What's wrong with commercial recordings

[semente]

This week I was zapping through the radio stations on my way home and this band was playing live from the BBC studios.

The vocals could do with a bit reverb but it sounds much more "lively" than most commercial recordings, crisper, better balanced, like being in a club:

 

http://www.bbc.co.uk/programmes/p05xjkzm

 

Why don't most commercial recordings sound as good as this?

Who's to blame, the producers?

 

 

[semente]

Here's another clip from a different session:

 

 

[Ralf11]

[jasenj1]

Don't know if this was a rhetorical question along the lines of "What's wrong with kids these days?" or if you're looking for serious answers/speculation, but I'll throw some input in.

 

Quote

Why don't most commercial recordings sound as good as this?

 

I played the video through MusicScope. The peaks are at -6dB. The loudness range (similar to dynamic range) is 10.0 - which is a pretty high value; this indicates the song has a mix of loud & soft passages, and even wide volume range within a passage. There is definitely some hard limiting at -6dB, but it is mostly the snare, kick, & bass that hit the limiter.

 

Compare that with Bruno Mars' "That's What I Like" from the album 24K Magic (winner of the 2018 Grammy for Best Engineered Album, Non-Classical): The peaks are at 0dB - all the way to the top. MusicScope says the "True Peak Level" is even higher than zero (this means after the DA conversion the signal may distort the analog stage; i.e. it's really loud). The LRA is 3.6 - pretty tight. And it's not just instantaneous signals like drums that hit the peaks; the vocals sit at 0dB most of the time, too. IMHO, this track is designed to be listened to in headphones in a gym or in a car; i.e. It is designed to block out ambient sounds - it is pretty white noise.

 

Also, the Bruno Mars track is not a live performance. It is assembled from individual, isolated recordings (stems), and heavily edited. It wasn't played, it was built.

 

Quote

Who's to blame, the producers?

 

If you are an this site, I expect you are aware of "The Loudness Wars". Music is no longer something you are expected to sit down & listen to in a controlled (read quiet) environment. It is something you listen to while going about your day. It must compete with all the other ambient noise going on around you. Therefore it is mixed loud & stays loud - you control the volume with a button. It's a culture problem. I don't think any one group can be blamed.

 

 

[Black]

The Loudness Wars are a bit silly and with some present indie bands they even make fun of it by doing it so LoFi and non-polished it's gotten critic's attention. Electric Studio's Forum (Steve Albini's Studio) has some hilarious but valid pro's and con's. Unless you're listening to "most" music from a reputable label, Production/Recording/Engineering staff it is what it is. It's hard to enjoy music if you know what goes behind it, or you're not hearing a tone stretch out due to it being cut off on the note to ether. That sucks. I hate it. But I also grew up on indie bands on labels that recorded like Pavement's original release of Slanted and Enchanted. I hated Nirvana's Nevermind because it was not the band's sound. It was more glossy than Appetite For Destruction.

[Rexp]

I doubt its the loudness wars to blame, more likely the AAC 320 the bbc use to stream just sounds better than flac/wav. 

[diecaster]

Sure, lossy AAC 320 sounds sounds better than lossless FLAC/WAV. Not in any reality I am aware of is that the case.....

[Rexp]

19 minutes ago, diecaster said:

Sure, lossy AAC 320 sounds sounds better than lossless FLAC/WAV. Not in any reality I am aware of is that the case.....

Sound quality is subjective isn't it? 

[Rexp]

Some of these bbc sessions seem to be streamed at 1700kbps, sound petty good too. 

[diecaster]

12 hours ago, Rexp said:

Sound quality is subjective isn't it? 

Sure. Please explain how a FLAC that would be a source of an AAC320 file would sound wise than the AAC320 file. I am really looking forward to how you explain that having less bits with a lower sample rate could sound better.

[Rexp]

54 minutes ago, diecaster said:

Sure. Please explain how a FLAC that would be a source of an AAC320 file would sound wise than the AAC320 file. I am really looking forward to how you explain that having less bits with a lower sample rate could sound better.

The AAC seems to filter out the objectional noise (bit like vinyl) that I hear in the original FLAC. Of course AAC aint perfect but I find it preferable. I am analog guy that thinks most digital playback sucks. Useful feedback anyone? 

[mkrzych]

It's been said many time about where the modern record production and mastering is going on, but what catching my eye also recently is when you go to DR Database you could see that for instance most of the 90s recordings where very dynamic and recent so called re-masters are (mostly!) shit! Of course, you could argue that's not only about the dynamic range, but still why to compress them so much!? Let's say from 11dB in average to 6dB average! This is insane and very stupid! You are getting 24/192 "re-master", when its originals in redbook quality sounds way better for less money, c'mon. 

