PDA

View Full Version : Logitech - please work with Pandora for 192kbps streaming!



jmpage2
2009-11-19, 14:21
I am a paying Pandora One account holder, which means that when I'm listening to Pandora on my computer I am treated to 192kbps audio that sounds notably better than 128kbps garbage.

Unfortunately this enhanced audio quality does not work on the Squeezebox platform.

I contacted Pandora support about this and they said that they have considered adding 192kbps audio streams to other devices but that it would also take work on the part of the device maker to support this as well.

So, if Logitech has any pull with Pandora whatsoever please try to get us 192kbps audio for the Squeeze products. I have three players and am considering a fourth but since Pandora is what I use most it's a real drag to listen to low quality garbage on my mid-fi and hi-fi setups.

andyg
2009-11-19, 14:26
The ball is in Pandora's court, we're waiting for an updated API.

jmpage2
2009-11-19, 14:52
Thanks Andy.

Maybe if more people complained to Pandora we could get this sooner rather than later.

JJZolx
2009-11-19, 15:14
Maybe if more people complained to Pandora we could get this sooner rather than later.

It's only going to come at the insistence of higher ups at Logitech. Not from one developer emailing another and asking "Dude, where's the new API?" and getting a response, "Dude, I'm swamped." My guess would be that the revenue being generated for Pandora from owners of Logitech products is pretty small, so 192k support is a pretty low priority.

garym
2009-11-19, 15:28
it's a real drag to listen to low quality garbage on my mid-fi and hi-fi setups.

Not to put too fine a point on it, but 128kbs is not exactly lo-quality garbage compared with 192. There is lots of evidence that most normal human beings can't ABX mp3 128 compared back to original CD in a blind test (visit hydrogenaudio forums for lots of info). That said, I'd love 192 simply because if it exists, I want it (same reason I listen to all my own files in lossless FLAC....but it is not because I think I can actually HEAR the difference between lossless FLAC and mp3--because I can't....in fact, I can't ABX lame mp3 at -V5 from original source on high-end equipment or good headphones except for a few known, rare samples.)

jmpage2
2009-11-19, 15:34
Not to put too fine a point on it, but 128kbs is not exactly lo-quality garbage compared with 192. There is lots of evidence that most normal human beings can't ABX mp3 128 compared back to original CD in a blind test (visit hydrogenaudio forums for lots of info). That said, I'd love 192 simply because if it exists, I want it (same reason I listen to all my own files in lossless FLAC....but it is not because I think I can actually HEAR the difference between lossless FLAC and mp3--because I can't....in fact, I can't ABX lame mp3 at -V5 from original source on high-end equipment or good headphones except for a few known, rare samples.)

I have consistently been able to pick up on the differences between even my --alt preset extreme LAME rips and the original CDs.

I have a modest sound system in my workout room that consists of an old (but very robust) Yamaha AVR and a pair of Polk budget speakers.

Numerous times when listening to Pandora (and busting my butt) I've noticed how flat, lifeless, distorted and dull the audio is. I will frequently then play this same track, from the same Squeezebox from my own library and the difference is extremely noticeable.

In fact I would say that it would be hard for someone NOT to be able to tell the difference.

garym
2009-11-19, 17:10
There's plenty of discussion of this on the web by very informed folks (in this forum and other places), so I don't want to turn this thread into a repeat of all that info. You may have amazing hearing, but there is lots of scientific evidence that clearly indicates that it is fairly rare for any individual to be able to ABX in a blind test the difference between a properly encoded 128 mp3 with the original CD. If you were posting over at Hydrogenaudio.org the folks would demand that you post your blind ABX test results to back up your statement about detecting differences (one of their terms of service). Of course, we're not over there and things are much looser here.

I can't speak to streams from pandora as there are so many more variables in the mix other than just the actual source file (and not sure how to easily do an ABX test with pandora stream since you don't have the source file in another format). Keep in mind that even extremely small differences in volume can make one source sound better than another (again, lots of hard evidence on this).

