Bass and subwoofers

[Sokel]

No matter who you are, I want to know on what evidence your claims are based.

Not sure how calling yourself a "Obsessive Compulsive Audiophile" makes any difference this way or that.

The effect might be real, and yes, Genelecs are nice (I enjoy the ones I have), but you still shouldn't trust subjective, uncontrolled perception.

Go to his channel,its not about calling himself,that's what I meant.

???
Subjective and uncontrolled?Did you read the paper?
It was just a note,use it or leave it,it's up to you.

 

[OCA]

Thanks, but as I designed my own system, I might actually know better than you how they were designed to be experienced. And as an engineer I know how fallible our ears (or rather the stuff between the ears) are, so I know not to use them for determining objective reality. Other people prefer to live in a subjective perception of what they believe reality is. I think my approach is referred to by the term "science" (that also appears in the name of this forum), whereas your approach is usually referred to as "belief".

I respect your expertise in designing your own system. However, I think you're conflating two different aspects here: the measurable and the experiential. Objective data is vital, but the ultimate goal of any system—no matter how well-engineered—is to provide an enjoyable listening experience. Dismissing the experiential side entirely seems counterproductive. The improvements I’ve referenced were relevant to the topic being discussed and also have been tested by a broad base of listeners in diverse environments. Statistically, that makes the results meaningful, especially when the improvements they report align with the intended design goals.

 

[Willem]

I don't think there is such a discrepancy in this area. The boomy sound of room modes is easily recognizable without any measurement. Human hearing resolution is low, but not that low. Similarly, the pleasure from listening to clean bass is real, and ultimately the point of the whole exercise. At the same time, measurement is important to confirm what we think we are hearing, and to help addressing problems.

 

[Julf]

Go to his channel,its not about calling himself,that's what I meant.

You were referring to his signature, not his channel, in your comment. No, his signature doesn't make me reconsider my question.

Subjective and uncontrolled?Did you read the paper?

I was not commenting on his paper.

 

[Julf]

I respect your expertise in designing your own system. However, I think you're conflating two different aspects here: the measurable and the experiential. Objective data is vital, but the ultimate goal of any system—no matter how well-engineered—is to provide an enjoyable listening experience. Dismissing the experiential side entirely seems counterproductive. The improvements I’ve referenced were relevant to the topic being discussed and also have been tested by a broad base of listeners in diverse environments. Statistically, that makes the results meaningful, especially when the improvements they report align with the intended design goals.

Not disagreeing with your basic premise, and not dismissing the experiential side entirely, just that I have learned to verify observations based purely on subjective perception. Sure, there are lots of situations where that verification is not needed - I don't measure the wavelength of the light coming from a traffic light before concluding it is green or red, but I have to say that I want verification *especially* in cases where the subjective perceptions align with the intended design goals. Somehow pretty much any tweak sounds subjectively better.

 

[NTK]

If Dr. Griesinger is correct, one is unlikely to be able to experience low frequency envelopment in the room haphazardly.
Source: http://www.davidgriesinger.com/vancouver_asa.ppt

From items 3&4: The recordings must contain the information. Mono bass (which basically means all vinyl) needs not apply.
From item 8: Both the room dimensions and speaker placements have to be right. Which means it is unlikely to happen by accident.

Using his room as example: Red shows the front/back room mode peak (high pressure). Green shows the left/right room mode null (high velocity/high pressure gradient). The overlap zones are where envelopment is audible.

Griesinger also came up with a method to measure low frequency envelopment which he called DFT (diffused field transfer function).
http://www.davidgriesinger.com/objmeas.pdf
His MATLAB code.
http://www.davidgriesinger.com/envelopment.zip

 

[bmc0]

Here are a couple more low-passed tracks on an XY display, in case anyone is interested. First, an older Florence + the Machine song, "All This and Heaven Too" from Ceremonials (2011):

Lots of stereo bass content below 80Hz throughout, which is somewhat unusual for a pop/rock/etc. song. Stereo bass is much more common in the classical genre. Here's an old recording (1960, Mercury Living Presence) of Mussorgsky's "A Night on Bald Mountain":

 

[Sokel]

Here's a nice sum of the (myth) of mono bass at ALL LP's and probably an agreement about the AE through it:

50 Years Of Record Cutting And Stereo Bass For Vinyl | RPM's Vinyl Blog

RPM Records article about record cutting and stereo bass for vinyl featuring Richard Simpson. The cutting engineer for RCA, Contemporary and Erika Records.

www.rpmrecords.dk

Coming from someone cutting vinyl for 50 years has some value.
There's more sources like that if one search around for pros that cut vinyl.
And that explains vinyl transfers from 60's (I know of classical,folks here may know others) who absolutely had stereo bass.

