Jump to content
IGNORED

HQPlayer Linux Desktop and HQplayer embedded


ted_b

Recommended Posts

3 hours ago, arglebargle said:

I would love for HQPlayer to handle volume leveling. I know its maybe not high on @Miska list being a big album listener but I think it would be a very popular feature. Especially if the volume target could be passed on by Roon. The sound quality of Roon's volume leveling leaves something to be desired. I don't use it for that reason. 

 

Since Roon already controls HQPlayer volume they could also do this control without changes to HQPlayer side. When Roon is used as playback source it is not possible to implement such solely at HQPlayer side.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
4 hours ago, audiofool said:

hqplayer minor Issue:
The Desktop settings shows channel_offset grayed out for NAA and it is not available at all on the embedded config page.
I inserted channel_offset="2" in the network address line of the xml files on both Desktop and Embedded versions and it works!

 

It is not available because it only works with most recent versions of both HQPlayer and NAA. There are still number of older NAA software versions out there where it doesn't work.

 

4 hours ago, audiofool said:

I have ebur128 average album volume tags in my music files.  What about an option to adjust the hqplayer volume control automatically at the start of the song relative to the ebur128 tag.  There would need to be a reference volume setting.  This would have the same effect as me physically changing the volume at the start of an album.

 

I don't know if there is any standard for such volume tags. At least I have not seen it in the metadata specifications. When HQPlayer's library is used for playback, it is technically doable. When source is for example Roon it is not possible unless Roon is the one adjusting the volume at a right moment.

 

4 hours ago, audiofool said:

I suspect the technical issue would be how to seamlessly change volume at the start of a song and avoid any clicks or gapless playback issues.

 

That is not really much of a problem, volume changes are already ramped using a smooth curve.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
14 hours ago, arglebargle said:

I would love for HQPlayer to handle volume leveling.

 

I'm using exact this with Roon. Be sure to set "Comfort Limit" to avoid loud noisiness. It's a very good sound!

 

38191107bb.png

Link to comment
On 3/31/2020 at 3:24 PM, Miska said:

 

I don't know if there is any standard for such volume tags. At least I have not seen it in the metadata specifications. When HQPlayer's library is used for playback, it is technically doable. When source is for example Roon it is not possible unless Roon is the one adjusting the volume at a right moment.

 

 

Thanks Miska,

I think REPLAYGAIN_ALBUM_GAIN tag is used by some software, probably not a standard though.  No idea what Roon is doing, maybe they are doing the calculation in real time by scanning the file?

Link to comment
49 minutes ago, audiofool said:

I think REPLAYGAIN_ALBUM_GAIN tag is used by some software, probably not a standard though.  No idea what Roon is doing, maybe they are doing the calculation in real time by scanning the file?

 

Roon calculates those gains by analyzing the content in background and stores this information in it's own database.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
On 7/14/2019 at 11:21 AM, Miska said:

As a note that may be interesting to some, the HQPlayer OS image supports running HQPlayer + NAA in "multihomed" mode. This is done by running all the Ethernet interfaces in HQPlayer OS in bridged mode. IOW, they all participate in the same network.

 

@Miska Yesterday I bridged the Intel dual fiber NIC's two SFP ports on my HQPe server:

SFP NIC 1 (connected to the main router) is now bridged with SFP NIC2. The latter is connected to the NAA endpoint.

I like the SQ change!

However, I see that my HQPe has reverted back to the trial version as the fingerprint has changed.

Is this to be expected in the case of bridging?

I will not make any rash decisions as I am also expecting a JCAT net card FEMTO, and I am not sure yet whether I will end up mounting this in the server or the endpoint. But I may have to request an updated key before long.

 

 

audio system

 

Link to comment
47 minutes ago, bodiebill said:

 

@Miska Yesterday I bridged the Intel dual fiber NIC's two SFP ports on my HQPe server:

SFP NIC 1 (connected to the main router) is now bridged with SFP NIC2. The latter is connected to the NAA endpoint.

 

NAA should be connected to a switch instead... Bridging turns computer into a software switch and helps avoid problems of multi-homed configurations. But it shouldn't have any impact on sound though.

 

48 minutes ago, bodiebill said:

However, I see that my HQPe has reverted back to the trial version as the fingerprint has changed.

Is this to be expected in the case of bridging?