 

Oh, it's about remasters, but almost all new releases how are still so much compressed. Don't even bother to mention them :-/

 

Please read this article as well if not pasted somewhere here: https://www.nytimes.com/2019/02/07/opinion/what-these-grammy-songs-tell-us-about-the-loudness-wars.html especially comments.

 

This is also good read: https://www.electricalaudio.com/phpBB3/viewtopic.php?f=4&t=68607

[mkrzych]

On 2/23/2018 at 4:58 PM, jasenj1 said:

Don't know if this was a rhetorical question along the lines of "What's wrong with kids these days?" or if you're looking for serious answers/speculation, but I'll throw some input in.

 

 

I played the video through MusicScope. The peaks are at -6dB. The loudness range (similar to dynamic range) is 10.0 - which is a pretty high value; this indicates the song has a mix of loud & soft passages, and even wide volume range within a passage. There is definitely some hard limiting at -6dB, but it is mostly the snare, kick, & bass that hit the limiter.

 

Compare that with Bruno Mars' "That's What I Like" from the album 24K Magic (winner of the 2018 Grammy for Best Engineered Album, Non-Classical): The peaks are at 0dB - all the way to the top. MusicScope says the "True Peak Level" is even higher than zero (this means after the DA conversion the signal may distort the analog stage; i.e. it's really loud). The LRA is 3.6 - pretty tight. And it's not just instantaneous signals like drums that hit the peaks; the vocals sit at 0dB most of the time, too. IMHO, this track is designed to be listened to in headphones in a gym or in a car; i.e. It is designed to block out ambient sounds - it is pretty white noise.

 

Also, the Bruno Mars track is not a live performance. It is assembled from individual, isolated recordings (stems), and heavily edited. It wasn't played, it was built.

 

 

If you are an this site, I expect you are aware of "The Loudness Wars". Music is no longer something you are expected to sit down & listen to in a controlled (read quiet) environment. It is something you listen to while going about your day. It must compete with all the other ambient noise going on around you. Therefore it is mixed loud & stays loud - you control the volume with a button. It's a culture problem. I don't think any one group can be blamed.

 

 

 

If I see this level of edition and compression is getting Best Engineered Album statue, I would like to vomit! It's insane, stupid and non-professional - ok, it's pure commerce :-(

[mkrzych]

This is slightly off topic, but watch this out till the end :-) 

 

[John Dyson]

On 2/23/2018 at 10:58 AM, jasenj1 said:

 

 

If you are an this site, I expect you are aware of "The Loudness Wars". Music is no longer something you are expected to sit down & listen to in a controlled (read quiet) environment. It is something you listen to while going about your day. It must compete with all the other ambient noise going on around you. Therefore it is mixed loud & stays loud - you control the volume with a button. It's a culture problem. I don't think any one group can be blamed.

 

 

There is a distribution architecture problem with older recordings (before the early 1990s') where apparently they were digitized without DolbyA decoding.  (Take a look at the various established digitization procedures -- DA decoding is sometimes not in the process.)  There is no GOOD software DolbyA decoder until recently and even then, the amount of CPU/disk to do it super well might have been pretty expensive before yr2000 or so.  A lot of material DOES languish in digital form without being DA decoded.  (It was common practice to simply copy analog DA encoded to another analog deck/bypassing the encode & decode cycle because of quality loss and complexity issues.)  The process for copying from DA encoded to digital was probably similar (back in the 'day', it wouldn't have been unreasonable for the DA decoding process be deferred just like on analog-analog transfers.)  Nowadays, vinyl is still probably done correctly when DA encoded, because it was always part of the 'procedure'.  With digital (making CDs, preparing downloads), that DolbyA step has been ignored because of the complexity, trouble, avoiding the use of real recording engineers (or using newer, less experienced recording engineers), or WORSE YET -- the choice of Beancounters avoiding the cost.

 

Playing out old digital tapes is MUCH faster than pushing the recording through an old DolbyA unit that works only in realtime.  Additionally, there is the matter of the HW setup, etc.  'Beancounters' control the world, and any dime saved in the process is worthwhile, because the public will purchase the defective recordings anyway.

 

So, a lot (I mean, really a lot) of the older recordings are leaked DolbyA with some EQ (-3dB@3k/Q=0.707 or -6dB@3k/Q=0.707), and call it 'listenable.'   Well, it is listenable if you are used to the heavy compression typically used in today's recordings.  However, the result of such 'decoding' is NOT the original album sound.  Maybe a few earlier CDs were made correctly -- back when there were engineers regularly using DolbyA.  Not any more.

 

THAT is where some of the 'harsh' digital sound has come from.

