, abound…
atsmile.gif

 
JJ


There are definitely additional considerations that come into play when you're hardware is field-upgradeable via software/firmware/OS/FPGA updates.  I know with the Linn systems, which I've been using since they came out, not every update was a step forward.  Some were technically better, but killed the musicality.  Some were just not that great.  But, overall, they've tended to be an improvement overall and they got better and better over time, with fewer releases that wound up getting a "thumbs down" from the Linn DS community.
 
The Linn units also have a "fallback" recovery mode, so that if an update/upgrade failed for any reason, you could get back to an operational state and then either try again, or install an earlier version of the software.  This put your system back to working condition and allowed you to experiment with newer software versions, while being able to get back to a prior version if you found you preferred it.  Though, to be honest, most of the time a release wasn't met with mass approval it tended to be very short-lived and replaced in pretty short order.
 
I've only done one upgrade cycle with the PS Audio units, and that was simple and straightforward (not quite as simple as the Linn stuff, but certainly something any normal user could easily handle). At the same time it gives me no insight into how upgrades might go over a longer period ... one hopes they'll always be positive, but I expect, like everything, there will be problems at times and not every update will be universally approved from a sound quality perspective.  But, again, it's easy enough to go back to an earlier version of the software.
 
I'll talk about how to communicate with these units in the next post, where I'll also talk a bit about overall configuration and protocols and so on.
 
Jul 18, 2016 at 9:01 PM Post #540 of 1,366
 
Any examples of "full featured" Ethernet players you could please provide? 
confused_face(1).gif

Could you please share your experience working with RedNet 3 Ethernet to Digital Audio: how does it SOUND (referring to Focusrite’s philosophy that ‘sound is everything’)? 
regular_smile%20.gif

I must confess I miserably failed trying to understand even some basics about this Ethernet-networked audio interface: for example, where the hell AES/EBU inputs/outputs are, what ADAT inputs/outputs are and how many of them there are: 32 channels or 16 channels, etc. 
frown.gif


The AES/EBU outputs on the RedNet units are via the DB25 connector.  This is pretty common in pro/studio setups, as it provides a much higher density connector for when you're running multiple channels (which you pretty much always are).  You can buy standard cables that'll give you up to 8 channels from that connector that'll terminate in a standard 110 ohm XLR connection.  These tend to be pro-audio focused and not necessarily paying attention to the things that audiophiles expect (and most audiophiles don't expect to be able to get 8 XLR connections on a cable for $60 or so ... an would probably be suspicious of it if they could!).
 
Don't confuse AES/EBU (a specific version of the AES protocol, namely AES3) with the physical connections used to carry it.  There are three standard physical implementations for AES/EBU define in the basic spec.  These are XLR, RCA-COAX and Optical Fiber.  There are some lesser used connection types too, but those are the three we care most about for home audio.
 
Sound ... I'll have to defer to @johnjen since I don't have my unit set up yet.
 

Users who are viewing this thread

Back
Top