Spotify - normalization vs dynamic range compression

[TrevC]

Hmmm. Spotify. Buy a Ferrari then fill it's tank with paraffin.

Or play records on a turntable. Would you put paraffin in a cheap car?

 

[steve_1979]

it's the optimisation of the time constants in order to minimise the audibility of the automatic volume control that is the important thing.

Could you expand on that a bit please?

I'm talking about the amount of time it takes for the levels to be auto adjusted. For streaming like spotify with no overmodulation issues I favour a long time constant for both increasing and reducing the volume, which would preserve most of the dynamic range in the music. The downside is that a loud track would take a long time to go down and vice versa.

Whether it works fast or slow the to auto adjusted dynamic range compression it's still going to end up compressing the size of the largest wave peaks down to the size of the meduim sized wave peaks. That's what dynamic range compession means. The wave shape of the largest peaks will be changed from their original form. The ratio of the waves in relation to each other are changed from thier original ratios. This dynamic range compression method effect sucks the life out of music and it sounds awful.

Normalisation works much better because just lowers the volume of everything equally without effecting the shape of the wave in any other way. The ratio and shape of the waves in relation to each other are totally uneffected by normalisation. The resulting sound using normalisation stays exactly the same as the original and is totally unchanged (apart from being quieter obviously).

 

[TrevC]

it's the optimisation of the time constants in order to minimise the audibility of the automatic volume control that is the important thing.

Could you expand on that a bit please?

I'm talking about the amount of time it takes for the levels to be auto adjusted. For streaming like spotify with no overmodulation issues I favour a long time constant for both increasing and reducing the volume, which would preserve most of the dynamic range in the music. The downside is that a loud track would take a long time to go down and vice versa.

Whether it works fast or slow the to auto adjusted dynamic range compression it's still going to end up compressing the size of the largest wave peaks down to the size of the meduim sized wave peaks. That's what dynamic range compession means. The wave shape of the largest peaks will be changed from their original form. The ratio of the waves in relation to each other are changed from thier original ratios. This dynamic range compression method effect sucks the life out of music and it sounds awful.

Normalisation works much better because just lowers the volume of everything equally without effecting the shape of the wave in any other way. The ratio and shape of the waves in relation to each other are totally uneffected by normalisation. The resulting sound using normalisation stays exactly the same as the original and is totally unchanged (apart from being quieter obviously).

Sounds perfect. How can it be achieved? In a player, it can't.

 

[steve_1979]

How can it be achieved? In a player, it can't.

Why do you think that?

 

[steve_1979]

How can it be achieved? In a player, it can't.

See post 6 on page 1.

 

[BigH]

I agree, I have tried it and it is excellent within it's limits on sound quality. I listen to internet radio at 192&320kbps streams and they can sound quite good too. But this is a suposedly a "HiFi" forum that endlessly debates often tiny improvements in their hardware, only to feed it a diet of compressed recordings. Regarding the "music" favoured by my daughters, it would make little difference as it has been compressed at source and delivered to the ipod generation with no regard for those who own proper HiFi. A SACD arrived today and I compared the CD layer on the fly. I feel sorry for those on here that insist that they can't hear the difference, because there clearly is one. An endless debate and ulimately futile. If anyone is interested, over on Sterophile.com there is an excellent technical report on the launch of Meridian's MQA. Perhaps renowned digital innovator, Mr Bob Stuart, has been wasting his time developing this exciting new approach to HiFi streaming.

The sacd is often different mastering from the cd layer, so its not always a good comparison. Many can't hear the difference between 320kbps and cds, even less can hear difference between 16 bit and 24 bit.

As for Stereophile, they promote hifi so i would not take too much notice, maybe it was the active speakers that blow his socks off?

 

[fr0g]

I agree, I have tried it and it is excellent within it's limits on sound quality. I listen to internet radio at 192&320kbps streams and they can sound quite good too. But this is a suposedly a "HiFi" forum that endlessly debates often tiny improvements in their hardware, only to feed it a diet of compressed recordings. Regarding the "music" favoured by my daughters, it would make little difference as it has been compressed at source and delivered to the ipod generation with no regard for those who own proper HiFi. A SACD arrived today and I compared the CD layer on the fly. I feel sorry for those on here that insist that they can't hear the difference, because there clearly is one. An endless debate and ulimately futile. If anyone is interested, over on Sterophile.com there is an excellent technical report on the launch of Meridian's MQA. Perhaps renowned digital innovator, Mr Bob Stuart, has been wasting his time developing this exciting new approach to HiFi streaming.