 

[levimax]

After reading through this I guess I better keep my collection of "original CD's and add to them as they contain both dynamic information which gets erased with most re-masters and they also may contain AE information which is now getting erased by lossy processing to Dolby Atmos.

 

[goat76]

After reading through this I guess I better keep my collection of "original CD's and add to them as they contain both dynamic information which gets erased with most re-masters and they also may contain AE information which is now getting erased by lossy processing to Dolby Atmos.

The actual Atmos productions are not lossy, it's just the current format used for streaming it that is lossy, but that will likely change if Atmos succeeds as a multichannel music format in the future.

 

[Andysu]

i played with lissajous with the laser and sub bass

 

[bmc0]

And that explains vinyl transfers from 60's (I know of classical,folks here may know others) who absolutely had stereo bass.

I've not personally looked at any classical releases on vinyl, but I don't doubt that at least some of them had stereo bass. The Mercury recording above is from the series of tape-to-digital transfers/remasters done in the '90s.

 

[Sokel]

I've not personally looked at any classical releases on vinyl, but I don't doubt that at least some of them had stereo bass. The Mercury recording above is from the series of tape-to-digital transfers/remasters done in the '90s.

I have posted examples at older threads but it's scattered,most from analog tapes but some from vinyl transfers too.
The point is that the mono bass at vinyl is not something that is absolutely necessary and that is coming from various sources from pro world.

Of course thats not the point of the thread as stereo bass is pretty common now and the ones who prefer it can find ways to enjoy it.

 

[Keith_W]

In Toole's book Ch 8.4 "Stereo Bass: Little Ado about Even Less" he dismisses the idea on two grounds: (1) psychoacoustic studies which shows the effects are either subtle, require trained listeners, or do not meet audible thresholds, even with contrived test signals; and (2) recordings, especially recordings mastered for vinyl, do not have stereo bass anyway.

In the spirit of open-minded investigation, I am prepared to put aside point 1 and accept the findings of Griesinger and Lund. So I investigated point 2 - do recordings contain stereo bass?

I decided to do a null test in Audacity. I invite ASR members to criticise my method - if there is a flaw in my reasoning or my method, I would like to hear about it. I started with a CD rip of a mono recording - a 1929 recording of Artur Schnabel playing Beethoven piano sonatas, remastered on digital by EMI in the 1980's. The intention was to confirm that the method works. This was my process:

1. Open the recording in Audacity

2. Split the stereo track into two mono tracks by clicking the down arrow and select "Split Stereo to Mono" as shown above
3. Select the second track then invert it with Effect - Special - Invert
4. Select both tracks with Ctrl-A ("select all")
5. Combine both tracks with Tracks - Mix - Mix and Render to New Track

6. Here we see a nicely nulled mono track, confirming that the operation has worked. Now select "Spectrogram" from the down arrow. Hovering the mouse on the vertical scale (the cursor will change to a +) we select 0Hz - 120Hz to zoom in.

7. Result. Anything black = nulled out. Anything coloured = contains sound energy.

So why does a 1929 mono recording have decorrelated bass below 40Hz and down to 0Hz? This particular remaster came from 78's. If you listen to the recording, you can hear clicks and pops from vinyl playback. I am convinced that the decorrelated bass is an artefact of the remaster, perhaps from rumble from the turntable, warped disc, or something. Indeed, if you zoom in you can see a regular periodic repeating band between 30-40Hz. I am not sure what that is, 78rpm should be 1.3Hz.

Regardless, I repeated the experiment by splitting the stereo track into mono and deleting the second mono track. I duplicated the first track then inverted and nulled it. The result (as you would expect) was a perfect null - black screen.

Anyway, now that I have come up with a method to examine for stereo bass, I analysed some of my other CD rips. I found that all of them did not null below 100Hz. In other words, every recording I looked at contained stereo bass.

For example, this is Hugh Masakela's "Stimela", a well-known audiophile favourite. We can see some areas of nulls, especially below 40Hz. You can see why audiophiles like it, there is no compression so it is not a victim of the loudness wars.

And this is Michael Jackson's "Wanna Be Startin Somethin" from "Thriller". In contrast, this track was mastered much louder so there is a lot of digital clipping (exceeds 0dB). Once again, the spectrogram shows decorrelated bass between 0Hz - 100Hz.

So - is this evidence to dismiss Dr. Toole's 2nd assertion - that recordings do not contain stereo bass? Of course, I would have to look at thousands of tracks for real evidence, and so far I have only examined half a dozen. But it's a 100% hit rate so far. I am always sceptical when I get an unexpectedly good (or unexpectedly bad) result - it tells me that my method is likely to be wrong. To validate this method, I should be examining a vinyl rip, but I can't think of any recordings off the top of my head.

