The difference between Qoboz stream vs purchased songs

@fredrik222  That's good information; I know some of the early services did that, and I know spotify does this as well.  I was curious if you download a 192/24 flac that fast, so I did some back-of-the-envelope calculation.

From the article, 192 kHz/24-bit music they allocate 41.9 MB/minute.

Looking through my library, a typical FLAC at 192/24 is about 14 MB for a 5 minute song.

That means that they'll send that song to a user in roughly 3 seconds, assuming the user has sufficient bandwidth.

If you assume a typical home user has a 300 Mb which is (mega Bit) not accounting for the overhead of the IP connection setup that would get you the song in around 3 seconds.   So yes, it's very possible they download first.   I'll have to try that when I get home this evening.

As a pure guess, I think it's likely they're using SCTP because you can let the protocol handle the error correction and breaking the music into chunks.   And then they can start playing as soon as they receive the first chunk of data. This avoids any interpolation of data in the application, and you don't have dropouts waiting for data.

All of this works because music files are tiny, they take almost nothing to move.  It doesn't even stress your internal network.   

It's far more complicated to stream video because we're talking 2-3 orders of magnitude difference in size.

@cleeds So when Qobuz tells you that it's cached first, literally, what do you think is more likely, 1) Your experiment is flawed 2) Qobuz is lying. The answer is obvious.  

They are not the same to me, I know the difference in the technologies, and that is why this matters. You don't. You have demonstrated this beyond any doubt. 

And, even with Qobuz telling you that they cache first, you still refuse to admit that you were wrong. That is ridiculous. 

@tomrk  Here's a nugget for you:

” See an example of that in the attached graph from my Roon core playing 96/24 from Qobuz as a track ends and the next starts: a steady send rate to the playing endpoint, and a big burst receiving the next track. It’s all on 1Gb wired local network to the router, so that 140 Mbps burst is no problem.”

https://community.roonlabs.com/t/qobuz-192-not-working-despite-over-100-mbps/165568/8

Also from the article "I compared latency streaming 320 kbps as an even benchmark across services. Loading a new track in Qobuz takes on average 1,25 to 1,5 seconds. " Which is more than enough to download a 320 kbps MP3.

fredrik222

what do you think is more likely, 1) Your experiment is flawed 2) Qobuz is lying. The answer is obvious.

Logical fallacy: excluded middle/false choice.

Please feel free to conduct your own experiments. I’ve now done three, with three different streamers, with similar results. The experiment is so simple, it’s a bit of a puzzle as to why you resist, other than it allows you to keep the "argument" alive.

Of course, you’re also free to copy/paste here, rely on press releases and the like. Suit yourself.

@fredrik222  Yeah, that's why I showed the math.  I surprised myself when I realized that you could literally download a high-definition FLAC from Qobuz in just a few seconds easily.

I use Spotify most of the time since I drive a lot and like to listen to headphones during the day.  However, I use Qobuz when I'm listening on my NAD M33 which has the BluOS built in.  I stream from both my internal network and Quobuz, and you really can't tell the difference between the two in terms of latency or the quality.

And the difference between digital and analog is while you can still argue about the quality of the sound between components (speakers, DAC's etc), digital data in terms of content and timing of data packets can be measured and quantified in a way that gives you something that is directly comparable.