For someone with the name you have and the tone you are taking, I can only assume you claim to see the beautiful suit.

I would bet my house that you could not tell apart a 320 Kbps recording from an HD track.

Delusion is commonplace, so no shame.

 

[busb]

it's the optimisation of the time constants in order to minimise the audibility of the automatic volume control that is the important thing.

Could you expand on that a bit please?

I'm talking about the amount of time it takes for the levels to be auto adjusted. For streaming like spotify with no overmodulation issues I favour a long time constant for both increasing and reducing the volume, which would preserve most of the dynamic range in the music. The downside is that a loud track would take a long time to go down and vice versa.

And there lies the problem with compression - it's almost impossible to get the right time constants for all types of music. What bothers me more than the concept of compression is the often horrid amplitude distortion artefacts that to my ears are just horrid. Maybe I'm particularly sensitive to it.

The problem is easly solved technically - just have every device aimed at punters to use compression as a default thus letting the minority who care, to switch it off - then produce all music largely uncompressed. Compression has its place - in cars or listening in noisy places like public transport, preferably with some control over the degree.

One answer is to get youth to listen to decent systems or the Hi Fi industry will shrink to point of being priced beyond most that would have the knock-on effect that uncompressed music would have too small a market share for anyone to bother producing it.

 

[steve_1979]

I've just noticed that with the new Spotify update yesterday they've now completely removed the option to 'play all tracks at the same volume' so I've retested the Bjork track to see if Spotify is forcing us to use dynamic range compression whether we want it or not...

...and the result is good (albeit very suprising!)

Spotify now seems to have more dynamic than it did originally with either the 'play all tracks at the same volume' option switched on or off so that's an improvement. The tracks also seem to have had their volume adjusted to that the peak reaches 0dB without clipping.

Here's the results to compare to those in post 1 of this thread.

There are several other comments and blogs around the web saying similar thing to what I've found. It seems that Spotify have been listening to their customers. *biggrin*

 

[peterpiper]

maybe an email to spotify to ask why they have reduced the qaulity, and show them your results, wonder what their excuse will be, another here considering leaving now, I 'upgraded' to the latest version and though it sounded a touch grainier and slighty less 'air' now I know why,

 

[steve_1979]

Hmmm. There's something odd going on here.
I've just deleted the Spotify cache and reinstalled the previous version of Spotify (0.9.15.27) and the result is the same as it was with the new one. TBH I'm a bit confused by these results now. Maybe they've just changed the version of the track 'Radio Protector' to a different one with different mastering? I'm not sure what to make of this. I'll investigate some more tomorrow but it's getting late now.

 

[andyjm]

I have always thought it a shame that providers don't include a test track or similar to allow users and installers to setup, calibrate and test. A series of tones at various known levels would allow a lot of us to accurately set up and align our equipment. It would also put a lot of these threads to bed as it would be clear what is going on with their processing. Perhaps they don't want to wash their dirty underwear in public.

Sky had a series of HD test card transmissions hosted by Myleene Klass (why?) which were great for setting up a Sky HD box. I recorded the transmission, but unfortunately lost it when my HD box packed up and was replaced. It would take very little effort for a Sky box to generate a basic set of test patterns. I guess the 3p they save per box by not including this adds up.

 

[peterpiper]

I just backgraded to an earlier version again , I have the 'set the same volume' option again, but now i dont know if it sounds better again , maybe its pyscological knowing the frequency range *might* be back up to 20k how do I know? *scratch_one-s_head*

 

[peterpiper]

I will be interested in your results steve, as will other spotify users with a discerning ear no doubt

 

[steve_1979]

I have always thought it a shame that providers don't include a test track or similar to allow users and installers to setup, calibrate and test...

That would be handy.

 

[steve_1979]

It's seems that I may have spoke too soon yesterday.

