Monday, July 18, 2011

Spotify Bitrategate, The Story So Far (Update: Problem Solved)

Important Update (23/09/2011): Spotify's Official Statement: "All music streamed through Spotify is of high quality (no less than 96 kbps for mobile and 160 kbps for desktop). We have a catalog of more than 15 million tracks and more than 99.9% are available in high bitrate (320 kbps) for our Premium users. Our catalog adds an average of 10,000 new tracks daily and we add the newly added tracks as quickly as possible."

---------------------

“Premium members deserve premium sound quality.” Spotify.com

“I guess premium members also deserve to be ignored…” A customer on GetSatisfaction

Spotify introduced 320 kbps streaming back in June 2009. Since then, no official statement on the conversion progress has been made. The closest I can find is on their support forum, GetSatisfaction, a Spotify employee Andres said about 60% of their music were available in 320 kbps. That was from one year ago.

So, as a premium user, what would you expect now? 80%, 90% or all Spotify music is available in 320 kbps streaming?

I did a simple test: After restarting my Macbook, I used a freeware, Magican Paster Lite to monitor the internet traffic. I saved the tracks one by one to an offline playlist, and the downloaded data shows the file size of the tracks.

I tested five tracks each from Spotify’s new singles/albums, and currently most popular singles/albums, taken from the official playlists and Top Lists in Spotify's desktop client. The results are very disappointing. The majority (13 out of 20 tracks) of the newest and most popular stuff are only available in 160 kbps, even the latest high profile release, Beyoncé’s 4, which was a Spotify Premium exclusive pre-release, is not in HQ.

I raised this question through Quora, GetSatisfaction and Spotify’s contact forum, and got two emails from custom service:

“We aim to have all the world’s music available at Spotify in HQ. We are still in the process of acquiring licenses to all music in the world (!), therefore it is possible that you won’t be able to find some of your favorite artists or tracks in HQ right now. We are signing new labels and adding a great amount of new tracks every week. Hopefully your favorite music will be up and running in 320 kbps soon.”

After I asked him why they needed a different license for HQ streaming, he then replied:
 

"I am sorry I was unclear, the quality issue has not that much to do with actual licenses, but rather which copies the labels hand us with. The goal for us is to obtain ALL music in HQ, but at the moment not every record is available in multiple qualities."

I find this answer rather vague. A friend in the licensing business, who has licensed their contents to Spotify, told me that: all streaming services, including Spotify, get music from content providers in lossless format. He said what he did with Spotify is uploading their contents in Flac to a FTP, with a xml file for metadata. I believe that’s the way Spotify got most, if not all their content. Otherwise, they will have to ask the content providers for 96 kbps (for mobile low-quality streaming), 160 kbps, and 320 kbps ogg files, and 320 kbps MP3 files for Spotify’s download service. Besides, Spotify’s ogg files are encrypted, how can  the content providers do that?

I did another test on July 6th. I used http://www.randomsong.se/, and Spotibot each to generate 40 random tracks. The first one seems to be randomly selected from a rather mainstream catalog, and the Spotibot one looks really random. I then switched to my other PC, with Spotify freshly installed (to make sure no cache interferes the test), and saved the tracks one by one into a offline playlist. I used a freeware Net Meter to monitor the size of downloaded data.

Since the file size difference between 160 and 320 kbps ogg files are obvious, a 4 minutes song in 160 kbps is about 5 MB and and normally more than 10 MB in 320 kbps, I believe my results are accurate.

See the statistical results of my investigations in this spreadsheet on Google Docs:



Only 35 of the 115 tested tracks are in high quality.

It is worth noting that, more than 100 of the 115 tracks are already available for purchase as 320 kbps MP3s on Spotify. I don’t know how much time Spotify spent on launching the MP3 store, but it seems obtaining their catalog in 320 kbps is not something they cannot get done in two years.  I can give you a rough estimate: an ordinary dual-core computer can convert an 10-track, 40-minute album from Flac to q9 ogg (320 kbps) in two minutes (at most). That is 12 seconds for one track.



15,000,000*12/(60*60*24*365*2)=2.854

So it only takes 3 home computers to convert Spotify’s whole 15-million tracks catalog into 320 kbps ogg files in two years time. Or, I guess, one 8-core workstation will suffice?

So why Spotify hasn’t done this yet? Honestly I have no idea. The only explanation I can imagine is to save bandwidth, but that seems to be too trivial a reason for not delivering goods to paid customers who took it for granted since 2009. Remember we are talking about a company who aims at getting 50 millions US users in first year. Since last week I asked many Spotify directors on Twitter, but no direct answer. CEO Daniel Ek said, to my big surprise, “Don’t know the answer but I’ll look it up for you!” Nothing came up so far, except for those two email quoted above. To me their reactions to this issue are like this:



I hope Spotify pay enough attention to this problem, it’s not a feature request, but giving Premium users what they are paying for. In another word, It might not be an exaggeration to say that Spotify is in danger of being accused of false advertising here.

I believe the majority of Spotify premium users are under the impression that most of Spotify’s catalog is already available in 320 kbps. Because after they announced the news of high bit rate in June 2009, most media press and online sources just refer to the Premium sound quality as 320 kbps, without noting that not all tracks are currently available in HQ.
 

A few examples:

Spotify Blog:

“Initially, not all tracks will be available at the higher bit rate. We’ve begun converting the most popular tracks first and over the next few of weeks and months we’ll be adding more and more high quality tracks until the entire catalogue is available in hi-fi.”

Wikipedia:

“…or optional q9 (approx ~320kbit/s) for Premium subscribers, the highest streaming rate for any online service.”

TechCrunch:

“Initially the most popular tracks will be available at the higher bit rate, with the entire catalogue converted to 320 kb/s over the next few weeks.”

and this quote from an interview:

“Currently we offer the option of listening at 320kbps on Spotify Premium and we use the Ogg Vorbis codec, which is considered to be very good amongst audiophiles,” defended Söderström.

Yes on their official site: Spotify do declare that “not all tracks are currently available in high bitrate”. But from a consequentialist’s point of view, it still looks like de facto false advertising. How many Premium users would expect that after two years, even the majority of new and Premium exclusive contents are only in 160 kbps?

If you are a premium user and also angry about this, please do something to persuade Spotify to speed up the long overdue conversion to 320 kbps streaming. Please ask questions on GetSatisfaction, Twitter, Facebook, Quora or contact their customer services, let’s hope our effort would make Spotify more transparent to users. Thanks.

Lastly, dear Spotify, I want to emphasize that: you are perfectly OK not giving users essential features like gapless playback or iPad app, at most you lose some customers, anyway you cannot please them all with scarce resources. But this is different: it could get into a lawsuit and you might even have to refund premium users if the case went wild. I really don’t want to see that happens, I still think that you are the best music service (though you lost the crown of the highest streaming rate to MOG for now) and hope you become even more awesome. Yours.

Ulysses Shi
A loyal Spotify user and supporter

Update (11/07/2011): After my constant inquires, Spotify costumer services sent me an new email this morning, it reads:

“Thank you for your message and your interest in clearing things out! You and the community posts you referred to, are correct in most cases. But I will however tell you a couple of things I just learned that will work as kind of an official response in the matter.

As you stated, all music we get is in lossless format. All music is intended to be converted to all three qualities. We get thousands of songs each day, and the converting process is always chewing new data. Surely we might need to upgrade and enhance our methods and equipment to make this go faster, but for the time being, I can tell you we're doing all we can to provide the users with 320 k ogg vorbis on every song!

Hope this answer was sufficient enough for you!”

It only confirms my speculation about how they get music from content provides, but the lacking of time and equipment is far from a satisfying, or even reasonable answer (see my calculation above). What Spotify promised in June 2009 is they would convert the whole catalog “over the next few of weeks and months”, now more than 100 weeks have past. Something else went wrong, but it is definitely not that Spotify (with $100m new funding) could not find enough computers to do this job, which by all means should be a high priority.

