Streaming reliability and streamer updating

[AnotherJohnson]

After last night’s positive experience running the ND8006 stream through the 5101 DAC, it has occurred to me that I should not have been surprised.

The sound quality should have been as good as the 5101’s output from CDs or SACDs if file bit depth and sampling rate were equal. Why wouldn’t it be? The 5101’s DAC should not care where its input came from. Equivalent input should lead to comparable output, no?

[Wildcat]

It really is unfortunate as I would probably be a Roon subscriber if they had Amazon music integration. My household is in the Amazon ecosystem (echo's, fire TV, and that stupid Amazon van coming to my house twice a week!). I don't like Tidal due to their support of MQA, and Quboz would be a viable alternative but again, it would mean breaking from the convenience of the amazon system.

I'm with you on Tidal--I've only ever had free trial accounts with them, and never saw a need to join. Hopping into the fray with MQA didn't exactly endear them to me either.

I know that the Qobuz app can cast to some devices and also works over DLNA, but as you say, it won't work within Amazon's ecosystem. In my case, the decision for Roon and Qobuz were separate. It's nice that Roon works with my Chromecast devices, and a handful of others, but it wasn't the main reason I bought it--I wanted a better quality front end to the lossless music on my server. Qobuz I'd be happy with separately, but it was nice to see how deeply it integrated into Roon and my own library. With one of my Tidal trials, it co-existed nicely with both my library and Qobuz. When looking at an album release, you click on "Versions" and it shows you the different versions in your own library, along with versions from either/both streaming services.

[AnotherJohnson]

After buying a Bluesound Node and feeding it via AQ Diamond digital coax to the DAC in the Levinson 5101, I experimented for a couple of weeks. Primarily have settled on QoBuz as preferred over Amazon.

The Node is great. BluOS is decent. But after a lot of variable sound quality, testing shows that Comcast is not reliably delivering anything close to the 400 Mbps service that they charge us for. The problems are on their side of our gateway.

So … with streaming, there are some variables over which you have no control.

Glad to be running with a $500 Node and not a $20k+ high end digital source. Garbage in is garbage.

[admin]

Glad to be running with a $500 Node and not a $20k+ high end digital source. Garbage in is garbage.

I think we need to come to the conclusion that although the "dip" into the streaming audio was fun and interesting, it is not going to replace your love and preference for vinyl. One of the reasons I did suggest the Node was that the buy-in price was minimal. You can sell that node tomorrow on ebay for $400.

It is hard to say what was behind the less than ideal performance. Was it the poor internet connection, not high enough level streamer or DACs in the downstream components, or simply a personal preference to the sound of vinyl? Most likely a combination of more than one of these things. At the end of the day I don't think this was a waste of time or money. It was a journey and we gain experience and knowledge from these endeavors.

[AnotherJohnson]

It is hard to say what was behind the less than ideal performance. Was it the poor internet connection, not high enough level streamer or DACs in the downstream components, or simply a personal preference to the sound of vinyl? Most likely a combination of more than one of these things.

I don’t think it’s hard to say at all.

Other sources are continuously good.

Streaming is all over the lot … from total failure to excellent.

Here’s the data. Garbage in a significant fraction of the time, but variable.

CD34E578-5FBC-41BB-8752-73F0AE380D44.png (696.86 KiB) Viewed 1955 times

Look at the latency and jitter numbers and think about how this would affect D to A reconstruction. Note that the photo is truncated at the bottom. Look at the scale on the left and note peak latency. 695 msec is at the human reaction time level. It would clearly be a challenge to recreate music.

This screen shot was a horrible mishmash of chaotic data transfer.

[AnotherJohnson]

8562D863-AF5F-467D-9518-16A34738F588.png (1 MiB) Viewed 1946 times

02E2E87D-340C-467F-BCA9-B48BA40D1012.png (592.42 KiB) Viewed 1946 times

Here are two more … nominal 400 Mbps from Comcast.

[Joe Appierto]

Please excuse the crudeness but that result sucks.

I wonder if Comcast would notice if you sent in your payments late 31.3% of the time?

[admin]

What are you complaining about, can't you see Comcast telling you that you have "Solid Performance?" What more do you want?

[Joe Appierto]

I don't know how to embed a local image so I'll have to just report the numbers. Sorry.

