• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Those of you who believe measurements aren't the whole story, do you have a hypothesis why that is?

kemmler3D

Major Contributor
Forum Donor
Joined
Aug 25, 2022
Messages
4,001
Likes
8,090
Location
San Francisco
The recordings I now make are amateur recordings of an amateur group; various adults playing for elementary school assemblies. I plug a handheld digital recorder into the tape output of a Mackie board used for the PA at the school, set the recorder for 24 bits and 44.1 sampling and leave plenty of headroom. When I later edit the songs from the assembly, I always apply compression. This allows the vocals to stand out as they should. It's not a lot of compression and it doesn't sound unnatural.

I would guess that the end product that a mastering engineer receives would have the dynamics manipulated in a way that is most musically useful. Most likely, if the engineer of the original project has done their job properly, any additional manipulation of dynamics would not be necessary. Of course, with so much modern music, the folks in the front office would demand additional compression because "the market demands it".
Sounds about right to me. I would agree that at the mastering stage, if the mixing engineer has done their job well, there shouldn't be much need for more compression. But I think it's quite rare to have a multi-mic recording go straight from the mic to the record without at least some of the tracks passing through a compressor or two...

To be honest I think this norm benefits mediocre speakers more than good ones. I'd love to see more recordings with crazy dynamic range that you just need to turn up a bit to get the average levels listenable.
 

Robin L

Master Contributor
Joined
Sep 2, 2019
Messages
5,492
Likes
7,968
Location
1 mile east of Sleater Kinney Rd
Sounds about right to me. I would agree that at the mastering stage, if the mixing engineer has done their job well, there shouldn't be much need for more compression. But I think it's quite rare to have a multi-mic recording go straight from the mic to the record without at least some of the tracks passing through a compressor or two...

To be honest I think this norm benefits mediocre speakers more than good ones. I'd love to see more recordings with crazy dynamic range that you just need to turn up a bit to get the average levels listenable.
My experience is that the sort of extreme dynamics that one finds in music generally do not work in most domestic environments. I'm thinking of percussion specifically. I've got a recording of Mahler's third symphony (Chailly, Concertgebouw, Decca) that simply has too wide of a dynamic range. Bernstein's first recording, with the New York Philharmonic Orchestra, from the early 1960s, has marvelous sound and does not have the same issue with dynamics. Recording is an art, after all.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
21,281
Likes
38,715
Sounds about right to me. I would agree that at the mastering stage, if the mixing engineer has done their job well, there shouldn't be much need for more compression. But I think it's quite rare to have a multi-mic recording go straight from the mic to the record without at least some of the tracks passing through a compressor or two...

To be honest I think this norm benefits mediocre speakers more than good ones. I'd love to see more recordings with crazy dynamic range that you just need to turn up a bit to get the average levels listenable.
Robin L has covered it. With most music with percussion which is not just drums you'll need compression. Otherwise your speakers might verge on bursting from the percussion peaks while everything else is not loud enough.
 

Anton D

Major Contributor
Joined
Mar 17, 2021
Messages
1,052
Likes
1,267
Those of you who believe measurements aren't the whole story, do you have an hypothesis why that is?

I must admit, most Hi Fi hobbyists I know have times where they are busy being happy and don't always have time to bother with wondering how an experience in measuring that day.

I would imagine most/some/a few of us roll that way.
 
Last edited:

Robin L

Master Contributor
Joined
Sep 2, 2019
Messages
5,492
Likes
7,968
Location
1 mile east of Sleater Kinney Rd
Robin L has covered it. With most music with percussion which is not just drums you'll need compression. Otherwise your speakers might verge on bursting from the percussion peaks while everything else is not loud enough.
Also, with "just drums" at times. There was a rock band (the Trike Shop, in Fresno California) where I would record their benefit concerts. The drums were unamplified but were still overbalancing the band, at least on the recordings I was making. As a live performing unit, they were fine, but I always had to tuck the percussion peaks down. There were performances of a medieval band I recorded that used only hand percussion, they always were perfectly balanced and didn't require any dynamic manipulation. My experience with chamber ensembles was that they didn't require any manipulation of dynamics, always managing dynamics by simply listening to each other.
 

sq225917

Major Contributor
Joined
Jun 23, 2019
Messages
1,397
Likes
1,671
It's obvious, measurements tell you how a device measures, it doesn't tell you how it sounds, because everyone's ears are slightly different and hearing has a psychological/emotional component to it, that we simply cannot eliminate.

There's also a language component, the terms we each favour to describe specific aspects of what we experienced and how the same terms might mean different things to different people.

That's it, in a nutshell.

Trying to get further into the weeds than that is pointless.