Update 2(12/07/2011): After I replied with my estimation that a couple of home computers can finish the job in two years, CS sent me another reply:

“Thank you for the further investigation. I am not however obligated to tell you exactly what our equipment is doing all the time, but I will forward this issue to the dev and content-departments to see what they have to say. I cannot promise you an answer or any further explanation but if I do learn such a thing, I will contact you again! “

Do you see the pattern here in all four replies? Whether deliberately or not, he always gave me a false answer, until I proved it wrong. I guess he also sent the previous answers to other people who asked the same question, but I might be one of the few who got this far. I will keep on digging, and please help to spread the news until Spotify cannot ignore this crucial problem anymore. Thanks.


Update 3 (18/07/2011): I wrote this article about ten days ago, and had to postpone the publication due to various reasons, one of them being Spotify told me, through a 3rd party, that an official reply to my query on GetSatisfation would be made shortly. Nothing happened except for the customer service John, whose emails I quoted above, repeated his answer that "we are doing all we can".

You see, this is why I am angry about this. I am perfectly happy to pay a tenner for Premium even without 320 kbps streaming, but the "ambiguity, dearth of information, and poor responsiveness" (to quote another customer) is what I don't want to see from a company that I greatly admire.

The U.S. market is a swim-or-sink place, and this problem is already hurting Spotify from the beginning. Techcrunch wrote in their comparison of Spotify and Rido:
 
"Spotify Bitrate Quality: 160 kbps with some tracks at 320 kbps for premium users."

That's about a more accurate description of the fact. Even if more than half of Spotify's catalog is in 320 kbps, it still doesn't mean a thing if you don't tell the users the exact proportion, and don't even let them see the bitrate of currently playing track. In my opinion Spotify's way of handling this problem so far devalued themselves and that makes me sad. I still believe they are the best thing ever happened to music since p2p file sharing, and I sincerely hope they will fix the problem in time, with dignity.

Note: Complete correspondences with CS can be found on this Google Doc. All 115 tested tracks are in this playlist: 320K Test.


Important Update (16/08/2011): Spotify converted a great number of recordings to 320 kbps in the past few weeks. See new comments here and in GetSatisfaction for more details.

Important Update (23/09/2011): Spotify's Official Statement: "All music streamed through Spotify is of high quality (no less than 96 kbps for mobile and 160 kbps for desktop). We have a catalog of more than 15 million tracks and more than 99.9% are available in high bitrate (320 kbps) for our Premium users. Our catalog adds an average of 10,000 new tracks daily and we add the newly added tracks as quickly as possible."

108 comments:

  1. I appreciate your diligence in looking into this, 160 kbps is not acceptable for a premium service.

    I am especially shocked that the CEO Daniel Ek didn't know about this.

    Why they wouldn't give you adequate or even truthful responses is a travesty. They really crossed a line there.

    It would be really unfortunate if Spotify again wasted your time and that of others who are soliciting the proper explanation for our undelivered premium service. They would be wise not to disrespect you with another false answer.

    ReplyDelete
  2. Great article but disappointing to find a low number of premium tracks, I listen a lot via headphones and can tell a difference on some tracks

    ReplyDelete
  3. This is 160kbps OGG, not MP3. There's a difference.

    ReplyDelete
  4. Good for you for making noise about this and keeping at them. I hope they will resolve this and live up to their advertising claims.

    I've had Spotify for just three days now in the U.S. and was seriously thinking about Premium. Now I want to know just what the hell that money would actually get me.

    ReplyDelete
  5. It is a shame that Spotify still hasn't fixed this problem. After I did a blind test between MOG and Spotify both on the computer and mobile phone, MOG clearly won out with better audio quality. Unfortunately, although Spotify has a larger classical music catalog, most of these albums are missing tracks in the US. MOG also has most of these albums, but all the tracks are playable in 320kbps. It is a shame that Spotify is such a disappointment in music quality and selection for classical music.

    ReplyDelete
  6. You cannot just recompress audio to make it sound better. E.g., if you have audio compressed at 128kbps, the data is already missing. Decompressing to a .wav won't bring back the missing data. MP3 files are called loss, because you loose data forever when you use the format.

    ReplyDelete
  7. Play 160kb ogg files and 320kb and see if you can hear the difference...

    ReplyDelete
  8. They're going to do what's best for their business. Unfortunately, 99% of their users will never care or ask, and even if they did they wouldn't be able to tell the difference. They're paying premium for the mobile streaming, not for the quality.

    The responses you've got exactly mirror that model. Spotify is doing exactly the right thing here business-wise: it's not worth it to them to spend the extra hundreds or thousands of dollars on bandwidth just to satisfy a couple squeaky wheels, sadly even if they lose them as customers. It's a sad reality of capitalism that it optimises for mediocrity.

    Frankly, you should be glad that they're running any of them in 320, and it's a good sign that they're working to move more of the library toward that goal; I believe that's true despite it clearly not being a very high priority.

    And they're upholding everything they've said, and not being particularly dishonest. From their "premium" details: "Premium members deserve premium sound quality. To be precise, you can stream music at a higher bitrate of up to 320kbps on your computer (not all tracks are currently available in high bitrate)".

    Overall, that's a pretty good description of exactly what the situation is. No one's lying, and no one's spilling any oil... for what we expect of a company these days, I think they're doing pretty damn good...

    ReplyDelete
  9. @rob t, I know. I didn't suggest Spotify to re-compress ogg q5 to q9. They got all the lossless files but didn't compress them into HQ ogg, that's the problem.

    @Anonymous, yes I can, and my headphone is merely an AKG-K450. I started to use headphone more and more recently, and found Spotify's sound quality unsatisfactory (on desktop client, I know Spotify Mobile doesn't offer 320 kbps at all). That's why I started the investigation in the first place.

    By the way, I think the ever-ongoing war between audiophiles and "nobody can tell 128/192/256/whatever kbps lossy files from CD" supporters is irrelevant here. I am not suggesting Spotify to give us a better quality that maybe doesn't mean too much for some other users, I am asking them why they didn't deliver the goods they promised more than two years ago.

    ReplyDelete
  10. @Tristan, thanks for the thoughtful comment. I know practically I am tilting at windmills here, and most likely things will end just as you predicted. But one of my major weaknesses is, I can't help but doing what I believe it's right, even if it's contrary to my own interests. Spotify helped to promote this blog a lot before, and I won't be surprised if they never do that again in a long time:(

    That being said, we can't say there's no risk on Spotify's side. Competition is much more stiff in the US market, and I can imagine it won't take long before other services like MOG start to look into this problem, even without me bringing up the story. After all, MOG is paying for the cost of their 320 kbps streaming, while Spotify enjoy an unfair advantage that, while they didn't offer maybe a substantial part of their catalog in HQ, many users still believe they are getting HQ streaming.

    ReplyDelete
  11. We should challenge Spotify to put an HQ icon next to all tracks available in high quality format. That would make it extremely clear what you're paying for.

    ReplyDelete
  12. Thanks for the great research. Now you're in the news (use Google translate): http://computersweden.idg.se/2.2683/1.395925/spotify-lurar-premiumkunder

    //Lars Karlsson

    ReplyDelete
  13. @ulyssestone - you're absolutely right, it's still a factor, and you're doing the right thing for calling attention to it.

    I think they've got quite a bit on their hands at the moment with the US launch (and certain other issues I've been experiencing as well, like skippy playback or songs outright not playing) but hopefully once they get everything in order they'll continue to expand their high-quality catalog. Hope so!

    ReplyDelete
  14. Thanks Lars for writing about this, nice picture on the IDG frontpage.

    http://twitpic.com/5sjkan

    The Google translate link to your article.

    ReplyDelete
  15. I really doubt that anyone here can tell the difference between 160kbps and 320kbps OGG files by listening to them. Sure, if you had a really high-quality audio setup at home and spent some time reading up on the compression artifacts typical of the OGG format, and then really strained yourself to try to pick out those faint artifacts, you might be able to separate them in blind tests. From my experience, though, 160kbps is perfect audio quality for modern compression schemes (AAC HC, WMA, OGG). I understand that you paid premium rates and expect something more, but the only way you'll know you're getting a higher bitrate is by looking at the file sizes. If you disagree, you can try running some ABX tests and see if you can distinguish 160kbps from 320kbps or WAV.