Bottom line, Pandora may very well not sound as good as your local files, but in general, 128 (created with a modern, quality encoder) is typically transparent for most users on most music. Your statement that the reverse is true flies in the face of lots of controlled experiments. I'm not trying to pick a fight by the way, I just hate to see myths about lossy encoding quality be perpetuated. That said, there are lots of reason to rip music to FLAC (or other lossless format), and I do so with all my own CDs.

ajkidle
2009-11-19, 17:39
It's only going to come at the insistence of higher ups at Logitech. Not from one developer emailing another and asking "Dude, where's the new API?" and getting a response, "Dude, I'm swamped."

Agreed. I thought the Squeezebox guys and the Pandora guys were tight, but maybe that relationship isn't what it once was. Either way, streaming 192 kbps would be a great enhancement to the Squeezebox experience.

-Stef-
2009-11-20, 10:28
I am more concerned about the tightness of the Logitech guys and the Pandora guys, but I agree that 192kbs would be an enhancement.

Especially when Pandora will be available for Dutch citizens. :(

chroma
2009-11-21, 17:11
more support for the appropriate folks to get in touch and make it happen. if pandora worked at 192 on my squeezeboxes, i would listen much more often. and am already a paying subscriber, but barely ever use it because the quality is so low.

jmpage2
2009-11-21, 23:18
There's plenty of discussion of this on the web by very informed folks (in this forum and other places), so I don't want to turn this thread into a repeat of all that info. You may have amazing hearing, but there is lots of scientific evidence that clearly indicates that it is fairly rare for any individual to be able to ABX in a blind test the difference between a properly encoded 128 mp3 with the original CD. If you were posting over at Hydrogenaudio.org the folks would demand that you post your blind ABX test results to back up your statement about detecting differences (one of their terms of service). Of course, we're not over there and things are much looser here.

I can't speak to streams from pandora as there are so many more variables in the mix other than just the actual source file (and not sure how to easily do an ABX test with pandora stream since you don't have the source file in another format). Keep in mind that even extremely small differences in volume can make one source sound better than another (again, lots of hard evidence on this).

Bottom line, Pandora may very well not sound as good as your local files, but in general, 128 (created with a modern, quality encoder) is typically transparent for most users on most music. Your statement that the reverse is true flies in the face of lots of controlled experiments. I'm not trying to pick a fight by the way, I just hate to see myths about lossy encoding quality be perpetuated. That said, there are lots of reason to rip music to FLAC (or other lossless format), and I do so with all my own CDs.

I agree with you that for the 'average joe' who thinks that an AM radio sounds "great" there is no need to go higher than 128kbps.

However....

It doesn't take amazing hearing to be able to tell the difference between 128 CBR MP3 and VBR MP3 that are both LAME encoded. All it takes is some musical knowledge and decent equipment.

Most of the listening test I've seen that show that users can't pick apart 128kbps vs 192kbps mp3 are usually performed using $5 computer speakers. Color me shocked on those numbers.

You might be mixing things up here and imply that 128 AAC (which is pretty much superior to LAME at the same bitrate) is "good enough" for most users.

All I can tell you is that to me it's pretty evident which I am listening to, at least with the Pandora service (and who knows how the music is encoded). They obviously offer the 192 kbps streaming for their paying customers for some reason, I doubt they would spend the money on placebo affect.

andynormancx
2009-11-22, 03:42
They obviously offer the 192 kbps streaming for their paying customers for some reason, I doubt they would spend the money on placebo affect.
Of course they might. Imagine a consumer who knows nothing about sound quality and is listening on $5 computer speakers. If they are thinking of signing up for a music service and they have a choice of one service that advertises the number 192 whereas another one uses the number 128, which one do you think they are most likely to choose ?