There's no reason why one person's preference in sound should correlate in any way to improved measured fidelity to the recorded signal, none at all. For example, there's no reason why I should like a harman curve iem just because it better matches an average preference of a test group than another iem that's a worse match.

It's just a pointless square peg and round hole exercise, and we're not going to change anyone's opinion, they can only do that themselves.
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,555
Likes
4,227
Location
Pacific Northwest
Those of you who believe measurements aren't the whole story, do you have an hypothesis why that is?
...
Wes covered this a few years ago: https://www.audiosciencereview.com/...ve-a-hypothesis-why-that-is.25416/post-865388

I'll add an example: consider SNR. It's usually measured at max volume. But nobody actually listens at max volume, and when you turn the volume down to a realistic level, the SNR always drops. But it doesn't always drop by the same amount, because different devices control volume / gain in different ways. So SNR as it is commonly measured is basically meaningless. The solution is to make a different measurement, SNR at a specific output level. This is why Amir measures SNR at 50 mV.

Here's a much better example: https://www.audiosciencereview.com/...tube-amplifier-review-and-measurements.44020/

So while everything we hear can in principle be measured, doing this is more complex than the measurements commonly made and published, which are not even close to the whole story.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
21,281
Likes
38,715
Also, with "just drums" at times. There was a rock band (the Trike Shop, in Fresno California) where I would record their benefit concerts. The drums were unamplified but were still overbalancing the band, at least on the recordings I was making. As a live performing unit, they were fine, but I always had to tuck the percussion peaks down. There were performances of a medieval band I recorded that used only hand percussion, they always were perfectly balanced and didn't require any dynamic manipulation. My experience with chamber ensembles was that they didn't require any manipulation of dynamics, always managing dynamics by simply listening to each other.
Oh, I was not excluding drums. I've recorded a group that used a large Bhodhran. It didn't sound that loud, but even with a high pass filter you needed some compression. Was recording a singer once with an Earthworks mic. They also used musical shakers (a lacquered gourd with sand in it). Didn't sound very loud, but the peak energy was around 14 khz and it was pegging out the meters. So needed a less vigorous shake.
 

Verig

Active Member
Joined
May 4, 2023
Messages
257
Likes
168
@MRC01 Ah, 300B. That living treble really is pretty magical. Still, a bit poor reference when looking at amps not-from-100-years-ago. Modern amps are very clean so what's missing anymore.
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,555
Likes
4,227
Location
Pacific Northwest
@MRC01 Ah, 300B. That living treble really is pretty magical. Still, a bit poor reference when looking at amps not-from-100-years-ago. Modern amps are very clean so what's missing anymore.
The example is not about the amp. It's about how measuring what we hear can get much more complex than what we get on spec sheets.

Put differently: given basic independent measurements like SNR (measured at max output, as typical), max power, and frequency response (20 dB below max output, as typical), the 300B phenomena is unexplainable. People who listen to the 300B say there is something special about the dynamics, but engineers who read the specs and believe "measurements are everything" would say measurements "prove" that the listeners are wrong so it must all be placebo. But more complex combined measurements reveals that the phenomena listeners describe is real - the frequency response changes with power output.

Of course, modern well engineered amps won't have this issue. The point is that measurements that truly represent "everything" would be more extensive and complex than what we read on spec sheets.
 

Verig

Active Member
Joined
May 4, 2023
Messages
257
Likes
168
I get your point but in the case of 300B you can take a look at sinad and see that it's completely horribly broken. The rest is just curiosity. Also, the high level of distortion and reactance makes it clearly audible to all. No one says that 300B is neutral or that there isn't something weird going on.
It would get a bit more interesting if people really heard difference between Hypex and Purifi. Then we needed to find something special to measure.
 

kemmler3D

Major Contributor
Forum Donor
Joined
Aug 25, 2022
Messages
4,001
Likes
8,090
Location
San Francisco
My experience is that the sort of extreme dynamics that one finds in music generally do not work in most domestic environments. I'm thinking of percussion specifically. I've got a recording of Mahler's third symphony (Chailly, Concertgebouw, Decca) that simply has too wide of a dynamic range. Bernstein's first recording, with the New York Philharmonic Orchestra, from the early 1960s, has marvelous sound and does not have the same issue with dynamics. Recording is an art, after all.
Yes, agree. Anyone that's been in the same room as a drum set knows that real-world dynamics are not necessarily desirable. But I think the best compromise (with modern equipment) would be a bit closer to live than what we actually get.
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,555
Likes
4,227
Location
Pacific Northwest
Yes, agree. Anyone that's been in the same room as a drum set knows that real-world dynamics are not necessarily desirable. But I think the best compromise (with modern equipment) would be a bit closer to live than what we actually get.
Yep. Full dynamics is not for everyone, nor for casual listening. But for serious audiophiles who have the room and equipment to handle it, it's awesome. Classical music recordings don't have the heavy-handed processing that plagues other genres, but they do occasionally use dynamic range compression and I wish they would use less of it. Or, at least, apply it to the CD but not to 24-bit high res. Then the consumer decides. Better yet, don't use any compression at all and let listeners apply it on playback. This is readily available in DSP.
 