    ReplyDelete
  16. Are the HQ files of constant or variable bit-rate? In the later case you cant link the file size to the bitrate, and the 320kbps is the peak bit-rate at one single time in the file.

    ReplyDelete
  17. This unethical conduct on Spotify's behalf is very shocking to me! Every Premium Spotify member should be 100% aware of being unethically treated like this for the purpose of making cash flow more efficiently to Spotify's business owners. This is damaging the competitors on the market as well who are competing on false ground.

    Thanks for a work well done! Reports like this help clean up the market and make it more efficient.

    ReplyDelete
  18. I still think the best digital sound you can get (apart from a real WAV CD) is to buy the download from Apple or Amazon and then burn to a CD-R (because most iPods and Smartphones aren't great music players when plugged into a good stereo). Agreed it stinks that Spotify is not being forthright, but I tried out MOG which does run at 320, and when I connected my iPod Touch to my stereo, the sound was disappointing. Truth is, all these digital services are about FM radio sound quality, which is perfectly fine most of the time. And with a lower bit rate, you use less bandwidth -- in the U.S., data caps are becoming more the norm. I subscribe to Napster, which is cheaper than Spotify if you subscribe for a year ($8/month) -- they're transparent about their bitrate which is 128 on the web and 64 on mobile. If I really love something, I buy it.

    ReplyDelete
  19. "No one's lying, and no one's spilling any oil... for what we expect of a company these days, I think they're doing pretty damn good..."

    One of the saddest things I've ever read. How much corporate crap will be excused with this exact logic: 'At least they didn't destroy the ocean.' My, how the bar has lowered.

    ReplyDelete
  20. note that there's no 320KB option on spotify's android app in any form - not the sync to phone (download the songs), not the streaming option, not over wifi asnd not over 3g.

    ReplyDelete
  21. "Anonymous said...

    Are the HQ files of constant or variable bit-rate? In the later case you cant link the file size to the bitrate, and the 320kbps is the peak bit-rate at one single time in the file."

    Yes Spotify use vbr and I was aware of that when I did my test. I think there is a reasonable file-size frame even for ogg files encoded in vbr, when a 4-minute track is only 5 MB in size, it can only be 160 kbps vbr, except for some rare exceptions like historical mono recordings of solo piano.

    There might be one or two misjudgements in my test, but it won't affect the result much.

    Another test shows a similar result.

    ReplyDelete
  22. I signed up for Spotify Premium in the US on July 18, and one of the main reasons that I opted for Premium was the 320kbps bit rate. I installed the desktop app and noticed right away that the sound was muddy.

    After some looking around in the program, I noticed that there was as "High quality streaming" check box under Preferences. Once this was checked, I could immediately hear a difference. I guess it was just a coincidence that the first album I tried was available as 320kbps.

    I have noticed quite a few tracks that don't sound like 320kbs. I have done a few A/B listening comparisons between Spotify and to the same album ripped from CD at 320bpbs and stored locally; I can absolutely tell the difference. My rips sound noticeably better.

    The fact that HQ streaming is disabled by default says that Spotify is obsessed with profit to the point that they're willing to cheat customers out of what they've paid for, and then simply hope the customers won't notice. That makes me very uncomfortable.

    @Ulysses, thanks for looking into this. Your efforts are *very* much appreciated.

    ReplyDelete
  23. I have the same experience with bitrates.... earlier playing around with despotify.

    I guess it would be possible do a small hack (based on despotify) that reports the available bitrates of all the tracks in a playlist

    ReplyDelete
  24. First, let me just say that I'm totally with you here. Spotify should live up to their promises. The converting from lossless to various bitrates of Ogg is obviously automated in some way, so the fact that all tracks don't exist in Q9 must be a concious and intentional decision.

    Secondly, a lot of people here think they can tell the difference between 160 kbps and 320 kbps. Having done real blind tests (using a special program that randomizes the tracks for me) I can say that even at 128 kbps (Q4) Ogg is *extremely* difficult to tell from the original wav. Calling 160 kbps "muddy" is just ridiculous and can only come down to one thing: placebo.

    ReplyDelete
  25. OK, I pay for something I don´t get. I will change.
    Bjorn B

    ReplyDelete
  26. @David Naylor:

    "Calling 160 kbps "muddy" is just ridiculous and can only come down to one thing: placebo."

    You are making the assumption that because you cannot perceive something, no one else can.

    Your logic is flawed.

    That's like a blind person insisting that colors are imaginary.

    The first album I listened to on Spotify, I assumed that because I bought HQ streaming, that that's what I was hearing. The fact that it sounded muddy prompted me to investigate the bit rate, and I was correct in my suspicion that it was not 320kbps.

    That's not a placebo; rather, it demonstrates a good ear for music.

    Alkettory

    ReplyDelete
  27. Spotify staff on Bitrategate:

    "You may find that something else suits you better at present. ..but run this 'Deathmatch' yourself. It looks like it's the only way."

    Link here. See first comment (cannot link to his comment directly for unknown reason).

    ReplyDelete
  28. Why do you fight for MP3 at 320 kb/s instead of lossless FLAC? That is the only thing that should be provided and advertised as HQ in my opinion.

    ReplyDelete
  29. @Anonymous, because I am not asking Spotify for a new feature like flac, I am just asking Spotify to deliver the goods they've been promoting for two years.

    ReplyDelete
  30. Many thanks for investigating this.
    I am a premium customer at Spotify and have had this feeling from the beginning that the quality i far from HQ.
    It has been very frustrating not to have proof of what my ears are telling me.

    /daniel

    ReplyDelete
  31. Spotify need about 20 years to do this simple thing? How awesome...

    ReplyDelete
  32. Hi Ulyssestone, I am h8ten from the 320k issue on the Spotify support forum. I suggest we make a backup of that discussion somewhere, just in case the discussion "disappears" from the forum for whatever reason. There is lots of valuable information on there. Could you do that?

    ReplyDelete
  33. Thanks for the reminder, Concertzender. I have all the emails from GetSatisfaction in my Gmail, and will try to save that GS page as a pdf or screenshot now.

    ReplyDelete
  34. Image backup (3.6 MB):

    http://dl.dropbox.com/u/371318/Why%20so%20little%20music%20available%20at%20320%20kbps%3F.png

    I also saved it as pdf, using Firefox's Print function, but the format is a bit messy, will try some other add-ons later.

    ReplyDelete
  35. 320k suggests a pristine master-quality rendition. This has some very awesome financial implications especially in publisher negotiations. I cannot believe a cloud-based streaming service is short on storage or converer-bot virtual machines in their data center.
    What I can believe is that they have been unsuccessful in securing HQ rights for the entire catalog. If I am a content owner/publisher, I would hold out for higher royalties/fees for higher quality because at 320k, this approaches master (ripp-off-able, pirate-able, made-in-china-able) quality. Who said a subscriber could not setup a program to record from the sound buffer on one's computer, this pristine 320k data stream and make CDs or bootlegs from it?

    I think they are letting us conclude that it is just a business decision so we will give up and stop asking questions.

    They get PCM files (uncompressed) from pubishers but they don't always have the right to offer them at HQ.
    Not sure they will want to respond to this.

    S

    ReplyDelete
  36. @salvideoguy

    Not a bad argument, but somehow MOG manages to stream 320kps in the US for the exact same monthly price as Spotify Premimum (US $9.99).

    Alkettory

    ReplyDelete
  37. Hi Ulysses,
    I must congratulate you to this excellent forensic study and article! I'm not a Premium user, but I have been using Spot for quite some time now (and happily so), and I certainly was under the general impression that if I were to pay them for Premium service, all or most of the streaming would be at 320 kbps.
    Like you, I tend to be quite "understanding" if businesses don't get everything right, and I also hate it when they don't own up to their responsibility when something's not going the way it should.
    Keep up the good work!