 

[Sokel]

In Toole's book Ch 8.4 "Stereo Bass: Little Ado about Even Less" he dismisses the idea on two grounds: (1) psychoacoustic studies which shows the effects are either subtle, require trained listeners, or do not meet audible thresholds, even with contrived test signals; and (2) recordings, especially recordings mastered for vinyl, do not have stereo bass anyway.

In the spirit of open-minded investigation, I am prepared to put aside point 1 and accept the findings of Griesinger and Lund. So I investigated point 2 - do recordings contain stereo bass?

I decided to do a null test in Audacity. I invite ASR members to criticise my method - if there is a flaw in my reasoning or my method, I would like to hear about it. I started with a CD rip of a mono recording - a 1929 recording of Artur Schnabel playing Beethoven piano sonatas, remastered on digital by EMI in the 1980's. The intention was to confirm that the method works. This was my process:

1. Open the recording in Audacity

View attachment 422496
2. Split the stereo track into two mono tracks by clicking the down arrow and select "Split Stereo to Mono" as shown above
3. Select the second track then invert it with Effect - Special - Invert
4. Select both tracks with Ctrl-A ("select all")
5. Combine both tracks with Tracks - Mix - Mix and Render to New Track

View attachment 422501

6. Here we see a nicely nulled mono track, confirming that the operation has worked. Now select "Spectrogram" from the down arrow. Hovering the mouse on the vertical scale (the cursor will change to a +) we select 0Hz - 120Hz to zoom in.

View attachment 422502

7. Result. Anything black = nulled out. Anything coloured = contains sound energy.

So why does a 1929 mono recording have decorrelated bass below 40Hz and down to 0Hz? This particular remaster came from 78's. If you listen to the recording, you can hear clicks and pops from vinyl playback. I am convinced that the decorrelated bass is an artefact of the remaster, perhaps from rumble from the turntable, warped disc, or something. Indeed, if you zoom in you can see a regular periodic repeating band between 30-40Hz. I am not sure what that is, 78rpm should be 1.3Hz.

Regardless, I repeated the experiment by splitting the stereo track into mono and deleting the second mono track. I duplicated the first track then inverted and nulled it. The result (as you would expect) was a perfect null - black screen.

Anyway, now that I have come up with a method to examine for stereo bass, I analysed some of my other CD rips. I found that all of them did not null below 100Hz. In other words, every recording I looked at contained stereo bass.

View attachment 422508

For example, this is Hugh Masakela's "Stimela", a well-known audiophile favourite. We can see some areas of nulls, especially below 40Hz. You can see why audiophiles like it, there is no compression so it is not a victim of the loudness wars.

View attachment 422505

And this is Michael Jackson's "Wanna Be Startin Somethin" from "Thriller". In contrast, this track was mastered much louder so there is a lot of digital clipping (exceeds 0dB). Once again, the spectrogram shows decorrelated bass between 0Hz - 100Hz.

So - is this evidence to dismiss Dr. Toole's 2nd assertion - that recordings do not contain stereo bass? Of course, I would have to look at thousands of tracks for real evidence, and so far I have only examined half a dozen. But it's a 100% hit rate so far. I am always sceptical when I get an unexpectedly good (or unexpectedly bad) result - it tells me that my method is likely to be wrong. To validate this method, I should be examining a vinyl rip, but I can't think of any recordings off the top of my head.

I suppose we can digitize some but of questionable quality,so...
I would instead suggest to go to the source to see what happens and deal with today's stuff.

Here's an example L an R channel) :




You can download the studio master here so to see if REW's results are legit and measure it your own way.

Linn Records

Born of love for music and superlative sound, Scottish label Linn Records boasts an impressive catalogue packed with fine musicians from the worlds of classical, jazz and folk music. Since its launch in 1982, the label has constantly been striving for high-quality audio, from vinyl to building a...

www.linnrecords.com

Edit:the above chart is electrical,not acoustical,with REW's WAV analyzer.

 

[Keith_W]

I suppose we can digitize some but of questionable quality,so...
I would instead suggest to go to the source to see what happens and deal with today's stuff.

Here's an example L an R channel) :


View attachment 422514

You can download the studio master here so to see if REW's results are legit and measure it your own way.

Linn Records

Born of love for music and superlative sound, Scottish label Linn Records boasts an impressive catalogue packed with fine musicians from the worlds of classical, jazz and folk music. Since its launch in 1982, the label has constantly been striving for high-quality audio, from vinyl to building a...

www.linnrecords.com

Edit:the above chart is electrical,not acoustical,with REW's WAV analyzer.

How do you know that the digital files which Linn have made available for download are mastered from vinyl, and not from tape or a digital master? I had a look at that link, and it does not say.

 

[Miley_B]

In Toole's book Ch 8.4 "Stereo Bass: Little Ado about Even Less" he dismisses the idea on two grounds: (1) psychoacoustic studies which shows the effects are either subtle, require trained listeners, or do not meet audible thresholds, even with contrived test signals; and (2) recordings, especially recordings mastered for vinyl, do not have stereo bass anyway.

