• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Streaming Services Are Comparatively Different, subjectively…

Headchef

Active Member
Joined
May 1, 2021
Messages
148
Likes
40
Location
In a jar, by the door
So I’m currently torturing myself to determine a simple thing, which streaming service gives the *best sounding* output.

objectively.

So I’ve been listening to Cachaito by Orlando Lopez as a reference album to compare the four services I currently use, all played at the same sample rates, exclusive mode when supported.

so far the sound reproduction seems to be split into two camps, and they are deter as to how they sound, the musical presentation through my system, now yours will no doubt be different and I’d be keen to understand how it does sound different to you…

so “as I hear it” this is how today’s listening has panned out…

Camp A

1. Qobuz
2. Apple Lossless

both soundstages around 2ft from the ground, width about 2-3ft either side of outside speaker edges, height 2-3 ft with clear central imaging, depth about 2-3ft from drivers, 1-2 ft behind. No discernible treble bleed or bass roll off.

Camp B

3. Tidal
4. Amazon

both soundstages around 1ft from the ground, width about 1-2ft either side of outside speaker edges, height 1-2ft with mostly clear central imaging, some blurring during complex passages, depth about 1-2ft from drivers, 1ft behind. Noticeable reduction in treble and bass resolution sometimes poor.

volumes checked with a constant 63db tone before playing to ensure equal output levels.

those numbers are ranked, Qobuz sounded the most realistic in reproduction with Apple very close behind, if a little over bright sounding in certain parts. Tidal and Amazon both gave very similar performances with Tidal did seem to have the edge over Amazon in the more demanding sections.

Anyone else care to compare them and share your thoughts? I have a couple of friends with professional recording studios, I might be able to borrow a few mics etc and actually record the levels at measured distances to see if it’s possible to determine if this is scientifically demonstrable.

it’s been, hot, perhaps too hot :D
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
14,316
Likes
23,047
What we need is someone to capture the digital stream. Easiest way to compare. Quite possible all are from different mastering as well. If you capture the digital stream then you can use software as well as listening comparisons.

I don't think recording with microphones will be good enough, but if you have them give it a try.
 
OP
Headchef

Headchef

Active Member
Joined
May 1, 2021
Messages
148
Likes
40
Location
In a jar, by the door
What we need is someone to capture the digital stream. Easiest way to compare. Quite possible all are from different mastering as well. If you capture the digital stream then you can use software as well as listening comparisons.

I don't think recording with microphones will be good enough, but if you have them give it a try.

There are other elements that could be at play too, Amazon I believe will alter the stream/bandwidth on the fly if the network capacity gets constrained. I’ve do doubt that various platforms also have their own DSP filters that go unmentioned.
 

escksu

Addicted to Fun and Learning
Joined
Jul 16, 2020
Messages
965
Likes
375
Hmm.... do you have anything ot back your claims? Some measurements? Also, what do you meant by bass resolution poor?
 

Beave

Addicted to Fun and Learning
Joined
May 10, 2020
Messages
614
Likes
1,123
Impressive how they can alter the height (both off the ground and from top to bottom) of the soundstages! :eek:

DSP sure has come a long way since I was in school.


:)))
 
OP
Headchef

Headchef

Active Member
Joined
May 1, 2021
Messages
148
Likes
40
Location
In a jar, by the door
Hmm.... do you have anything ot back your claims? Some measurements? Also, what do you meant by bass resolution poor?

well as someone that plays both guitar and bass I’m quite used to scrutinising the sound of my guitars coming out of my amps, especially if I’m modifying the tone. There’s a million and one ways that you can alter the tone, sometimes you might want to actually break up or corrupt the sound, other times you might want to make it as pure and accurate. Whilst doing that there’s a sweet spot between it sounding great or terrible and that sweet spot is usually where you still have the note I ringing true, with enough (let’s say for example) distortion, attack, delay, fuzz, reverb, etc so that it still give a definable change in pitch or tone amongst all the noise. That’s what I mean by bass resolution poor, it’s as though elements of the bass notes are lost in the reproduction, it sounds as though there’s a notable LF roll off.

As for the “where’s the measurement, I can’t see any measurements, youre not valid without a full suite of measurements“ read the damn post, Prior to conducting detailed measurements its customary to conduct what is known as a preliminary study and gather opinion. That’s what I’m doing now. Would you like me to draw you a graph?
 