    ReplyDelete
  38. Can someone post a simple chart--or a link to something--describing the quality differences in these "bitrates?" If a commercial music CD is X, and (iTunes?) MP3 is Y, where is 160kbps or 320kbps, and FLAC? What are other digitial streaming media in comparison; example, Sirius/XM OTA or through the internet, FM radio (which is analog, so?), any scientific studies of human perception (or lack there) of recorded sound at the various "bitrates?"

    ReplyDelete
  39. Jean-Paul BertrandJuly 25, 2011 at 3:13 PM

    I am in France and I use Qobuz.com
    The catalogue was at a time limited but did grow quickly to an extend that now enlarge a lot and the quality is splendid. Try it !

    ReplyDelete
  40. There seems to be a lack of understanding here with the compression of audio formats.

    "Can someone post a simple chart--or a link to something--describing the quality differences in these "bitrates?" If a commercial music CD is X, and (iTunes?) MP3 is Y, where is 160kbps or 320kbps, and FLAC?"

    160kbps does not tell you good, bad or ugly. It can be used as an indication though. It is similar to describing the operating frequency of a processor as the determining factor in how much work it can do in a given space of time. This is simply not a viably accurate assessment, but again, does provide SOME general indication. The quality, perceived quality, or lack there of (to the guy earlier, it's less about ADDING artifacts, and more about LOSING fidelity, primarily on the top and bottom of the spectrum) is a direct result of the amount of information contained in the recording. Described another way, the `dynamic-ness`.

    If you compress someone speaking (without music) at 320 and 160, you'll likely not be able to perceive a difference, even with stellar audio equipment. Because quite simply there is plenty of bandwidth available to fit within the lower bit rates data/sec. Conversely, in something like very busy orchestral music with extremely varied/dynamic sound across the spectrum, you'll eat up that bandwidth per sec easily, and the result is (essentially) some audio, which could be described as fidelity,.. the subtlety on the high and low ends, being chopped out, muted, etc.

    The result is something slightly hollow, less warm, or deep compared to the lossless (essentially) source.

    Please forgive all the generalities I've used here, but hopefully the point is understood. 160 isn't inherently bad, but it certainly isn't as high fidelity as 320. But the bottom line is that it depends greatly on the material!

    ReplyDelete
  41. I am very angry with Spotify over this deception. There can be no doubt that they announced in June 2009 that premium subscribers would get High Quality for all the catalog. The CEO himself said back then that the conversion would be done "in weeks".

    Clearly that hasn't happened, yet. Some people have suggested that this is like internet companies advertising their speeds as "up to ....". It isn't the same, though. For them, there are always a few people who can get the advertised speeds (even if not many!). But here, NO-ONE can get access to the full Spotify catalog in High quality. It just hasn't been done. No, this isn't optimistic advertising - I think that this is very close to being false advertising.

    Sadly, I wonder whether Spotify are happy to lose a few premium members over the issue, hoping that the noise will die away, and they can continue to mislead others in the future.

    Personally, though, I don't want to leave - but simply to get what I thought I was paying for.

    The voice of the internet can be a very powerful lever. If this were picked up by a few high profile blogs in the USA, then surely Spotify could not afford to ignore it. Keep spreading the news !

    ReplyDelete
  42. I had Spotify Free version and upgraded to Premium, as I listen through my fairly decent home audio setup (Tech details: From my PC via a Yamaha Audiogram 3 USB sound card, fed via high-quality cables thru a Yamaha RX-V663 HT Receiver in "Pure Direct" mode (no signal processing) to my NAD 2400 THX amp, then to my custom-cabled Dynaudio Audience 122 speakers and Dynaudio Sub-10 250w subwoofer). I can definitely hear the positive difference since switching to premium, as I have A/B tested many songs against their CD or LP counterparts which I own and with which I am fully familiar. ALL sound as good or better than the CD or LP! Madonna's 'Vogue' is amazing; Diana Krall's "Let's Face The Music and Dance' is every bit as good as the very well-recorded CD version, etc.

    Bottom line: I have had NO issues with sound quality once switching to Premium. Where one will notice the difference first in a lower bit rate recording is in the bottom end. A low bit rate recording produces a muddy sounding bass. It also reduces the mid-high crispness and detail and restricts the "air" around each instrument/vocal. It's quite audible on a good system to a discriminating listener. But it is no doubt inaudible/irrelevant on the vast majority of computer headphone and desktop speaker setups.

    I trust my own ears. ☺

    ReplyDelete
  43. @Anonymous, I'm glad you are enjoying Spotify. Since I can see that sound quality is important to you, maybe you should read into my original post more closely. According to my test, perhaps only one third of Spotify's catalog is available at 320 kbps. You are lucky to stumble upon some HQ albums, but I'm sure you will notice the difference when you listen to more albums on Spotify. Try Radiohead's King Of The Limbs, Coldplay's Viva la Vida, Stones's Exile on Main Street, or most of the classical music there.

    Again, I'm not complaining about the sound or overall quality of Spotify's service, I'm questioning why they don't deliver on the promise they made more than two years ago. The fact that you are satisfied with what you've heard so far (in a week?), or other users might not be able to tell the difference between 160 and 320 kbps, is irrelevant here.

    ReplyDelete
  44. re: "You are lucky to stumble upon some HQ albums"

    I would say that if a song or album sounds as good or better than the CD/LP version using high quality audio equipment, bit rate is irrelevant. All I know is that switching to Premium (and engaging the 'High Quality Streaming' option under PREFERENCES) instantly improved everything. Perhaps Spotify "Free" is squeezed even further than we think and 160kbps is the absolute best one might ever hope to hear for free.

    If I did not hear an audible difference I would not pay for Premium. I have not been disappointed. (Currently listening to John McLaughlin/Shakti's album 'A Handful of Beauty', which I have listened to at least 1000 times on LP and CD. It sounds GREAT!)

    Cheers! ☺

    ReplyDelete
  45. @Anonymous,

    1, you haven't heard every album on Spotify or all albums you want to listen to, if you think what you heard so far is great automatically guarantees everything is great, good for you.

    2, Enjoy the music:)

    ReplyDelete
  46. re: "you haven't heard every album on Spotify"

    No one has -- or can.

    I have fairly eclectic tastes when it comes to music -- everything from jazz to country to club/DJ to hip hop. I also know that there is an inherent quality to any given recording. This is why for my sonic assessment of Spotify Premium I have referenced high quality recordings that I own.

    Now, I have definitely heard a few recordings on Spotify that would not be classified as top quality, but that is a function of the original recording itself. On some older recordings that have not been remastered there would probably be no loss of "quality" if they were streamed at a lower bit rate. But I have yet to hear an inherently high quality recording that sounded bad. Indeed, I've been blown away at how good some of the older recordings sound. I listen to a lot of Bill Evans stuff from the 60s and 70s that has in some (but not all) cases been remastered. The remasters sound terrific and the others sound just like the original recordings that they are. In the case of classical, there is also a wide range of inherent recording quality. It depends mostly upon when the recording was made and whether it has been remastered. I was listening to some Mahler the other night and it was just gorgeous.

    If I do run across something that I know should sound better than it does, I will post it.

    ReplyDelete
  47. @Anonymous, I don't understand why you would think that all music You heard on Spotify so far sounded great to You is relevant here.

    If you are suggesting all music on Spotify is at 320 kbps, look at the stats above, you are wrong.

    If not, thanks for sharing, It's great that you enjoyed your music, but that's irrelevant here.

    If you think "320 kbps or not, it sounds great anyway". Well, sound quality is highly subjective, I don't want to quote other people's bad comments here, I just want to repeat that: what You think people should care (the actual sound quality, the kbps figure) is not important, the original post is not about music appreciation, it is about whether or not Spotify's 320 kbps is false advertising.