 

No, it shouldn't happen, only changes to actual hardware devices should change it.

 

Please check with "ifconfig" that the same actual hardware devices are present in both cases.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
1 hour ago, Miska said:

NAA should be connected to a switch instead... Bridging turns computer into a software switch and helps avoid problems of multi-homed configurations. But it shouldn't have any impact on sound though.

 

Not sure if I understand: after bridging the HQP server, NAA is connected to a switch, namely to the server operating as a switch. Correct? So this should be OK? Or do you still recommend an additional HW switch between the server and the endpoint?

 

Could not the perceived SQ improvement be the result of less devices between the server and the endpoint?

There was no doubt to my ears that SQ was better after bridging, however I always admit the possibility of placebo when we are getting into these small incremental so called improvements.

 

1 hour ago, Miska said:

No, it shouldn't happen, only changes to actual hardware devices should change it.

Please check with "ifconfig" that the same actual hardware devices are present in both cases.

 

Ah, then I know what happened. After bridging I also deactivated the on board LAN, so that was the HW change changing the fingerprint. Thanks Miska!

 

audio system

 

Link to comment
5 hours ago, bodiebill said:

Not sure if I understand: after bridging the HQP server, NAA is connected to a switch, namely to the server operating as a switch. Correct? So this should be OK? Or do you still recommend an additional HW switch between the server and the endpoint?

 

Yes, but software switch is not the same as a proper hardware one. I always recommend having just one network interface that connects to a central switch (to which your internet router is also connected).

 

5 hours ago, bodiebill said:

Could not the perceived SQ improvement be the result of less devices between the server and the endpoint?

 

How was the setup before bridging?

 

5 hours ago, bodiebill said:

Ah, then I know what happened. After bridging I also deactivated the on board LAN, so that was the HW change changing the fingerprint.

 

OK, just keep it enabled even if it is not used and the fingerprint should remain the same, bridging or not.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
46 minutes ago, Miska said:

How was the setup before bridging?

 

before

main modem router => CAT7 => 1st MikroTik SFP switch => fiber => Server SFP NIC 1;

1st MikroTik SFP switch => fiber => 2nd MikroTik SFP switch => CAT6 => NAA endpoint

 

after

main modem router => CAT7 => 1st MikroTik SFP switch => fiber => Server  SFP NIC 1 <BRIDGED WITH> SFP NIC2 => fiber => 2nd MikroTik SFP switch => CAT6 => NAA endpoint

 

And soon I will receive a Cisco WS-C2960G-24TC-L switch (SFP + RJ45) and the JCAT net card FEMTO to play with.

Where would you put the latter?

 

audio system

 

Link to comment
6 minutes ago, bodiebill said:

Before:

main modem router => CAT7 => 1st MikroTik SFP switch => fiber => Server SFP NIC 1;

1st MikroTik SFP switch => fiber => 2nd MikroTik SFP switch => CAT6 => NAA endpoint

 

After:

main modem router => CAT7 => 1st MikroTik SFP switch => fiber => Server  SFP NIC 1 <BRIDGED WITH> SFP NIC2 => fiber => 2nd MikroTik SFP switch => CAT6 => NAA endpoint

 

In this kind of setup I would go with a HP or Cisco switch that have two SFP slots (remember to check that 802.3x flow control is enabled for all ports in switch settings!). But OTOH, the server doesn't need to be behind optical.

 

Remember to check that your copper network cables don't have metal connector body that would be connected to a shield!

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
12 minutes ago, Miska said:

In this kind of setup I would go with a HP or Cisco switch that have two SFP slots (remember to check that 802.3x flow control is enabled for all ports in switch settings!). But OTOH, the server doesn't need to be behind optical.

 

I will receive a Cisco WS-C2960G-24TC-L switch soon. You suggest I put it instead of the 2nd MikroTik? 

 

12 minutes ago, Miska said:

Remember to check that your copper network cables don't have metal connector body that would be connected to a shield!

 

All CAT cables are unshielded, thanks to your caveat 🙂

Suppose I would only use copper between the server and the NAA: would it be a problem if the cable is 23 meters long? 