In the spirit of open-minded investigation, I am prepared to put aside point 1 and accept the findings of Griesinger and Lund. So I investigated point 2 - do recordings contain stereo bass?

I decided to do a null test in Audacity. I invite ASR members to criticise my method - if there is a flaw in my reasoning or my method, I would like to hear about it. I started with a CD rip of a mono recording - a 1929 recording of Artur Schnabel playing Beethoven piano sonatas, remastered on digital by EMI in the 1980's. The intention was to confirm that the method works. This was my process:

I'm following this thread with great interest.

Null test are a great way to demonstrate correlation or lack of, I use them frequently.
To quote from the book "Most of the bass in common program material is highly correlated or monophonic to begin with..." He does not precisely assert that all LF information in common program material is monophonic.
Your Spectrograph view does not show an amplitude scale and even if it did I find colours are hard to accurately interpret.
How do the nulled mono sums compare to the original stereo tracks mono summed? If there is a 30dB difference below 120Hz between the nulled and the original is the uncorrelated portion significant? Upon listening to the nulled sum is the bass information still easily audible?
Looking at the spectrograph of the 1929 mono vinyl rip I think not too much should be read into the uncorrelated noise below 40Hz, it does not appear to be musical content, this is likely just demonstrating the limitations of the medium of vinyl.

My experience of the music recording industry for contemporary style music (ie. closed mic'd or electrical sources) is that rarely are engineers or producers actively making the LF content monophonic, rather that it is monophonic in the first place. A bass drum and bass guitar are the primary sources of low frequency content in many music tracks, both are usually single sources and as such will generally be placed central in the mix. Only augmentation such as effects or processing will decorrelate the LF information and that will be at a much lower level compared to the original. Of course the more acoustic or ambient based the recording the more spatial LF information will have been captured in the first place.

 

[Sokel]

How do you know that the digital files which Linn have made available for download are mastered from vinyl, and not from tape or a digital master? I had a look at that link, and it does not say.

I don't think they are mastered from vinyl,that's not the point.
Does it matter at the end of the day?Medium is irrelevant to the effect I think,or it should be .
As soon as there's music out there made this way...

Now,totally anecdotal that's a 24/192 vinyl rip from Dire Straits "Heavy Fuel" as we have the clue that they did stereo bass.
Take it with a mountain of salt until we find a legit way to see it properly other than listening to it.

REW;s full results:



*


Again,just a hint,nothing more.

*Edit:FFT size for better resolution low

 

[Keith_W]

Your Spectrograph view does not show an amplitude scale and even if it did I find colours are hard to accurately interpret.

Sorry, I should have said that in my post. The spectro's x-axis is time, and the y-axis is frequency in Hz. Amplitude is in colours - the brighter the colour, the more sound energy is present.

How do the nulled mono sums compare to the original stereo tracks mono summed? If there is a 30dB difference below 120Hz between the nulled and the original is the uncorrelated portion significant? Upon listening to the nulled sum is the bass information still easily audible?

Yes, that is a weakness in my methodology. I can not quantify the difference with the spectro. However, @Sokel suggested loading the .WAV files into REW. So this is what I did:

- In REW, File-Import. I imported a music .WAV track as audio data. REW automatically truncates the first 100s and creates two graphs representing L/R.
- I inverted the phase of the second graph
- Then Trace Arithmetic A+B

Here is the result for one track I examined with 1/48 smoothing. Red/Green = L/R. Blue = sum.

This particular piece (modern recording of a pipe organ in a church) has decorrelated bass about -20dB below the main signal.

Re: listening to the nulled track. I did not try that before you suggested it. So I just went and had a listen. I can barely hear the bass, but if I touch the subwoofer it's definitely moving.

I don't think they are mastered from vinyl,that's not the point.
Does it matter at the end of the day?Medium is irrelevant to the effect I think,or it should be .
As soon as there's music out there made this way...

The reason I want to test the method against a vinyl rip is to check that the method works for identifying decorrelated bass frequencies. What I want to see is nulled out bass <50Hz from a vinyl rip.

I think your vinyl rip of that Dire Straits track does conclusively demonstrate that YOUR method works, so I will be using it from now on.

 

[Sokel]

The reason I want to test the method against a vinyl rip is to check that the method works for identifying decorrelated bass frequencies. What I want to see is nulled out bass <50Hz from a vinyl rip.

I think your vinyl rip of that Dire Straits track does conclusively demonstrate that YOUR method works, so I will be using it from now on.

We have a saying here that age does not come alone,and forgetting the best tool on the planet for such comparison is not a good thing

So here's Deltawave in all its glory including phase info.
(tested with "bird on a wire",L to R just to see) :