    What Spotify promised in 2009 was not "a great sound", it was "320 kbps for their complete catalog, with the next few weeks or months". I hope I made myself clear enough this time:)

    ReplyDelete
  48. Look, I'm not stupid and I understand what you are saying. But ultimately it boils down to this: Is the sonic quality of Spotify Premium superior to their free version, thus worth 10 clams a month? Over the wide range of recordings I have listened to, the answer is yes. Is the quality of Spotify Premium on par with the equivalent CD/LP version of the same recording? Again, yes. Therefore, to focus solely on bit rate is academic, because it does not matter if it is not audible on a given recording. As another poster pointed out, 160kbps would be indistinguishable from 320kbps for certain -- even many -- recordings. Perhaps Spotify knows this and is targeting 320kbps for those recordings where it would make an audible difference.

    If you can point me to a recording where you have done A/B testing using a decent audio setup and have found an audible difference between Spotify and the CD/LP, please do so. I have not found one yet.

    I should also point out that in addition to the aforementioned audio system on which I do most of my listening, I also listen through two different pairs of high quality headphones, plugged directly into the Yamaha Audiogram 3 USB sound card: Phiaton PS 20 NC Active Noise Cancelling Earphones, and Allen & Heath XONE XD-53 Professional Monitoring Headphones.

    ReplyDelete
  49. Quick answer:

    1, Personally I think Spotify Premium well worth a tenner per month. I never said it didn't.

    2, Spotify should give their customer what they promised and advertised. Again, not just "a great sound", not "I will give you all the albums that I think worth to be presented at 320 kbps at 320 kbps", but complete catalog in 320 kbps.

    It's their promise, not my request. If you think that's unreasonable, ask Spotify to change, not me. Thanks.

    It is as simple as that, asking a business organization to fulfill their promise.

    I have no interest in talking about audible difference here.

    I've seen dozens of people complaining about the sound quality in the past week, go search for it by yourself, please.

    ReplyDelete
  50. re: "I have no interest in talking about audible difference here."

    Because there isn't one.

    Nothing else matters. Splitting hairs over the bit rate for any given recording would be meaningful only if it impacted the sound negatively.

    As far as "I've seen dozens of people complaining about the sound quality in the past week", I dunno. Some people just like to complain. I'm fairly discriminating when it comes to sonic quality and I've given my results.

    Best of luck in your pursuit of bit rate compliance.

    ReplyDelete
  51. You can't hear a difference doesn't mean nobody can. Try to understand that, please.

    I didn't start my article with "I think the sound is bad" because there is no universal standard. Some think Youtube mono 32 kbps is perfectly fine, some think Blue-ray 5.1 Surround 96kHz/24-bit/ is not good enough. Everyone has their own standard for "good enough". You can't tell people how they should feel.

    In order to make my case clear and prevent people like you from jumping over it saying "it's good enough FOR ME (so everyone should feel the same) why should you complain?" (as it happened a million times before) I based my argument on one simple principles:

    A business organization must fulfill their promise to customers, because it is a legal obligation.

    If you deny that, I have nothing more to say.

    Honestly I don't care for your results or feel the need to show you mine, because I never said that's what I'm looking for. If that makes you feel I can't hear a difference, feel free to think so.

    Just beware that, if a judge asks you why you didn't perform your obligation under the written contract, "I think I've done good enough" might not be an answer that's good enough. It's just an fictional example, sorry if it offends.

    ReplyDelete
  52. re: "A business organization must fulfill their promise to customers, because it is a legal obligation."

    Here is the exact "promise" made by Spotify for their Premium service:

    'To be precise, you can stream music at a higher bitrate of UP TO [emphasis mine] 320kbps on your computer (not all tracks are currently available in high bitrate).'

    http://www.spotify.com/us/get-spotify/premium/

    Seems pretty clear and straightforward to me. This is what I understood when I signed up. I think it would also be clear and unambiguous to a judge.

    Have a nice day ☺

    ReplyDelete
  53. If you see someone says a product is "not all xxx" and feels it's "pretty clear and straightforward " and buy it right away, good for you.

    Ummm, look, are you interested in buying this adorable statuette? It's not all gold.

    If you can't bother to read the OP to see where my argument comes from, please don't bother to comment either.

    "Initially, not all tracks will be available at the higher bit rate. We’ve begun converting the most popular tracks first and over the next few of weeks and months we’ll be adding more and more high quality tracks until the entire catalogue is available in hi-fi."

    http://www.spotify.com/us/blog/archives/2009/06/25/bumping-up-the-bitrate/

    ReplyDelete
  54. re: "...we’ll be adding more and more high quality tracks until the entire catalogue is available in hi-fi."

    Let's see: "Hi-Fi". Hmmm. seems to me that I have shown that they are in fact streaming very high fidelity music.

    In any case, I see no time limit on this. There is nothing contractual here, it is merely a statement of intention, made back in 2009! The only statement one should concern themselves with in 2011 is the one I quoted which describes honestly what the Premium service offers.

    ReplyDelete
  55. What seems to you is hi-fi, is irrelevant (I've told you that before, and you seems agreed becasue you dropped that argument for a while. Would you please show some consistency when you speak, instead of going back and forth?). In the context of that blog post, "hi-fi" meant 320 kbps, otherwise they are contradicting themselves. If they meant 160 kbps is hi-fi enough, why the conversion? Why "adding more and more high quality" while they were all 160 kbps from the beginning?

    I find it ridiculous that an alleged customer like you would try so hard to speak for a business organization. Is your ego so big that you can't stand other people criticizing a product that you enjoyed? And you can't even see that I am not even criticizing the product itself?

    I'm not a judge and I'm not convicting Spotify for any crime. I'm an independent blogger trying to call attention to this problem. This misleadingly advertised feature. It's not a problem for you, fine. The problem exists, for everyone who's complaining, and as customers they are all entitled to do so.

    If you feel so strongly that your argument is the ultimate and indubitable answer to this issue, don't bother to post here, email Spotify's PR and they will award you for solving this problem. But I doubt they dare to tell the users what you said above.

    ReplyDelete
  56. I've pointed you to Spotify's current statement re: precisely what to expect from their Premium service. It is not in any way misleading. It's about as clear and plain a statement as one could ask for. You seem to be hung up on "that blog post" as though it were a stone tablet handed down from on high. It isn't. You certainly have every right to complain if you are dissatisfied, but don't try to make it seem as though Spotify is being deceptive or is in some way violating some supposed legal agreement regarding what, exactly, they offer in their Premium service.

    As far as our respective posts here on this matter, I am confident that I have made my points. Each reader will have to make their own judgment as to the validity of what has been posted by either of us.

    Have a nice day. Really.

    ReplyDelete
  57. Please allow me to summarize:

    For you, “not all tracks are currently available in high bitrate” is “pretty clear and straightforward”and "as clear and plain a statement as one could ask for", after seeing that, you are perfectly aware of what this product will offer. And everyone should feel the same.

    If some very picky costumers felt that’s not clear enough, and in order to find a more clear description of the product, they found that 2009 official blog post (the most quoted article on Spotify Premium sound quality, the original source of thousands of media press that refer to Premium sound quality as 320 kbps "within weeks or months since June 2009). You think they should interpret that post as follows:

    1, In the context of an announcement for the availability of 320 kbps streaming, everything they promised there didn’t necessarily mean 320 kbps.

    2, “over the next few of weeks and months we’ll be adding more and more high quality tracks until the entire catalogue is available in hi-fi.” meant “we will try to make our entire catalogue sounds great (but not necessarily means 320 kbps), and no timeline was given”. And when a Spotify employee said “It is true we say all our music will be available in 320 kbps”, he was wrong, but Spotify is always right.

    3, If a user didn’t think in the way you think they should, it’s their own fault that put them under the impression that most/all of Spotify’s catalog is at 320 kbps. Regarding this topic, Spotify has always being honest and straight, therefore no complainant or inquiry is reasonable (though tolerable, as you kindly pointed out) , even if some stats disclose that maybe only 30% of their catalog is at 320 kbps, two years after the announcement.

    Did I miss something?