Today I remembered that we have another computer in the house with Spotify on it so I booted it up and sure enough it still has the old version of Spotify installed so I've been able to do some direct comparisons and the results show a big drop in sound quality!

Using the track 'Radio Protector' here are the results.

The right pictures shows the old version Spotify with the 'Set the same volume level for all tracks' option turned off.

EDIT - Images removed because I made a mistake.

 

[fr0g]

I just did a 7 day trial to TIDAL (Which I now realise is basically WiMP which I have used before).

Through my headphones I am struggling to say if there is any difference between that (on Hifi mode) and Spotify.

To be honest Steve, I am past caring. Both sound great, one is cheaper and with better functionality. The other is pandering to those people who actually believe they can tell a 320 Kbps MP3 from lossless (which they can't).

 

[TrevC]

My Spotify has automatically updated to the new version. It passes the Pink Floyd 'Shine on Crazy Diamond' test with flying colours, where the previous version came on too loud at the beginning and cut down when the drums started with the 'play tracks at the same volume' enabled.

Keep calm and use Spotify.

Btw I'm currently listening to WUMB Boston via the internet at 128k. That sounds easily as good as FM to me.

 

[fr0g]

Btw I'm currently listening to WUMB Boston via the internet at 128k. That sounds easily as good as FM to me.

I am not surprised. The current LAME codec is excellent. While I can occaisionally ABX 128K, it isn't 100% and it's surprisingly transparent...

Try this and see if you can differentiate 128 and 320...dare you!

http://mp3ornot.com/

 

[wilro15]

Probably a stupid question, but does this affect just the PC app? Or would it also affect iOS, Android, Spotify Connect, etc users?

 

[MajorFubar]

Steve your detective work with Spotify is to be applauded, but don't ever lose sight of the fact that Spotify is first and foremost a consumer-grade streaming service with no pretentions of super high fidelity. When its developers ****-about with the settings and codecs behind the scenes, someone like you (or collectively us) couldn't be more distant from their thoughts. Digital streaming services like Spotify only sound as good as they do because with digital audio you really have to try incredibily hard to fcuk it up to the point of it becoming unlistenable. The acceptable quality they achieve is a happy accidental consequence of modern digital file-compression and distribution. They don't have to try very hard to get the SQ acceptably right, and there's no audiophile in the back office QA'ing their streams through range-topping amps and speakers that cost as much as a Ferrari.

Interesting observation though that I'll throw in the mix: the latest Mac version still retains the check-box to disable volume normalisation, via Preferences > Playback, directly under 'crossfade tracks'.

 

[MajorFubar]

Probably a stupid question, but does this affect just the PC app? Or would it also affect iOS, Android, Spotify Connect, etc users?

Interesting observation though that I'll throw in the mix: the latest Mac version still retains the check-box to disable volume normalisation, via Preferences > Playback, directly under 'crossfade tracks'.

I haven't tried with a mobile device.

 

[davedotco]

Probably a stupid question, but does this affect just the PC app? Or would it also affect iOS, Android, Spotify Connect, etc users?

Interesting observation though that I'll throw in the mix: the latest Mac version still retains the check-box to disable volume normalisation, via Preferences > Playback, directly under 'crossfade tracks'.

I haven't tried with a mobile device.

My OSX desktop client updated yesterday and the checkbox has gone.

Similarly my iOS apps updated two/three days ago, gone from there too.

 

[iQ Speakers]

I have to say Steve my hat is off to you for providing some brilliant information, wish I understood it all! I have tried Tidal and Qobuz I wish I could say they sound better, they don't hit me instantly as sounding better. The biggest issue is the albums for me are just not there unlike on Spotify.

 

[davedotco]

I have to say Steve my hat is off to you for providing some brilliant information, wish I understood it all! I have tried Tidal and Qobuz I wish I could say they sound better, they don't hit me instantly as sounding better. The biggest issue is the albums for me are just not there unlike on Spotify.

I have always found Spotify to be very slightly dull, possibly something to do with the way it is streamed which seems way more complex than I can understand.

I really discovered this when comparing to uncompressed Qobus, which sounded slightly more 'focused' and precise. A tiny touch of treble lift in the Spotify eq seemed to make all the difference, just a touch mind, the 'Treble lift' preset is far to much.