It doesn't matter if they don't know anything about bitrates or the fact that the 128k service could sound as good as the 192k service depending on the encoder used. They are still likely to be swayed by the bigger number.

garym
2009-11-22, 06:44
No, I'm talking about very serious BLIND ABX listening tests with high quality equipment conducted by serious folks. I'm not talking about average joes happy with AM radio. And I'm not saying that no one can detect the difference between an mp3 128 and original source, just that it typically takes very good hearing and songs where potential artifacts are likely. These are not the typical case for most listening. If you really want to learn more about this sort of thing go over to hydrogenaudio.org and read the material on this issue and the many forum posts. But I'm getting the idea that you're not actually interested in learning anything that changes your initial beliefs and you're happy to accept the heuristics and biases that may color your beliefs. That's ok--to each their own. We're just listening to music, not curing cancer. And as I said, I personally prefer all my music in FLAC (lossless) if at possible (even streams if they were available)....but not because I think I can hear any difference (in my library of over 100,000 songs I've heard less than a dozen songs that were not transparent at 128).

garym
2009-11-22, 06:50
It doesn't matter if they don't know anything about bitrates or the fact that the 128k service could sound as good as the 192k service depending on the encoder used. They are still likely to be swayed by the bigger number.

Agree! And this affects the average listener (who could care less as long as they can download the song and it plays on their nano) as well as the "audiophile" that pays $3,000 for a single 3 foot audio cable because it adds more "air" to the music.

tomjtx
2009-11-22, 07:32
No, I'm talking about very serious BLIND ABX listening tests with high quality equipment conducted by serious folks. I'm not talking about average joes happy with AM radio. And I'm not saying that no one can detect the difference between an mp3 128 and original source, just that it typically takes very good hearing and songs where potential artifacts are likely. These are not the typical case for most listening. If you really want to learn more about this sort of thing go over to hydrogenaudio.org and read the material on this issue and the many forum posts. But I'm getting the idea that you're not actually interested in learning anything that changes your initial beliefs and you're happy to accept the heuristics and biases that may color your beliefs. That's ok--to each their own. We're just listening to music, not curing cancer. And as I said, I personally prefer all my music in FLAC (lossless) if at possible (even streams if they were available)....but not because I think I can hear any difference (in my library of over 100,000 songs I've heard less than a dozen songs that were not transparent at 128).

I have done quite a few blind tests that I have distinguished between 128 and 192 or higher listening to well recorded classical and jazz. Most commercial pop recordings are so bad that it doesn't seem to make much diff.

It takes an experienced listener with a well trained ear to hear the difference.

At 128 the soundstage is flatter, less 3 dimensional. There is less hall ambience which is so helpful in creating spacial cues. There is less air around instruments etc.

192VBR seems to be the starting point for adequate hi-fi. It is much harder to hear differences above that rate, although I have been able to do that on some recordings.

At 128 it is actually very easy to hear the diffs blind on well recorded music especially if you are very familiar with the recording.

One should remember that any listening test is testing the listeners "ear training" as well as the bit rate.
Not everyone will hear the diffs. However the majority of people can be trained to hear those diffs, as proven through years of musicians improving their aural skills in ear training classes.

upstateaudio
2009-11-22, 07:38
I emailed Pandora regarding 192kbps and received a prompt response that they are a small organization and need to allocate their resources as such. They did thank me for the feedback. Perhaps if enough of us contact Pandora, it would persuade them to devote some resources to this project.

garym
2009-11-22, 07:49
I have done quite a few blind tests that I have distinguished between 128 and 192 or higher listening to well recorded classical and jazz. Most commercial pop recordings are so bad that it doesn't seem to make much diff.

It takes an experienced listener with a well trained ear to hear the difference.

At 128 the soundstage is flatter, less 3 dimensional. There is less hall ambience which is so helpful in creating spacial cues. There is less air around instruments etc.

192VBR seems to be the starting point for adequate hi-fi. It is much harder to hear differences above that rate, although I have been able to do that on some recordings.

At 128 it is actually very easy to hear the diffs blind on well recorded music especially if you are very familiar with the recording.