    My comments? Did you miss something?;)

    ReplyDelete
  58. Guys - there's lots of words here - but there is a fundamentally false premise in Anonymous's arguments.

    He asks : "Is the sonic quality of Spotify Premium superior to their free version? .....Over the wide range of recordings I have listened to, the answer is yes".

    Simply put - that is impossible. The majority of recordings are at 160kbps - so will sound THE SAME whether you are Premium or Free. Thats not an opinion, that is a fact.

    If anonymous maintains that he cant tell the difference - that fine. Thats down to his hearing, so we can accept that. It doesn't mean that others can't, though.

    And as to the argument that a promise from the Spotify CEO in 2009 to convert all the tracks "within weeks" is just a "statement of direction" - well, it reads pretty precise to me. The words were "will convert within weeks", not "might convert sometime".

    Sorry, Anonymous, everyone is entitled to a viewpoint. But in this case, the facts are against you.

    ReplyDelete
  59. re: "The majority of recordings are at 160kbps"

    Spotify has >15,000,000 tracks. In order for you to say what percentage Spotify Premium users are getting at 160kbps with a confidence level of 95% and an interval of 5, you would need a sample size of 384. And your testing methodology would have to be independently verified. Neither of these conditions has been met in your testing.

    I also believe that this is worth noting (from vorbis.com):

    Vorbis' audio quality is not best measured in kilobits per second, but on a scale from -1 to 10 called "quality". This change in terminology was brought about by a tuning of the variable-bitrate algorithm that produces better sound quality for a given average bitrate, but which does not adhere as strictly to that average as a target.

    This new scale of measurement is not tied to a quantifiable characteristic of the stream, like bitrate, so it's a fairly subjective metric, but provides a more stable basis of comparison to other codecs and is relatively future-proof. As Segher Boessenkool explained, “if you upgrade to a new vorbis encoder, and you keep the same quality setting, you will get smaller files which sound the same. If you keep the same nominal bitrate, you get about the same size files, which sound somewhat better.” The former behavior is the aim of the quality metric, so encoding to a target bitrate is now officially deprecated for all uses except streaming over bandwidth-critical connections.

    For now, quality 0 is roughly equivalent to 64kbps average, 5 is roughly 160kbps, and 10 gives about 400kbps. Most people seeking very-near-CD-quality audio encode at a quality of 5 or, for lossless stereo coupling, 6.

    http://vorbis.com/faq/#quality

    [NOTE: Spotify Streams are q5 quality. Premium subscribers can enable higher bit rate streaming at q9 from the preferences menu.]

    https://www.spotify.com/us/help/faq/tech/#codec-quality

    ReplyDelete
  60. SEE ALSO (from "An Introduction to Compressed Audio with Ogg Vorbis"):

    Ogg Vorbis format encoders don't normally consider bitrate at all (the default mode of operation is VBR), instead using a "quality" rating, which ranges from -1 to 10 in increments of 0.01 or so. This quality rating is a measure of how close to the original the compressed file should sound; the encoder uses as many or as few bits as necessary to satisfy the quality requirement. Each quality setting results in a rough average bitrate for a piece of average music, but this is a by-product of how the encoder has been tuned; the encoder does not aim at any particular bitrate.

    Thttp://goo.gl/MVxTQ

    ReplyDelete
  61. Thus, any discussion of bit rate -- or supposed measurement of same --as it pertains to the quality of the audio for a given piece of music on Spotify is totally irrelevant.

    To quote Geoffh: "in this case, the facts are against you."

    ReplyDelete
  62. Anonymous said:
    "Conversely, in something like very busy orchestral music with extremely varied/dynamic sound across the spectrum, you'll eat up that bandwidth per sec easily, and the result is (essentially) some audio, which could be described as fidelity,.. the subtlety on the high and low ends, being chopped out, muted, etc."

    As I understood it, classical music requires a lower bitrate for the same SQ compared to a heavily dynamically compressed pop track, because the dynamic compression works "against" the tone masking effect that psychoacoustic models use to decide what not to encode.

    This may be relevant to the discussion of bitrates & quality ratings.

    ReplyDelete
  63. @ Phileas re: "This may be relevant to the discussion of bitrates & quality ratings."

    Through this discussion, we have learned that the Ogg Vorbis method of encoding is NOT based upon forcing a given bit rate; rather, it is based upon a quality scale that is "a measure of how close to the original the compressed file should sound; the encoder uses as many or as few bits as necessary to satisfy the quality requirement."

    Thus, to merely look at file size and use that to make a claim that "this must be 160kbps (ergo not high quality) because it is only X mb in size" is NOT possible. How many bits it might take to achieve q9 for one piece of music may actually be fewer than the number needed to achieve q5 on another.

    This renders the entire line of investigation used in this controversy highly suspect, if not invalid on its face.

    It is also quite possible that this technical fact is not well understood by those at Spotify who have responded to the queries about bit rate.

    It would explain why Spotify states that its Premium members will be able to listen "up to 320kbps" -- because only God and Ogg Vorbis knows what the kbps might actually be for any given piece of music at q5 or q9.

    It would also explain my observation that switching to Spotify Premium (and engaging "High quality streaming" under PREFERENCES) immediately produced superior sound. I am listening at q9 instead of q5.

    Further, I have shown that a certain rather large sample size would be required in order to say with a 95% confidence level that Spotify's 15M+ song catalog contains X percentage of songs with some certain characteristic -- assuming that that characteristic could even be measured and had any impact at all upon sound quality.

    One more thing -- I do NOT work for Spotify. To toss this around as a way of trying to invalidate my arguments is baloney.

    ReplyDelete
  64. @Anonymous:
    I'm not trying to invalidate your argument - I think you make a valid point.

    Using variable bitrate encoding, the mean bitrates of different tracks would depend on the type of music.
    Classical music requires lower bitrates for the same SQ
    The evidence suggests that classical tracks on Spotify have lower mean bitrates.

    ReplyDelete
  65. I should add, I'm not sure the facts can be fully explained with this way of thinking!

    ReplyDelete
  66. @ Everyone:

    Any specs can be misleading in terms of assessing sound quality. This is the heart of the age-old debate between audiophiles and those who focus solely on technical specs. For example, two different amplifiers can have virtually the same specs, in terms of power, THD, etc., but sound vastly different from one another. As to which sounds "better" is a qualitative judgment. Stereophile magazine, for example, has enlisted highly qualified reviewers to assess such quality. They very pointedly do not rely on "specs" to determine sound quality, as it is quite possible for a cheap receiver to have similar specs to a piece of high-end electronics. But one listen by a discerning, experienced listener can easily distinguish between the two.

    Trust your ears -- and make sure you do any critical listening on a decent sound system. ☺

    ReplyDelete
  67. I subscribed to Spotify last night and noticed that most of these supposedly higher quality tracks I tried to listen to (many being mainstream rock artists) just sounded terrible. Thanks for investigating this further, and I don't think I will continue my subscription after this month. Now I know why they don't offer a free trial like Rdio.

    I like the artist selection a lot better on Spotify as well as the offline playlists, and those are the main reasons I chose them in the end, but this isn't worth it. I'd rather actually buy the tracks in proper quality than rent them when they aren't even fun to listen to.

    ReplyDelete
  68. And in response to Phileas, I was actually thinking the same thing about bitrate while reading this post, but ultimately I did trust my ears and I have come to the same conclusion as the author of this blog post.

    ReplyDelete
  69. From the Wiki on sound quality:
    Bit Rate

    Recorded sound is stored in many formats. The need to save space on the storage device is opposed to sound quality. The smaller the sound file for a given recording the poorer the sound quality. This is not true for lossless compression methods, where the quality is preserved also in smaller sound file sizes.

    The sound quality is determined by the bit rate, which is the number of samples per second times the number of bits per sample. It also depends on number of audio channels (mono, stereo, multichannel) and compression method.

    The most common are lossy compression formats as MP3, OGG Vorbis, AAC and many others. For example, MP3 files commonly have a bit rate of 128 kbit/s, because it typically offers adequate audio quality in a relatively small space. MP3 has a maximum bit rate of 320 kbit/s.