OP
Headchef

Headchef

Active Member
Joined
May 1, 2021
Messages
148
Likes
40
Location
In a jar, by the door
Impressive how they can alter the height (both off the ground and from top to bottom) of the soundstages! :eek:

DSP sure has come a long way since I was in school.


:)))

Digital Signal Processing (aka DSP) is a continually developing field, and I’d imagine that it probably has come on a long way since you went to school on the assumption that you used that phrase, which I’d assume puts your year of birth somewhere between the 60’s-70’s? :D

theres this thing called Moores law which actually has been exceeded a few times, and as processing capabilities have improved so too has the capabilities of digital signal processing, it’s funny to wonder what Joe Meek might have made of Apples latest capabilities, which can now (on the fly) identify voice, instrument, sounds, etc and reposition them within the mix.

so yes, your point is quite right, unless you were at school yesterday, then no, it’s not changed all that much.
 

Beave

Addicted to Fun and Learning
Joined
May 10, 2020
Messages
614
Likes
1,123
You don't have to draw a graph, but it would be great if you could explain how two-channel audio can have differing vertical soundstaging.
 

Leporello

Active Member
Joined
Oct 27, 2019
Messages
220
Likes
488
So I’m currently torturing myself to determine a simple thing, which streaming service gives the *best sounding* output.

objectively.

So I’ve been listening to Cachaito by Orlando Lopez as a reference album to compare the four services I currently use, all played at the same sample rates, exclusive mode when supported.

so far the sound reproduction seems to be split into two camps, and they are deter as to how they sound, the musical presentation through my system, now yours will no doubt be different and I’d be keen to understand how it does sound different to you…

so “as I hear it” this is how today’s listening has panned out…

Camp A

1. Qobuz
2. Apple Lossless

both soundstages around 2ft from the ground, width about 2-3ft either side of outside speaker edges, height 2-3 ft with clear central imaging, depth about 2-3ft from drivers, 1-2 ft behind. No discernible treble bleed or bass roll off.

3. Tidal
4. Amazon

both soundstages around 1ft from the ground, width about 1-2ft either side of outside speaker edges, height 1-2ft with mostly clear central imaging, some blurring during complex passages, depth about 1-2ft from drivers, 1ft behind. Noticeable reduction in treble and bass resolution sometimes poor.

volumes checked with a constant 63db tone before playing to ensure equal output levels.

those numbers are ranked, Qobuz sounded the most realistic in reproduction with Apple very close behind, if a little over bright sounding in certain parts. Tidal and Amazon both gave very similar performances with Tidal did seem to have the edge over Amazon in the more demanding sections.

Anyone else care to compare them and share your thoughts? I have a couple of friends with professional recording studios, I might be able to borrow a few mics etc and actually record the levels at measured distances to see if it’s possible to determine if this is scientifically demonstrable.

it’s been, hot, perhaps too hot :D
How did you do level matching (to 0.1 db)?
 

bluefuzz

Addicted to Fun and Learning
Joined
Jan 17, 2020
Messages
607
Likes
986
which streaming service gives the *best sounding* output
Why not Spotify as well?

Some thoughts:

For the test to be meaningful you need to be sure all the tracks are from the same master and compare them to a CD version or FLAC rip of same played through the same signal chain. Assuming 'sounds best' in this context means closest to the original CD. It should also go without saying that tests need to be blinded too and probably use a larger selection of different music styles ...

I think it should be taken as given that there may be differences between different streaming services due to different compression algorithms, different transport methods, different masters, different watermarking schemes etc. But even if you can detect differences or even a clear preference in any one track that doesn't necessarily mean those differences will hold for the whole library, i.e. sometimes Apple will be 'better', sometimes Tidal etc. Then what?

Unless the differences are 'night and day' (in which case something is broken) then does it matter? When I had Tidal I informally compared Tidal, Spotify and CD-rips of a few tracks and couldn't hear any difference - nor did I expect to. Perhaps I could have detected a difference if I had done proper level matched, blinded tests but If the differences are that small I simply don't care.

I'm currently evaluating Apple Music lossless - not because I think it will sound any different to Spotify (it doesn't) - but I'm interested to see if it has better useabilty, library coverage or different discovery features which are the only meaningful differences between these services IMO.
 