kemmler3D

Major Contributor
Forum Donor
Joined
Aug 25, 2022
Messages
4,001
Likes
8,090
Location
San Francisco
Better yet, don't use any compression at all and let listeners apply it on playback. This is readily available in DSP.
Not that most consumers (or even professionals who aren't mixing engineers, if I'm honest) can operate a compressor successfully, but if the track had the compressor settings in the metadata and it set the compressor automatically, it could be interesting.

The difficulty with this is that compression is usually applied on a per-track / instrument basis and sometimes the master channel gets another compressor also. So unfortunately this is a bit of a pipe dream unless music is delivered as stems instead of full mixes.
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,555
Likes
4,227
Location
Pacific Northwest
Not that most consumers (or even professionals who aren't mixing engineers, if I'm honest) can operate a compressor successfully, but if the track had the compressor settings in the metadata and it set the compressor automatically, it could be interesting.
Most consumers aren't audiophiles and don't care that much about sound quality anyway. All they want is to hear the quiet parts over ambient noise, or limit the peaks so they don't blow their cheap wimpy speakers. All they need is a playback compressor with settings for "on/off", or at most, "off, low, high". They can be trusted to operate that ;) I've seen this in car stereos, DVD players, and other cheap playback devices.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
21,281
Likes
38,715
Not that most consumers (or even professionals who aren't mixing engineers, if I'm honest) can operate a compressor successfully, but if the track had the compressor settings in the metadata and it set the compressor automatically, it could be interesting.

The difficulty with this is that compression is usually applied on a per-track / instrument basis and sometimes the master channel gets another compressor also. So unfortunately this is a bit of a pipe dream unless music is delivered as stems instead of full mixes.
Oh it is more complex than that. Read about parallel compression as one example. Some of the track is compressed, some is not. So undoing that at playback end is going to be "wrong" no matter what you do.

 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,555
Likes
4,227
Location
Pacific Northwest
Oh it is more complex than that. Read about parallel compression as one example. Some of the track is compressed, some is not. So undoing that at playback end is going to be "wrong" no matter what you do.
...
Indeed. All the more reason for studio engineers to use a lighter hand with compression, and if anything more is needed, that is playback specific and should be done on playback.
 

atmasphere

Addicted to Fun and Learning
Technical Expert
Audio Company
Joined
May 26, 2021
Messages
535
Likes
836
The example is not about the amp. It's about how measuring what we hear can get much more complex than what we get on spec sheets.

Put differently: given basic independent measurements like SNR (measured at max output, as typical), max power, and frequency response (20 dB below max output, as typical), the 300B phenomena is unexplainable. People who listen to the 300B say there is something special about the dynamics, but engineers who read the specs and believe "measurements are everything" would say measurements "prove" that the listeners are wrong so it must all be placebo. But more complex combined measurements reveals that the phenomena listeners describe is real - the frequency response changes with power output.

Of course, modern well engineered amps won't have this issue. The point is that measurements that truly represent "everything" would be more extensive and complex than what we read on spec sheets.
The dynamics are caused by distortion. The ear uses higher ordered harmonics to sense how loud sounds are. Above about 20-25% of full power with any SET, the higher orders start to show up on transients where the power is. The ear responds to that and presto! 'dynamics'. But its really distortion masquerading as 'dynamics'. It is higher ordered harmonics that cause such amps to sound 'loud' which is why so many people think that 4 or 7 Watts is enough power. The harmonics are just making it sound louder than it is; a sound pressure app for your phone will sort that out easily enough.
 

dasdoing

Major Contributor
Joined
May 20, 2020
Messages
4,405
Likes
2,893
Location
Salvador-Bahia-Brasil
you guys wouldn't like music without any compression, this I can guarantee.
It doesn't sound natural at all since we don't hear music with ou ears 10cm from the instruments....and that's where the mics are
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,555
Likes
4,227
Location
Pacific Northwest
you guys wouldn't like music without any compression, this I can guarantee.
It doesn't sound natural at all since we don't hear music with ou ears 10cm from the instruments....and that's where the mics are
I disagree. I listen to music without any compression, both live and recorded, and I think it sounds fantastic.
 
Top Bottom