    By contrast, uncompressed audio as stored on an Audio CD has a bit rate of 1,411.2 kbit/s, or four times the best MP3 can do, and twelve times that of common MP3 files.

    The WAV, AIFF or AU audio file formats can store audio using various bit rates, in accordance with the number of channels and compression. They often store uncompressed PCM audio with a bit rate of 1,411.2 kbit/s.

    There are also lossless compression formats, such as FLAC, ALAC, Monkey's Audio and many others, which preserve quality of original uncompressed audio, but can reduce the needed storage space. Advantageously, these lossless formats can also store metadata (eg. images, title, artist, album, label, etc.) like lossy formats can, which cannot be done with pure uncompressed formats like WAV, AIFF, AU.

    From the Wiki on Bit Rate:
    Audio

    [edit] MP3
    32 kbit/s
    96 kbit/s
    100–160 kbit/s – Standard Bitrate quality; difference can sometimes be obvious (e.g. lack of low frequency quality and high frequency "swashy" effects)[citation needed]
    192 kbits/s is the highest level supported by most MP3 encoders when ripping from a Compact Disc.
    224–320 kbit/s – VBR to highest MP3 quality.

    [edit] Other audio
    800 bit/s – minimum necessary for recognizable speech (using special-purpose FS-1015 speech codecs)
    2.15 kbit/s - minimum bitrate available through the open-source Speex codec
    8 kbit/s – telephone quality (using speech codecs)
    32-500 kbit/s -- lossy audio as used in Ogg Vorbis
    256 kbit/s - Digital Audio Broadcasting (DAB) MP2 bit rate required to achieve a high quality signal [13]
    400 kbit/s–1,411kbit/s – lossless audio as used in formats such as Free Lossless Audio Codec, WavPack or Monkey's Audio to compress CD audio
    1,411.2 kbit/s – Linear PCM sound format of Compact Disc Digital Audio
    5,644.8 kbit/s – DSD (A trademarked implementation of PDM) sound format of Super Audio CD[14]

    ReplyDelete
  70. There are any number of us posting as "Anonymous", so this can be confusing to those reading the thread, as one "Anonymous" may disagree with another.

    In the end everyone will need to make up their own minds about all of the various posts.

    From this "Anonymous", good night and good luck.

    ReplyDelete
  71. I've had enough of this, "this Anonymous".

    After my summarization, you dropped all those arguments you've being making in the past two days, and started it all over again, citing statistical theory and extended audio tech articles. If you are so confident that all my test results are inadequate, why didn’t you say that from the very beginning? Instead you went into lengthy arguments around the clock, at 4 in the morning and in the midnight, based on my “false premise”. I don't have time to debate with you, who showed no consistency in your posts (before long you went back to "trust your ears" again. I repeat last time, your ears are good for you, but they, or anyone's ears, are not the topic of the OP), or any intensest in debating other than arguing. Because you don't read or just ignore everything that's all ready written here, before you make you argument. I've said specifically about the reasonable file size frame for q5 and q9 ogg. Vorbis doesn't mean a 4 minutes q9 tracks can be compressed into 5 MB (as I noted, except for rare examples).

    And for you, in order to to raise a reasonable doubt towards a business organization, it is the customer's responsibility to conduct a full-scale scientific research and be verified independently. A 115 random tracks sampling doesn't tell ANYTHING (and you are the only one who seems think so, after 15k pages views of this post and a few media coverage). Next time you find the some of the eggs you bought in the supermarket are bad, don’t complain, they ship millions of eggs every month, you need to sample a least hundreds of them and get verified independently, before you can even reasonably question the company.

    And while I’m writing this, you went back even further to the relationship between sound quality and bitrate again, despite that argument was deemed irrelevant and your short silence after I pointed that out.

    I don’t know why I bother after all this.

    ReplyDelete
  72. Hi,

    Nice analysis! Just for an added point, to my knowledge Spotify uses some Bittorrent-derivative to stream music onto clients machines, meaning it caches dataon your computer and uses a very small part of your bandwith to supply other people with their data too.

    I just wanted to point that out because it does put some perspective on the "320 kbps bandwith issue". It is completely possible that going into a massive HQ streaming frenzy could bog down their current system entirely. They would need for example more peers per user, store more data on client machines, or even a complete revamp of their current technology to support the need of streaming 320 kbps to a bigger amount of people. When considering a single machine this is completely benign, but in a network of this size this king of small difference can crush a robust protocol down.

    I'm just throwing ideas here, I didn't analyze this clearly nor did I run any numbers. The current Spotify P2P delivery system is actually very clever and works pretty well, my guess is that the HQ transition got halted along the way because issues arose other than converting raw data to 320 Kpbs OGG.

    ReplyDelete
  73. Anonymous said:
    "They very pointedly do not rely on "specs" to determine sound quality"

    But someone must decide what average bitrate is required for every track to be at least q9 SQ? I.e. they have to rely on "specs"

    Sorry, bit off topic this, but Anonymous also said:
    "two different amplifiers can have virtually the same specs, in terms of power, THD, etc., but sound vastly different from one another."

    Frankly, I think this is baloney. Properly designed modern amps sound very similar, as shown by mumerous properly conducted blind listenimng tests.

    ReplyDelete
  74. Well I for one can easily tell the difference between the low and high quality streams. To me, Spotify is not listenable as is. Its not some desire to see a 320 instead of a 160 its that non premium sounds like crap played on a home stereo.

    So great pay the 10 buck a month... and... you still get crap sound because very few songs are in HQ. Would be great if they listed the bit rate of the songs on the interface and you could filter out the low quality files.

    If you cant hear the difference between 160 and 320 well good for you. I can and for now its disappointing. I paid my 10 bucks and am not getting the higher quality sound.

    Spotify is nice for music discovery for sure, but to actually relax and listen it does not cut it for me.

    ReplyDelete
  75. Nice work ulysses! Last may i performed some tests myself and found out that classic 'Harvest' by Neil Young was low bitrate only. Also Spotify heavily advertised with the new R.E.M. album 'Collapse into Now'. Still, that was also low bitrate only.
    At the moment, both these albums have been upgraded to 320kpbs. So Spotify is doing something for sure. I think your campaign surely helps to speed things up.

    I would also prefer more transparency in the bitrate of the songs. On the Mac, when you make a playlist available offline, the 'Console' application shows the bitrate for those songs. I would like to see that kind of information in the Spotify application itself.
    Btw i love Spotify, but the bitrate issue has been bothering me just like it did to you.

    ReplyDelete
  76. Thank you arjandr, it's amazing! Console shows every track's bitrate! Do you know how does it do this for encrypted files used on Spotify?

    ReplyDelete
  77. I just performed arjandr's console bitrate check, and so far, every song I have tested shows up as 320kbps. Music from all genres appear as 320 kbps in my console, including 'Harvest' by Neil Young. I am using Spotify US. Do tracks appear as 320 kbps in Console in other countries?

    ReplyDelete
  78. I rechecked some tracks from the 115 tracks in my OP, Console shows new tracks like Beyonce now are in 320 kbps, while old tracks like those random tracks generated by Spotibot are still 160 kbps.

    I think Spotify is making a progress!

    ReplyDelete
  79. Sounds to me Radiohead's TKOL is still 160 kbps (or it's just the cached version?), since I've already saved it to offline, I'm not able to check it now... Can anyone try to check this? Thanks.

    http://open.spotify.com/album/3EkYAh7JiJNSUxzhVLJqnL

    ReplyDelete
  80. Just checked The King of Limbs - Radiohead. All tracks are in 320 kbps.

    ReplyDelete
  81. Thanks 177, I deleted TKOL from my offline playlist, restarted my Spotify, and add it back to offline, now it starts to download again, and in 320 kbps.

    Before I restart Spotify, it won't re-download even if I deleted it and added again, so it's the problem of the cached version I download a couple of days ago. And that version was 160 kbps.