OP
Headchef

Headchef

Active Member
Joined
May 1, 2021
Messages
148
Likes
40
Location
In a jar, by the door
How did you do level matching (to 0.1 db)?

quite a rough and ready method to be honest, when testing different components I play a continuous tone (the same note from the same digital key) and set the volume control so that each set up is reading an average of 61db. (I was supposed to be comparing DAC’s but got side tracked when I accidentally discovered that the same track sounded different through the same dac only to realise I’d accidentally played it on a different streaming transport) on an app called dbmeter (iPhone), the phone is sat (with the mic pointing towards the speakers) on a stand just in front of my listening position. Much to the better half’s annoyance I was using a track called Brass Monkey by The Beastie Boys, described by the wife as “possibly the most annoying thing you could have picked” to set an average equal volume when comparing the different streams.
 
OP
Headchef

Headchef

Active Member
Joined
May 1, 2021
Messages
148
Likes
40
Location
In a jar, by the door
Why not Spotify as well?

Some thoughts:

For the test to be meaningful you need to be sure all the tracks are from the same master and compare them to a CD version or FLAC rip of same played through the same signal chain. Assuming 'sounds best' in this context means closest to the original CD. It should also go without saying that tests need to be blinded too and probably use a larger selection of different music styles ...

I think it should be taken as given that there may be differences between different streaming services due to different compression algorithms, different transport methods, different masters, different watermarking schemes etc. But even if you can detect differences or even a clear preference in any one track that doesn't necessarily mean those differences will hold for the whole library, i.e. sometimes Apple will be 'better', sometimes Tidal etc. Then what?

Unless the differences are 'night and day' (in which case something is broken) then does it matter? When I had Tidal I informally compared Tidal, Spotify and CD-rips of a few tracks and couldn't hear any difference - nor did I expect to. Perhaps I could have detected a difference if I had done proper level matched, blinded tests but If the differences are that small I simply don't care.

I'm currently evaluating Apple Music lossless - not because I think it will sound any different to Spotify (it doesn't) - but I'm interested to see if it has better useabilty, library coverage or different discovery features which are the only meaningful differences between these services IMO.

I think I’ll add Spotify to this, I just didn’t have it on the MacBook but it’s easy enough :)

Essentially I just trying to find out which of the myriad of different services actually sound the most realistic through my system, it’s a purely subjective and somewhat selfish motive to be honest :)

Any you’re spot on with the reasons why they probably sound different, that’s my thinking too, and they are noticeably different. Play Brass Monkey by The Beastie Boys through a couple of different streaming services and you’ll see almost immediately. Bass & Treble vary quite widely.

This is purely a preliminary study to see if this is worth spending time evaluating, and I’m beginning to think that it is.

it bugs me that a service might be being billed as hires, master, lossless, etc but if that’s not actually the case the whole tower of cards falls down.

Try it for yourself, play Brass Monkey by The Beastie Boys through a could of different streaming services and tell me if they all sound similar (to the point it’s not really noticeable) or if you can see a marked difference in bass & treble reproduction …
 
OP
Headchef

Headchef

Active Member
Joined
May 1, 2021
Messages
148
Likes
40
Location
In a jar, by the door
You don't have to draw a graph, but it would be great if you could explain how two-channel audio can have differing vertical soundstaging.

ok, I can go one better than that, here’s a little exercise that I’d like you to try and report back your experience.

listen to the opening of Roger Waters Amused to Death Track 6 Late Home Tonight pt. 1

on both Apple Music and Tidal

and let me know what you hear, if there’s any difference to the imaging of the bird song.
 

kyle_neuron

Active Member
Joined
Jun 18, 2021
Messages
145
Likes
241
If you want to truly do this objectively, your best bet may be to find & use tools to ‘obtain’ the actual file from the services in question. You can find downloader tools, or could use something like SoundFlower or VoiceMeeter to digitally loopback to do a recording of each, then carefully extract the exact same clip. Studies show that 12 seconds is about the threshold for ABX comparisons, so it may be worth grabbing a few choice passages from your chosen tracks.

You can then put those into a folder and make a MUSHRA test using webMUSHRA, with randomised sample selection and a hidden anchor and reference. That way it’s double blind, at least.

As for Brass Monkey, I’m pretty sure there’s a bunch of masters of that. Tidal used to let you review which edition was being played but no longer seems to list that in the track credits.
 
OP
Headchef

Headchef

Active Member
Joined
May 1, 2021
Messages
148
Likes
40
Location
In a jar, by the door
If you want to truly do this objectively, your best bet may be to find & use tools to ‘obtain’ the actual file from the services in question. You can find downloader tools, or could use something like SoundFlower or VoiceMeeter to digitally loopback to do a recording of each, then carefully extract the exact same clip. Studies show that 12 seconds is about the threshold for ABX comparisons, so it may be worth grabbing a few choice passages from your chosen tracks.