 

John

[mkrzych]

43 minutes ago, John Dyson said:

There is a distribution architecture problem with older recordings (before the early 1990s') where apparently they were digitized without DolbyA decoding.  (Take a look at the various established digitization procedures -- DA decoding is sometimes not in the process.)  There is no GOOD software DolbyA decoder until recently and even then, the amount of CPU/disk to do it super well might have been pretty expensive before yr2000 or so.  A lot of material DOES languish in digital form without being DA decoded.  (It was common practice to simply copy analog DA encoded to another analog deck/bypassing the encode & decode cycle because of quality loss and complexity issues.)  The process for copying from DA encoded to digital was probably similar (back in the 'day', it wouldn't have been unreasonable for the DA decoding process be deferred just like on analog-analog transfers.)  Nowadays, vinyl is still probably done correctly when DA encoded, because it was always part of the 'procedure'.  With digital (making CDs, preparing downloads), that DolbyA step has been ignored because of the complexity, trouble, avoiding the use of real recording engineers (or using newer, less experienced recording engineers), or WORSE YET -- the choice of Beancounters avoiding the cost.

 

Playing out old digital tapes is MUCH faster than pushing the recording through an old DolbyA unit that works only in realtime.  Additionally, there is the matter of the HW setup, etc.  'Beancounters' control the world, and any dime saved in the process is worthwhile, because the public will purchase the defective recordings anyway.

 

So, a lot (I mean, really a lot) of the older recordings are leaked DolbyA with some EQ (-3dB@3k/Q=0.707 or -6dB@3k/Q=0.707), and call it 'listenable.'   Well, it is listenable if you are used to the heavy compression typically used in today's recordings.  However, the result of such 'decoding' is NOT the original album sound.  Maybe a few earlier CDs were made correctly -- back when there were engineers regularly using DolbyA.  Not any more.

 

THAT is where some of the 'harsh' digital sound has come from.

 

John

 

Very interesting, but for what the sake reason we need DolbyA in the chain of the recording?

[John Dyson]

4 hours ago, mkrzych said:

 

Very interesting, but for what the sake reason we need DolbyA in the chain of the recording?

Back between the 1960's through the early 1990's, DolbyA was very often used for noise reduction.  Encoded recordings made back then (even archived recently) are often not DolbyA decoded.  So, those old recordings are left with the DolbyA imprint.   Decoding is tedious (due to the decoding process/ancient equipment/nothing-good-in-DAW), so it isn't done.  I have numerous examples (really numerous), and sometimes I even find material from the timeframe that isn't encoded.  MFSL stuff is often done correctly -- and some other cases of correct decoding.  However, I have Brasil'66, Simon&Garfunkel, Herb Alpert, Carly Simon, Carpenters, ABBA, Petula Clark, Queen -- all DolbyA encoded shall I go on?  Even Carpenters from HDtracks was left DolbyA encoded.  There is even the possibility that I have found 'Shake it off' used the DolbyA for a sound effect (not really sure, but sounds like it.)  I can make an example or two (short) that show the before/after results.  IT IS A REAL ISSUE.

If you look at the archiving procedures for old tape material, it does NOT include DolbyA decoding when appropriate!!! (You or I can find the references for the standardized process for transferring to digital for places like LOC if  look around enough.)

DolbyA isn't used on newer recordings except maybe for effect.  I do know that there are some situations where DolbyA has still be used, but NOT VERY OFTEN anymore on new material.

 

This is probably 50% (at least) the reason why some people might judge vinyl sounding better than digital (because, in the sense of undecoded material, they are 100% correct.)

 

John

[John Dyson]

PS:  regarding my previous post.  Our team just completed a DolbyA decoding project (now starting Telcom C4), and are getting interest from major label/recording studios and archives of old-time recordings for decoding DolbyA.  Currently, the market is focused solely on the professional frankly for one reason only -- support matters.  Dealing with consumers not having access to DolbyA tones can make the calibration pretty frustrating for support people.  The decoder currently needs computer and pro audio savvy, and is NOT GUI based.  (For example, there is a project rumor that pop group reissues are looking at the DHNRDS for decoding some of the old material.)  The SW is nice for the computer savvy pro, but would suck badly if command lines are painful for the user :-).

 

When carefully reviewing the results from the very latest (still experimental -- not the released version yet)  DA compatible decoder, it doesn't appear FOR LISTENING that there is ANY case where a true DolbyA performs as well as the software decoder.  The DolbyA Hw does have a flatter audio region frequency response (the DHNRDS has approx +-0.35dB freq response), but the listenability/distortion/clarity is much better with the DHNRDS.

I only speak of the technical behaviors or aspects, but mostly mentioning that the 'leaked' DolbyA irritated me enough that I started the project 4yrs ago.  I was grossed out once I realized why some of the old material sounded overly 'enhanced.'