 

audio system

 

Link to comment
24 minutes ago, bodiebill said:

 

before

main modem router => CAT7 => 1st MikroTik SFP switch => fiber => Server SFP NIC 1;

1st MikroTik SFP switch => fiber => 2nd MikroTik SFP switch => CAT6 => NAA endpoint

 

after

main modem router => CAT7 => 1st MikroTik SFP switch => fiber => Server  SFP NIC 1 <BRIDGED WITH> SFP NIC2 => fiber => 2nd MikroTik SFP switch => CAT6 => NAA endpoint

 

It seems like you are using your second Microtik as an SFP ... that's fine. The "before" config is correct.

 

24 minutes ago, bodiebill said:

 

And soon I will receive a Cisco WS-C2960G-24TC-L switch (SFP + RJ45) and the JCAT net card FEMTO to play with.

Where would you put the latter?

 

If you are substituting different cards and switches, there is no "correct" config beyond what you started with, so you can place different devices in different places and the ones that produce music will be "correct"

Custom room treatments for headphone users.

Link to comment
55 minutes ago, jabbr said:

The "before" config is correct.

 

So that puts me before a difficult choice: doing what is 'correct', or what sounds 'better' to my ears... I will keep experimenting...

There are many advocates on AS for a bridged setup (the 'after' setup), but apparently you have a different view? Is this based on theory or perceived SQ, or both?

 

audio system

 

Link to comment
48 minutes ago, bodiebill said:

 

So that puts me before a difficult choice: doing what is 'correct', or what sounds 'better' to my ears... I will keep experimenting...

There are many advocates on AS for a bridged setup (the 'after' setup), but apparently you have a different view? Is this based on theory or perceived SQ, or both?

 

I have briefly tested various brands of fiberoptic NICs on various servers, all 10Gbe or greater (e.g. 25,40,100 Gbe) and cannot tell a bit of SQ difference. Nor can I tell an SQ difference among professional quality switches.

 

When I was testing endpoints 5 years ago I thought I heard an SQ improvement going from copper RJ45 ethernet to fiberoptic but that might have been bias. In any case I'm only using fiber and wi-fi at this point. For HQPlayer, wired is far more reliable at DSD512, and wifi works well at DSD256 and if the wind is blowing the right way in my house DSD512 -- I can usually pull 280 Mbs over WiFi to my NUC/NAA

Custom room treatments for headphone users.

Link to comment
33 minutes ago, jabbr said:

I have briefly tested various brands of fiberoptic NICs on various servers, all 10Gbe or greater (e.g. 25,40,100 Gbe) and cannot tell a bit of SQ difference. Nor can I tell an SQ difference among professional quality switches.

 

So you do not believe in the concept of an audiophile ethernet switch?

As for me, I still just don't know. Just trying to find out... 

 

Quote

When I was testing endpoints 5 years ago I thought I heard an SQ improvement going from copper RJ45 ethernet to fiberoptic but that might have been bias. In any case I'm only using fiber and wi-fi at this point. For HQPlayer, wired is far more reliable at DSD512, and wifi works well at DSD256 and if the wind is blowing the right way in my house DSD512 -- I can usually pull 280 Mbs over WiFi to my NUC/NAA

 

Could it have been the same bias made me embrace fiber optics some time ago?

I do know that I now prefer copper ethernet downstream, i.e. as the last connection before the endpoint. I even removed the fiber NIC from my NAA PC.

Never tried WiFi though.  

 

audio system

 

Link to comment
14 hours ago, bodiebill said:

So you do not believe in the concept of an audiophile ethernet switch?

 

Certainly not. I try to choose switches that have the features I need and try to stick to HPE and Cisco.

 

14 hours ago, bodiebill said:

Could it have been the same bias made me embrace fiber optics some time ago?

 

I don't know about bias. Fiber matters for the last link to the NAA. But if that is fiber, rest of the network infra on the other side doesn't matter since there's no noise conduction over fiber. It doesn't harm either, but from NAA's perspective it just looks the same, no difference.

 

It is just important to make sure all involved networking gear support things like 802.3x, 802.1p and 802.3az (last one has limited applicability to fiber). And that cables don't have shields that would connect device grounds together. Otherwise some negative effect somewhere may spoil the advantage gained elsewhere.

 

14 hours ago, bodiebill said:

I do know that I now prefer copper ethernet downstream, i.e. as the last connection before the endpoint. I even removed the fiber NIC from my NAA PC.

 

That is counter-intuitive, that's the place where fiber matters as it keeps the NAA fully isolated from rest of the network.