One should remember that any listening test is testing the listeners "ear training" as well as the bit rate.
Not everyone will hear the diffs. However the majority of people can be trained to hear those diffs, as proven through years of musicians improving their aural skills in ear training classes.

agree on all your points (in particular the "ear training"...with some basic training to listen for mp3 artifacts, you can train yourself to detect differences on even some of highest bitrate mp3 files). And in fact for lossy files (used on my IPODs) I actually convert to mp3 192 from my FLAC files just because the possibility of artifacts is much reduced at that level. My only point in all these postings is that the initial statement that inferred that just about anyone with just about any song should be able to easily hear differences in 128 vs the original CD is factually wrong and perpetuates certain myths.

tomjtx
2009-11-22, 08:27
agree on all your points (in particular the "ear training"...with some basic training to listen for mp3 artifacts, you can train yourself to detect differences on even some of highest bitrate mp3 files). And in fact for lossy files (used on my IPODs) I actually convert to mp3 192 from my FLAC files just because the possibility of artifacts is much reduced at that level. My only point in all these postings is that the initial statement that inferred that just about anyone with just about any song should be able to easily hear differences in 128 vs the original CD is factually wrong and perpetuates certain myths.

Oh No ! It sounds like we are in complete agreement...............Oh wait.........this isn't the audiophile section so it's OK if we agree :-)

For me it becomes very difficult to hear diffs at 256VBR which is what I use for my Itouch.
My 160GB Ipod classic has all lossless but with 160gb it's no prob.

To the OP, Rhapsody streams at 192VBR through SB or TP so you might want to consider that.
Their download file for purchase are at 256VBR.

toby10
2009-11-22, 09:00
As has been suggested, how well the file has been encoded matters as much as the bitrate. I hear many 128k stations that do not sound as good as some 96k stations, and I'm guessing the encoding method is likely the main difference. As a general rule (and as I do) searching for the higher bitrate streams will usually render better overall SQ, but it is not true 100% of the time, unfortunately. :)

pablolie
2009-11-22, 09:49
For the sake of the OP, let me say that I fully agree and would love to get the 192k version of Pandora on my multiple SB equipment.

On the subject of increased quality, any well recorded jazz or classical album will truly allow one to hear the difference between 128k MP3 and the original. The difference between 256k/320k and the original will be much harder to detect unless you are very familiar with the original and know your MP3 artifacts (typically a hissy nature in the highs, sometimes even pre-echo and similar), but it can be detected. But, again, it takes a good recording and familiarity with it.

From listening to Pandora 128k and 192k streams, I can attest to the fact the 192k stream seems less tiring. Mind you, the 128k channel sounds great for an Internet radio channel - but if we can get more, why not? It is in the nature of the digital consumer to demand a lot for free, and the impossible if we pay a subscription - add that to the cluetrain manifesto. :-D

jmpage2
2009-11-22, 10:03
No, I'm talking about very serious BLIND ABX listening tests with high quality equipment conducted by serious folks. I'm not talking about average joes happy with AM radio. And I'm not saying that no one can detect the difference between an mp3 128 and original source, just that it typically takes very good hearing and songs where potential artifacts are likely. These are not the typical case for most listening. If you really want to learn more about this sort of thing go over to hydrogenaudio.org and read the material on this issue and the many forum posts. But I'm getting the idea that you're not actually interested in learning anything that changes your initial beliefs and you're happy to accept the heuristics and biases that may color your beliefs. That's ok--to each their own. We're just listening to music, not curing cancer. And as I said, I personally prefer all my music in FLAC (lossless) if at possible (even streams if they were available)....but not because I think I can hear any difference (in my library of over 100,000 songs I've heard less than a dozen songs that were not transparent at 128).

Please get off the pedestal. I know all about hydrogen audio and have been using the LAME encoder and the Fraunhofer encoder before that. I've been encoding my music for over twelve years and have participated in numerous ABX tests that demonstrate that I usually can pick the higher quality sample (at least between 128kbps and source, it is much harder with -alt preset extreme recorded material to tell the difference).