 

John

[semente]

On 3/22/2019 at 4:03 PM, John Dyson said:

Back between the 1960's through the early 1990's, DolbyA was very often used for noise reduction.  Encoded recordings made back then (even archived recently) are often not DolbyA decoded.  So, those old recordings are left with the DolbyA imprint

 

Haven't they all been digitised a few times already?

[John Dyson]

On 4/1/2019 at 11:29 AM, semente said:

 

Haven't they all been digitised a few times already?

I am sure that they have been digitized, but if you digitize a tape without decoding it (usually wait until the very last step before decoding), and then just play it out (with EQ, because it sounds 'tinny'), then you have something that is left encoded.  I have a formula for recovering DolbyA encoded material (it has changed over time, but I have settled on, before decoding):

shelf +3dB@3kHz/Q=0.707 or +5dB@3kHz/Q=0.707 or +6dB@3kHz/Q=0.707.  I have found no material that needs +4dB EQ, but do have one that likes +4.5dB.

After doing the EQ, then you have the 'used to be' well known, DolbyA sound.  Then, decode.

Sometimes a DolbyA decode is NOT an improvement because of all of the IMD created into the recording.  My decoder scrubs the IMD through a very proprietary mechanism -- it doesn't just hide it like the Orban patent.  So, you can play recordings with a characteristic DolbyA distorted sound result in fairly clear sounding results. (ABBA is such a group -- sadly.  Multiple female vocals, with wall of sound and moderately heavy compresson on some material is a real problem for DolbyA.)

However, by far most of the time, if something is DolbyA encoded -- it sounds better decoded instead of being EQed.

 

John

[mkrzych]

6 hours ago, John Dyson said:

I am sure that they have been digitized, but if you digitize a tape without decoding it (usually wait until the very last step before decoding), and then just play it out (with EQ, because it sounds 'tinny'), then you have something that is left encoded.  I have a formula for recovering DolbyA encoded material (it has changed over time, but I have settled on, before decoding):

shelf +3dB@3kHz/Q=0.707 or +5dB@3kHz/Q=0.707 or +6dB@3kHz/Q=0.707.  I have found no material that needs +4dB EQ, but do have one that likes +4.5dB.

After doing the EQ, then you have the 'used to be' well known, DolbyA sound.  Then, decode.

Sometimes a DolbyA decode is NOT an improvement because of all of the IMD created into the recording.  My decoder scrubs the IMD through a very proprietary mechanism -- it doesn't just hide it like the Orban patent.  So, you can play recordings with a characteristic DolbyA distorted sound result in fairly clear sounding results. (ABBA is such a group -- sadly.  Multiple female vocals, with wall of sound and moderately heavy compresson on some material is a real problem for DolbyA.)

However, by far most of the time, if something is DolbyA encoded -- it sounds better decoded instead of being EQed.

 

John

 

Very interesting @John Dyson, thanks for your comments on it. Another thing which is annoying me, but this is related to modern recordings are two things: boxed sound and artificial stereo placing/panning. First one is kinda like thin percussion timbre and boomy double basses - it sounds like hitting a shoes box, not the real instruments - I found it recently a bit with Joshua Redmann latest recording from Nonesuch. Second one is like spreading the sound of the percussion or piano across the whole stereo soundstage - usually on stage, instrument is placed in one place and you simple cannot hear it starting at left and ending up on right. 

[John Dyson]

4 hours ago, mkrzych said:

 

Very interesting @John Dyson, thanks for your comments on it. Another thing which is annoying me, but this is related to modern recordings are two things: boxed sound and artificial stereo placing/panning. First one is kinda like thin percussion timbre and boomy double basses - it sounds like hitting a shoes box, not the real instruments - I found it recently a bit with Joshua Redmann latest recording from Nonesuch. Second one is like spreading the sound of the percussion or piano across the whole stereo soundstage - usually on stage, instrument is placed in one place and you simple cannot hear it starting at left and ending up on right. 

yikes!!!  Why don't they realize that a bit of realism is appreciated?

The only (minor) justification might be for old material -- or mastering people who just don't know how to produce a good image.  More than likely microphones were just randomly placed and someone liked to play with panning...

I am NOT a recording engineer -- but know a few -- and even I learned about proper microphone layout so that the material sounds real (or nearly so) back in high-school (45yrs ago.)

Silly panning is inexcusible unless the artist him/herself wanted a special effect of some kind.

Geesh!!!  Thanks for that info.

 

John

[Ralf11]

Silly panning is often found in psychedelic-enhanced engineering...

 

That reminds me of some Grateful Dead tracks where they tossed the band's sound into a Mix-master or something.

 

 

[Solstice380]

2 hours ago, Ralf11 said:

Silly panning is often found in psychedelic-enhanced engineering...

 

That reminds me of some Grateful Dead tracks where they tossed the band's sound into a Mix-master or something.

 

 

 

Can you tell the difference between panning and phase shifting and QSound?