 

14 hours ago, bodiebill said:

Never tried WiFi though.

 

WiFi has same isolation effect as fiber, but with some added radiated RF noise. Whether that matters depends on implementation of NAA.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
1 hour ago, Miska said:

That is counter-intuitive, that's the place where fiber matters as it keeps the NAA fully isolated from rest of the network.

 

I understand. However I have the impression that FMC's and fiber NIC's are noisy and that could be worse than the isolation advantage.

 

My post was in reaction to @jabbr's post, but I appreciate your answers! Learning all the time...

 

audio system

 

Link to comment
18 minutes ago, bodiebill said:

 

I understand. However I have the impression that FMC's and fiber NIC's are noisy and that could be worse than the isolation advantage.

 

My post was in reaction to @jabbr's post, but I appreciate your answers! Learning all the time...

 

Hi, don't want to delve too far off topic here but I don't use FMCs. I use Mellanox which outsources to HPE and Dell. I have also used Intel and Solarflare NICs. They all work great. Newer Fiber NICs are very low noise and must meet extremely tight noise standards. 

Custom room treatments for headphone users.

Link to comment

@bodiebill ... and others ... Let me give my overall impressions:

 

1) The best thing you can do with HQPlayer is NAA which gives Ethernet isolation and a low powered quiet machine attached to the DAC.

2) I use fiberoptic Ethernet for isolation

3) I haven't yet fully evaluated the myriad of filters and algorithms including kernels for room correction and digital crossover -- I expect that these will have a much greater difference in SQ than various switches and because my brain has a limited ability to compare things that mostly sound very similar I am selective in what I compare. I got my Mellanox (100gbe) switch because I got a fantastic deal on it, and I got my prior Brocade (10Gbe) switch for the same reason.

4) I find that Wifi works reliably for me up to DSD256 where wired is reliable to DSD512, so I'm currently looking at whether ASDM7EC/DSD256 compares with AMSDM7/DSD512 that said, I think the dropouts/latency with Wifi is likely much more of a factor regarding SQ than perhaps EMI (my house has a *very* active Wifi situation going on right now with multiple kids doing home schooling/videoconferencing on laptops, also TVs which wifi netflix/amazon etc) 

 

Thoughts on ASDM7EC/DSD256 compared with AMSDM7 etc./DSD512?

Custom room treatments for headphone users.

Link to comment
4 hours ago, jabbr said:

@bodiebill ... and others ... Let me give my overall impressions:

 

4) I find that Wifi works reliably for me up to DSD256 where wired is reliable to DSD512, so I'm currently looking at whether ASDM7EC/DSD256 compares with AMSDM7/DSD512 that said, I think the dropouts/latency with Wifi is likely much more of a factor regarding SQ than perhaps EMI (my house has a *very* active Wifi situation going on right now with multiple kids doing home schooling/videoconferencing on laptops, also TVs which wifi netflix/amazon etc) 

 

I've tested WiFi up to 8 channels of DSD256 (to exaSound e28 DAC). But it naturally depends on type of the WiFi and how much other activity there is. Mine is two access points (PoE powered HPE stuff) 802.11ac with simultaneous 2 and 5 GHz bands and MIMO antenna. These are run from a Cisco switch with PoE ports.

 

4 hours ago, jabbr said:

Thoughts on ASDM7EC/DSD256 compared with AMSDM7 etc./DSD512?

 

Depends on the DAC, but it is likely that the EC modulators give best performance. If you are on ESS chip though, you may still want to consider running DSD512 instead, but you can compare.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
1 hour ago, Miska said:

I've tested WiFi up to 8 channels of DSD256 (to exaSound e28 DAC). But it naturally depends on type of the WiFi and how much other activity there is.

Sure but the e28 has no WiFi or ethernet connections.  What was connecting to the LAN?

Kal Rubinson

Senior Contributing Editor, Stereophile

 

Link to comment
21 minutes ago, Kal Rubinson said:

Sure but the e28 has no WiFi or ethernet connections.  What was connecting to the LAN?

 

Mac Mini running as HQPlayer NAA, connected to gigabit Ethernet. HQPlayer running on my Lenovo T470p laptop connected to the WiFi (running Ubuntu Desktop 18.04 LTS as OS).

 

On the way in network infra there are those HPE wireless access points, Cisco switch in central patchbay and then HPE switch in my listening room.

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...