    So Spotify updated that album in the past few days, a good progress!

    ReplyDelete
  82. Tested first five albums (first track only) from Spotify's official new album releases playlist, 3 are 160 kbps.

    http://twitpic.com/5x8lkn/full

    I hope Spotify is secretly busing converting their complete catalog now:)

    ReplyDelete
  83. Could you link a playlist? So far my tests have only found 320kbps songs

    ReplyDelete
  84. It's an official playlist:

    http://open.spotify.com/user/spotify/playlist/3Yrvm5lBgnhzTYTXx2l55x

    I tested the first tracks of the first five albums.

    ReplyDelete
  85. Spotify missed this... Coldplay's most popular track is still 160 kbps, at least for now;)

    http://open.spotify.com/track/1mea3bSkSGXuIRvnydlB5b

    ReplyDelete
  86. Although some of the albums in the playlist were not available in the US, out of the 10 albums I tested, only 3 were 160kbps, suggesting that most albums are of high quality. The 160 kbps albums may be due to label restrictions, as within their own playlist only a few albums are low quality. So far a large majority of music tested was 320kbps.

    ReplyDelete
  87. About 'Console' : it shows logs generated by other running applications. The bitrate information for the offline files is 'logged' by the Spotify application itself and that log can be read using Console.

    ReplyDelete
  88. Thanks, arjandr, wish I knew this before I started my test, it would save a lot of time!

    Tested this Hyperdub playlist, first track is already cached so can't tell. Out of last 23 tracks, 9 are 160 kbps. Looks like what we've done here and on the support forum are not in vain:)

    http://open.spotify.com/user/hyperdub/playlist/78QhXlbvgtiJYR8Zm5zCuJ

    ReplyDelete
  89. Using console, I recheck my 115-track 320k test playlist, five tracks already cached so can't tell. Out of the 110 remaining tracks, 60 of them are now at 320 kbps.

    And the pattern is clear, most popular tracks have been converted in the past few weeks since my initial test, or, more likely, since the original post was published. The relatively less popular tracks generated by Spotibot are still mostly at 160 kbps.

    I think Spotify is heading to the right direction, regarding this issue.

    But it doesn't mean the campaign is already over, after all, their commitment was to convert the whole library within a few months of June 2009.

    Look out, Spotify, now we can find out the bitrate of any track at ease;)

    ReplyDelete
  90. Hi Ulysses, thanks for all your hard work and diligent reporting on this "Bitrategate" issue. I was searching for "spotify artifacts" and finding not many useful results; now, finally, I've found discerning people who understand the problem.

    Let me assure you that before I ever heard of your blog I immediately noticed nasty artifacts in some songs on Spotify (the artifacts are unacceptable even for 160kbps, I might add). And upon upgrading to a premium account I was certainly disappointed that there was no noticeable improvement, and I wrote about this on my blog: http://www.mattmontag.com/music/weird-spotify-compression-artifacts.

    It should be clear this is not a "pet issue" of purist geeks that can be abrogated by hiding the bitrate of content. The artifacts are obvious! And it's too bad because Spotify has otherwise created an awesome service built on state-of-the-art networking technology.

    ReplyDelete
  91. BTW, great advice from arjandr. Any way to do something similar on Windows?

    ReplyDelete
  92. Sorry, yet another "Anonymous" poster to confuse matters.

    Well done on this Ulysses. I have been confused for some time as a Premium subscriber as to why some tracks seemed to stream in excellent quality and others not. I have had quite an education by reading through this conversation.

    Bottom line is that while Spotify is in my opinion an excellent service I feel that you are absolutely right to hold them to promises made and services advertised.

    The most disappointing aspect of Spotify is the way they simply go to ground when problems arise. A few years ago they launched with great fanfare in Ireland - premium service only mind. Then a few months later withdrew the service without any explanation. I have seen no explanation to this day why they withdrew. I should state that those of us lucky to have subscribed and kept up payments are still subscribers - "international subscribers" it appears.

    They seriously need to get themselves a good press officer or PR advisor as they are needlessly pissing people off in the way they handle legitimate queries.

    ReplyDelete
  93. Nice post! Any idea on how to get around the annoying "Appears On" issue? Well, if you have shuffle and repeat on and listen to a quite famous artist like "Norah Jones" you will not get her songs but istead all the crap on the "Appears On" albums. I did a test and for "Norah Jones" I got 19 subsequent songs that were not "Norah Jones"! Searching for "Artist:Norah Jones" doesn't work.

    ReplyDelete
  94. Two solutions:

    1, Save all official releases of the artist into a playlist. Then play shuffle from that playlist.

    2, On the artist page, press Ctrl (CMD) + F to bring out the filter bar, input artist name. Then play shuffle on that page.

    And... you commented on the wrong post:) The right one is here:

    http://www.spotifyclassical.com/2011/07/how-to-search-for-classical-music-on.html

    ReplyDelete
  95. This blog got noticed by CNN:

    http://www.cnn.com/2011/TECH/gaming.gadgets/08/12/htc.beats.audio/index.html

    Alkettory

    ReplyDelete
  96. For those who what to use this console trick on Windows, run spotify.exe with the -console command option.

    "c:\Program Files (x86)\Spotify\spotify.exe" -console

    or

    "c:\Program Files\Spotify\spotify.exe" -console

    depending on where you have Spotify installed. A small console window will appear where the bit rate messages will be logged.

    ReplyDelete
  97. Thank you for this post. I was considering upgrade to premium and now plan to postpone until more music is available in high quality. To me quality matters, not mobile access.

    ReplyDelete
  98. I will quit Spotify Premium, 10€ was not much but when you listen flac with foobar you just can't listen Spotify again,it's sad that they're lying about quality for Premium users.

    ReplyDelete
  99. Je viens de faire le test avec Spotify.console, merci pour l'astuce
    dans mon cas, 80% de mes playlist (principalement de la pop et du rock)sont encodées en 320kbps? Honnêtement j'ai du mal à faire la différence entre du 160 et du 320, pourtant j'ai une bonne installation hi-fi.
    Par contre la différence entre du 320 et du flac est très nette.

    ReplyDelete
  100. Thanks for exposing this scam. The technobabble about artefacts in the streams by other posters is completely irrelevant in this case.

    The fact is that when I subscribed for the Premium service, I expected to get q9 streams, based on Spotify's own claims. Basically I have paid for a service that I have not been getting and that annoys me. I hope this consumer pressure will force Spotify to act swiftly.

    I am seriously considering asking for a refund for the difference between the Premium and the regular subscription for the whole duration of my Premium subscription so far.

    ReplyDelete
  101. Je vous conseille de vérifier par vous même en ouvrant Spotify en mode console et en activant le mode "hors connexion"sur vos playlist , vous verrez le bitrate :

    "C:\Program Files\Spotify\spotify.exe"-console

    Presque toutes vos chansons sont encodées en 320kbps, tout va bien.

    ReplyDelete
  102. Premium users just got raped by Spotify, on now they are on the "Oh, I'm soooo sorry, didn't mean to" wagon.

    ReplyDelete
  103. Hi, About that console thing.. i understood how to do it. But its dosn't say anywhere what the kbps is. Just "ms". I'd be glad for responses :-)

    Cheers.

    ReplyDelete
  104. You need to save playlists to offline to see their bitrate. I posted a few screenshots of Console showing where the bitrate info is. Please search for Twitpic in this page.

    ReplyDelete
  105. Looks like your campaign had effect. All the tracks I've tried have had high bitrate Vorbis files for them. More surprisingly, suddenly (as of December 17th 2011) all music is available stored in LAME-encoded 320K MP3 as well as Vorbis.

    Not sure if this will be to allow a new client to play from MP3 or to support the purchase feature, which previously used a different mechanism. Mp3 play out would be a good idea - Vorbis works exceptionally well at 160K but there is little or no audible improvement at 320K, while LAME MP3 works well at higher bitrates. And yes, one can hear the difference between MP3 and Vorbis 320K.

    ReplyDelete