At the moment I'm using JRiver to stream files to the PS Audio DAC. The playlist contains files from 44.1 to DSD64.

At the same time Packet Loss was run using the Georgia server and these are the results:

Sent: 149/150 99%
Time: 10 sec
Received: 149/150 99%
Total Packet Loss: 0/149 0.0%
Average Latency: 39.19 ms
Average Jitter: 2.51 ms

Edit:
Fast.com: 370 Mbps

[Wildcat]

That much packet loss, high latency, jitter, etc., is due to using WiFi. It's also good to test WiFi, but to see what the provider is contributing, without the interference of WiFi, you would need to do the test from a computer connected via Ethernet. (I check my connection weekly by using https://speedtest.net .) In addition, if the test is done from a specific device (phone, tablet, laptop), it also does not reflect what a different device might return for results--a streaming box with a poor WiFi antenna built in could get worse results, or its built-in antenna could be better than what is found in a phone and get better results.

I always run streaming, whether it's audio or video, over Ethernet, for that reason--it's reliable and consistent. On the common 2.4GHz band, there is too much interference from other, older devices in the house, plus over a dozen neighbors with weak signals, all competing for the same handful of channels. 5.0GHz has less congestion but due to the higher frequency, sometimes it is not as strong depending on where the equipment is located. WiFi is convenient but I don't always find it to be the optimum connection--for casual use it's fine, but I won't use it for audio or video if I want to ensure full quality.

It might not explain all your streaming issues, but if you tried Ethernet and the results changed (or not), that might help find out where the problem is. In my systems here, the main audio system and my desktop system are both running from Ethernet, so the streaming quality has been consistent from Qobuz. At worst, if there's any issue with the network speed, Roon simply refuses to load it, and will advance to the next track (it seems as though Roon buffers the incoming data for a short time).

I ran that test myself on the computer (over Ethernet) just now, default settings, and got all zeros. I also changed the default settings to give it more of a torture test (100 pings/second is way more than any streaming device would use, and a packet size around 480 bytes is about an average size for an audio data packet):

packet-loss-2.jpg (87.58 KiB) Viewed 1939 times

I also ran the test using my cheap Kindle Fire tablet (which I use as a remote for Roon, lighting, etc.), and had 7.3% late packets, with all the rest at zero. But, I'm sitting right behind the room that has the access point. Stepping out the side door into the driveway, the results are worse. Upload packet loss was 0.3%; total packet loss 5.7%; download packet loss 5.5%; late packets 15.3%. Average latency was 49.92ms and jitter 21.93ms.

Unless I had a bad Ethernet cable or connection on my network, it would be rare to ever see packet loss on my wired network. But wireless (WiFi) uses the same TCP/IP protocol, so a lot of the packets can be dropped to/from a wireless device, due to interference issues. If I tested my AT&T connection with Ethernet and had a high packet loss, I would have to eliminate any source of problems inside the house (including the modem/gateway) before I would suspect the AT&T connecting having a problem.

The jitter value is misleading--it won't directly relate to audio jitter, based on how the TCP/IP protocol works. A very low ping speed (latency) is more important for gaming, but a good average latency of 50ms is fine for most purposes. (A ping is a packet of data sent over the network to a device on the other end; the time it takes for you to send the packet and receive the "echo" reply from the remote device is your ping time.)

Without getting too technical, TCP/IP is self-healing, to a point. The device receiving the packet will send a signal back to the sending device whether it received its packet intact. If not, the sending device will transmit it again. But there is a limit to the number of times it will retransmit the data, at which point the packet is dropped. In essence, then, the receiving device receives all the packets and processes them in order before data is sent further along. Almost like a buffer, in other words.

This could be kind of technical, but it gives an overview of how the TCP/IP protocol works and how errors are handled:

https://www.dynatrace.com/news/blog/det ... -services/

[AnotherJohnson]

Regardless, it has to be reconstructed to analog, and 3/4 second late packets with 1/4 second jitter is problematic.
As I’d a sub 3% delivery of paid for speed.

[AnotherJohnson]

I literally just took a call from a guy with a heavy accent who says they found the problem. There was apparently something specific that they have sorted out on their side. FWIW, it is working perfectly… although since the problem has been variable, working now doesn’t guarantee working later.