You can then put those into a folder and make a MUSHRA test using webMUSHRA, with randomised sample selection and a hidden anchor and reference. That way it’s double blind, at least.

As for Brass Monkey, I’m pretty sure there’s a bunch of masters of that. Tidal used to let you review which edition was being played but no longer seems to list that in the track credits.

thanks, if this is to move forwards into a close analysis study that’s probably a good way to proceed. It does seem strange that we’re all comparing the equipment in such excellent detail but then in actual use folk are unaware that the source material they’re listening to seems to be being fed through some rather destructive or enhancing filters/DSP?
 

MarcosCh

Addicted to Fun and Learning
Joined
Apr 10, 2021
Messages
828
Likes
535
Hm, i bet only one digital mastering of that recording exists (no idea though), so good for this experiment. To me, personally, the real deal would be to see what sources each service uses for more popular music for which more than one remaster exists, but not that popular that the different remasters are so well known that they need to be disclosed (ie beatles and the like).
 

Goodman

Active Member
Joined
Jun 1, 2021
Messages
169
Likes
61
So I’m currently torturing myself to determine a simple thing, which streaming service gives the *best sounding* output.

objectively.

So I’ve been listening to Cachaito by Orlando Lopez as a reference album to compare the four services I currently use, all played at the same sample rates, exclusive mode when supported.

so far the sound reproduction seems to be split into two camps, and they are deter as to how they sound, the musical presentation through my system, now yours will no doubt be different and I’d be keen to understand how it does sound different to you…

so “as I hear it” this is how today’s listening has panned out…

Camp A

1. Qobuz
2. Apple Lossless

both soundstages around 2ft from the ground, width about 2-3ft either side of outside speaker edges, height 2-3 ft with clear central imaging, depth about 2-3ft from drivers, 1-2 ft behind. No discernible treble bleed or bass roll off.

Camp B

3. Tidal
4. Amazon

both soundstages around 1ft from the ground, width about 1-2ft either side of outside speaker edges, height 1-2ft with mostly clear central imaging, some blurring during complex passages, depth about 1-2ft from drivers, 1ft behind. Noticeable reduction in treble and bass resolution sometimes poor.

volumes checked with a constant 63db tone before playing to ensure equal output levels.

those numbers are ranked, Qobuz sounded the most realistic in reproduction with Apple very close behind, if a little over bright sounding in certain parts. Tidal and Amazon both gave very similar performances with Tidal did seem to have the edge over Amazon in the more demanding sections.

Anyone else care to compare them and share your thoughts? I have a couple of friends with professional recording studios, I might be able to borrow a few mics etc and actually record the levels at measured distances to see if it’s possible to determine if this is scientifically demonstrable.

it’s been, hot, perhaps too hot :D
 

Goodman

Active Member
Joined
Jun 1, 2021
Messages
169
Likes
61
Once Spotify goes FLAC (very soon), It's game over for all the above mentioned.
 

bluefuzz

Addicted to Fun and Learning
Joined
Jan 17, 2020
Messages
607
Likes
986
and they are noticeably different.
They may well be different but so what? They may be sourced from different masters or otherwise have different pedigrees. I don't think anyone knows for sure how the various streaming services source their tracks and in what format - or what happens to the tracks after the services get them. No doubt it's a constantly changing hodgepodge of formats and distribution methods depending on country, record company, streaming service or licensing deal.

Every service has undoubtedly some standardized 'master' format they use for distribution. There will be automated systems to convert tracks from one format to another, add metadata and watermarks etc. I think most of us know how challenging it can be just to maintain consistent metatagging and cover-art of a personal flac library. I imagine it's a bit more complicated to maintain a library of hundreds of millions of tracks sourced from god-knows-where in any consistent manner.

My point is whatever 'difference' you may find between streaming services in any one track or album will not necessarily be the same for any other album. And since every service is constantly changing features, updating licensing, formats etc. even if today one service sounds consistently 'better' for a sample of tracks it may not do next week. Or it may just be some upstream encoding error that is not representative at all ...

it bugs me that a service might be being billed as hires, master, lossless, etc but if that’s not actually the case the whole tower of cards falls down.

This is why I just stick with Spotify. They have the largest library, the best discovery, decent useability and competitive pricing. And of course sound just fine. Do you really expect perfection for $10 a month? If you want the guaranteed best unadulterated quality then buy the CD.
 
Top Bottom