It also might interest you to know that while I don't purport to be an audiophile (self purported audiophiles with their CD stabilizer rings and $600 power cables tend to creep me out), I used to sell extremely high end stereo gear. Listening to some top of the line Martin Logan speakers powered by Krell amps and high end DACs will go a ways towards getting you comfortable with what is possible in audio and you can set your expectations from that.

The entire point to this thread before you insisted on trying to derail it with these arguments is that Pandora paid subscribers are already paying for the higher quality audio, and we should be able to enjoy it on as many devices as possible.

To any other Pandora One subscribers who would be interested in obtaining the higher quality streams for the Squeezebox, I would advise you to email support@pandora.com and politely request this feature.

Maybe if enough people pester them about it they will actually put it on the road map for 2010.

dts
2010-01-18, 12:44
I emailed Pandora regarding 192kbps and received a prompt response that they are a small organization and need to allocate their resources as such. They did thank me for the feedback. Perhaps if enough of us contact Pandora, it would persuade them to devote some resources to this project.

Just emailed Pandora.

DaveWr
2010-01-18, 13:04
Why you are already a subscriber, what do they gain?

Dave

pablolie
2010-01-20, 06:04
Why you are already a subscriber, what do they gain?

Dave

Happier customers. A higher guarantee of renewed subscriptions. Offering a better, more consistent service (what else are they going to do, offer FLAC streams but only to browser based clients?). :-)

I think it compelling enough: customers ask for it, plus more music with very good quality (which 192 represents, 128 is a tad bland for good jazz and classical recordings) to more clients... sounds like a common sense bet for them.

DaveWr
2010-01-20, 07:07
I agree with your second statement, that indeed would affect revenues. However trying to sort their international side might provide significantly greater opportunities. i believe licensing is now slightly better.

On a more positive note looks like AndyG has been on Twitter with their developers, hinting at where is the API.

Dave

maggior
2010-01-20, 07:41
However the majority of people can be trained to hear those diffs, as proven through years of musicians improving their aural skills in ear training classes.

Sometimes ignorance is bliss :-). It's for this reason that I've never done serious ABX testing.

On my home setup, I used 190 kbs VBR LAME files initially. When I upgraded it to FLAC, I noticed a subtle yet very real difference with my headphones. The music sounded fuller with more presence.

On my iPod, I use LAME mp3 files. I use this mostly in my car. Yes, I could train my ear to hear the difference in my car, but I don't want to :-). Given the noisy environment of a car, it would be difficult to train my ear there anyway.

bobkoure
2010-01-20, 21:00
...what else are they going to do, offer FLAC streams but only to browser based clients?). :-)
Actually, what I want, as a paid subscriber, is the ability to use their "Music DNS" with my own music, the way I use MusicIP.
I realize that this could be a big development job for them - but then they have a product that produces income without having to pay licensing fees - and there'd be less bandwidth required so it might pay for itself.
It'd also be something they could provide into all markets, even when they hadn't yet negotiated licensing agreements so they could start marketing pre licensing.

I'm guessing that track identification is where the bulk of the work would be. I think there are already some music fingerprinting algorithms out there - but they could start out just assuming tag info is correct and go from there. Maybe they could work a deal with MusicBrainz or the current owner of MusicIP.

Once they had something like this, I'd hope to also be able to say "I'm looking for new music like this track/artist/album - don't play anything already in my library".
And I'd love to be able to switch between a MusicIP mix and a Pandora mix.

Fleury
2010-02-10, 16:02
Rather than wait for Pandora or Logitech to blink, try this work around - Use the Pandora One desktop applet (a very nice impovement over Open Pandora). Because it's coming off your PC and not SqueezeCenter your Pandora stream should be 192. Pipe it to the SB via your wave input entry in "Favorites", using the WaveInput plugin (a pain to configure, but worth the effort):

http://wiki.slimdevices.com/index.php/WaveInput_plugin

This should take the Pandora stream right off your sound card, with whatever quality level you have set for wav in your SC/SB Settings - enjoy.