He’s going to call again tonight to check and make sure it’s still on track. If it’s not, a technician will come out to my site tomorrow.

FWIW, this seems responsive to me.

I’m streaming a familiar program at 24/96. It is sounding very good… including the bass.

Maybe the variability can be tamed.

FWIW, when I went over to XFinity/Comcast this morning, they were unable to get remote diagnostic info from my site … so they made tomorrow’s appointment. I guess they kept working on it.

If I can keep 400 at the router, I can keep 150 -200 at the system. And that’s more than enough for better than highly adequate performance.

[AnotherJohnson]

Another For What It’s Worth …

The native English speaking young fella who helped me at the store was actually empathetic beyond the normal modern customer care template.

He said he listened to lossless hi-rez music too, and that he could hear differences as speeds and loads changed too.

Their appointment scheduling algorithm would not let him make a “truck roll” commitment, but by the time his supervisor came to over ride it, it had allowed the appointment on its own.

I’m streaming some over processed 24/192 Christmas leftovers right now. Best I’ve heard them.

I like 24 bit depth. But 192 kHz seems to my ears to introduce artifacts that are annoying in the long term.

YMMV

[admin]

That's a pleasant surprise. I'm glad it looks like it's being taken care of. Beside the music streaming side, you deserve the internet speed and reliability that you are paying for.

[AnotherJohnson]

That's a pleasant surprise. I'm glad it looks like it's being taken care of. Beside the music streaming side, you deserve the internet speed and reliability that you are paying for.

Yes, we’ll see how it goes. But they took an aggressive approach by not waiting for the tech appointment tomorrow.

With the speed and packet loss tools, I have a way to quantify my issues. They really took notice of the sub 10 Mbps screen shots at the router. The variability has been a big deal … 10 could become 600 ten minutes later.

I don’t know what they fixed, but they found something and they think they fixed it.

The 24/96 stream I’m listening to right now is so nice …. Literally - Diana Krall’s “So Nice.”

[AnotherJohnson]

Just got an automated call from Comcast explaining that they found and fixed the problem on their side of my router.

So they’re definitely staying with it. I can’t help but give them positive points for their attention to my problem.

My streaming today continues to be the best I’ve heard during these trials.

[admin]

Just got an automated call from Comcast explaining that they found and fixed the problem on their side of my router.

So they’re definitely staying with it. I can’t help but give them positive points for their attention to my problem.

My streaming today continues to be the best I’ve heard during these trials.

Glad to hear. "Unexpected" considering it's comcast but I'm glad they got on top of it so quick.

It's nice that at least you can give the streaming a fair shake without internet problems getting in the way. Even if you decide not to keep the streaming component in the main system, at least you can see what your current hardware is capable of with reliable internet.

It will also be interesting to see if the new speakers (which should be arriving soon) will augment the positives of digital streaming,... or highlight its deficiencies.

[AnotherJohnson]

I’m streaming Madonna’s True Blue at 24/192 right now.

It is quite good. Better by a large margin compared to my previous trials with this program.

I’ve had one brief glitch that could be WiFi related as WiFi speed did drop back to 40ish briefly. It quickly returned to the 150 range.

There are instruments and tones on this highly produced (some would say over produced) multilayered recording that I’ve keyed into for years in comparisons, especially on the cut “live to tell.”

This album gives the bass a good workout too, and it is much more natural than it’s been in prior streaming tests here.

[AnotherJohnson]

Now that the speed at the router may be stable, or at least within reasonable expectations, I’ve added the extender/bridge to the listening room and have set the Node on the extended network.

Speed has improved at the Node. Packet loss is a bit higher than at the router, but does not seem audible.

The midrange on True Blue (24/192) has improved.

I will try a hardwire connection to the bridge later. Right now I’ve got to get dinner going for this evening’s weekly Wednesday night family dinner.

Maybe we’ve made some progress today.

[AnotherJohnson]

It will also be interesting to see if the new speakers (which should be arriving soon) will augment the positives of digital streaming,... or highlight its deficiencies.

The new speakers will probably not get drop shipped to my job site. Apparently Wilson frowns on that. The speakers will go to the dealer who will have the responsibility for final delivery and set up. I guess Wilson is taking no chances. The dealer is the next link in the chain of custody.

They’re on track to probably ship to the dealer on some date after Friday.