Vinyl LPs - Expanding Frequency Range?

[cegreen]

Hello,

I'm new to the world of audio tools, although I'm very experienced at sophisticated image restoration, which I suspect is not terribly dissimilar in its basic concepts.

I've been working on digitizing and cleaning-up some mono classical music LPs from the early 1950's. Not only is the dynamic range compressed, but the frequency range is significantly compressed as well.

I understand about RIAA equalization, and about using expanders and/or equalizers to adjust the dynamic range and the relative loudness of various portions of the sound. I think I'm OK with this part.

Adjusting the frequency range, however, is another matter. I suspect I'm probably overlooking something simple that is staring me in the face, but I cannot seem to find a suitable AP tool or plug-in that will allow me to adjust the frequency range.

I don't want to make the low's louder--I want to make them lower--and I want to make the high's higher.

Essentially what I want to do to the sound is the equivalent of increasing the contrast in a very flat (low-contrast) image--making the darks darker and the lights lighter and adjusting the grays in-between accordingly.

Are there tools available that will allow me to do this?

Many thanks for any suggestions, and please forgive me if I'm overlooking the obvious.

-Chris

Update: After doing a little more experimenting with the RIAA and LP/78 equalizers in AP, it appears I didn't really understand how they work after all. I'll play with them some more, and see if I can do what I want.

Meanwhile, though, I cannot get the Preview function on any of the the LP/78 presets to work. I'm using AP v1.4.5 (688) on a Mac Pro running OS X 10.4.11.

Thanks.

[CDJonah_alt]

Unfortunately, I don't think the analogy follows. I really doubt that
they compressed the frequency range, they just demphasized the higher
frequencies so they are no longer there. What you need to do is amplify
the highs and lows -- but of course there isn't anything there to
amplify any more; the information has been lost. It would be as if, in
imaging terms, one compressed the color range say from 24 bit color to
256 colors. Once it is done, the data are lost and you can't go backwards.

I think if they did what you suggest, the frequencies for A1 A2 A3 and
A4 (A for different octaves) would no longer be multiples of two from
each other.

Chuck

cegreen wrote:

Hello,

I'm new to the world of audio tools, although I'm very experienced at sophisticated image restoration, which I suspect is not terribly dissimilar in its basic concepts.

I've been working on digitizing and cleaning-up some mono classical music LPs from the early 1950's. Not only is the dynamic range compressed, but the frequency range is significantly compressed as well.

I understand about RIAA equalization, and about using expanders and/or equalizers to adjust the dynamic range and the relative loudness of various portions of the sound. I think I'm OK with this part.

Adjusting the frequency range, however, is another matter. I suspect I'm probably overlooking something simple that is staring me in the face, but I cannot seem to find a suitable AP tool or plug-in that will allow me to adjust the frequency range.

I don't want to make the low's louder--I want to make them lower--and I want to make the high's higher.

Essentially what I want to do to the sound is the equivalent of increasing the contrast in a very flat (low-contrast) image--making the darks darker and the lights lighter and adjusting the grays in-between accordingly.

Are there tools available that will allow me to do this?

Many thanks for any suggestions, and please forgive me if I'm overlooking the obvious.

-Chris




_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

--
Charles D Jonah
Building 200
Chemical Sciences and Engineering Division
9700 S. Cass Avenue
Argonne, IL 60439
630-252-3471 CDJonah@anl.gov

_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

[philxm]

I have two feature additions to suggest for Martin's consideration. Although
I am no programmer, it seems to me that both would be fairly straightforward
to implement, as they both involve only simple mathematics.

1. A way to discover the amount of + or - amplification Amadeus has applied
following a Normalize move.

2. A feature similar to the Fade function in Photoshop, where, following the
application of an effect (or other modification), that effect can be "dialed
back" by a user-determined proportion.

For example, a user applies Suppress Noise to a selection. He or she then
applies this new feature (which I'll call Fade for now*). Selecting Fade
opens a small box, into which the user enters a number from 0 to 100. If,
for example, he or she selects "70", Amadeus then mixes 70% of the noise
suppression it had just performed with 30% of what the selected signal had
previously been; in other words, dialing back the suppression effect so that
it applies only 70% of its original alteration.

(*Although Martin would, of course, have to choose a different name, both to
suit his own whim and to avoid confusion with Fade-In and Fade-Out
functions.)

The beauty of this method is that it provides an easy way to alter the
degree of application of an effect, or any other sort of signal
modification. Because it can only downscale the initial modification, the
user is encouraged to initially apply a bit more than he or she is likely to
want, and then use Fade to gradually adjust it to the desired amount.

Thanks for your attention to these ideas,
--Phil M.


_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

[philxm]

Charles Jonah:

Unfortunately, I don't think the analogy follows. I really doubt that
they compressed the frequency range, they just demphasized the higher
frequencies so they are no longer there. What you need to do is amplify
the highs and lows -- but of course there isn't anything there to
amplify any more; the information has been lost. It would be as if, in
imaging terms, one compressed the color range say from 24 bit color to
256 colors. Once it is done, the data are lost and you can't go backwards.

True, but can't that simulated, via an effect such as Excitifier?

--Phil M.


_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

[cegreen]

Hi Chuck,

Many thanks for your quick response. I suspect that when the music was originally recorded, the engineers re-mapped as much as they could from the wide frequency range of the live instruments into a narrrower range that could be captured on the vinyl LP.

That is, the moderately low sounds and high sounds weren't simply truncated or lopped-off, but rather, they were squeezed into a narrower range. Now I want to unsqueeze them.

Your analogy to changing from 24-bit color to 8-bit color doesn't quite fit here. As you note, doing this changes the data from containing millions of different shades of color to 256, and, as you note, in making the change, data has been discarded.

Having said that, however, the brightness range of the image is not changed by doing this. Only the number of steps between light and dark and the subtlety of the shades of gray/color in-between are altered.

I'm currently restoring a heavily-damaged black-and-white photograph from 1890. It is very faded, and the blacks have become a medium gray and the whites a lighter gray. The brightness range is now very narrow, and the image is both faint and flat. Using Photoshop, though, it is not difficult to take that faded, low-contrast image and make the blacks black again and the whites white.

That's very much what I want to do to the sound recordings.

Best,

-Chris

[CDJonah_alt]

cegreen wrote:

Hi Chuck,

Many thanks for your quick response. I suspect that when the music was originally recorded, the engineers re-mapped as much as they could from the wide frequency range of the live instruments into a narrrower range that could be captured on the vinyl LP.

Almost certainly not -- first because it doesn't make sense musically
and second, at that time they almost certainly didn't have a good way to
do that.

That is, the moderately low sounds and high sounds weren't simply truncated or lopped-off, but rather, they were squeezed into a narrower range. Now I want to unsqueeze them.

No, they weren't truncated; they were rolled off.

Your analogy to changing from 24-bit color to 8-bit color doesn't quite fit here. As you note, doing this changes the data from containing millions of different shades of color to 256, and, as you note, in making the change, data has been discarded.

Unfortunately, it does make sense; they gave up information on higher
frequencies; data were lost.

Having said that, however, the brightness range of the image is not changed by doing this. Only the number of steps between light and dark and the subtlety of the shades of gray/color in-between are altered.

I'm currently restoring a heavily-damaged black-and-white photograph from 1890. It is very faded, and the blacks have become a medium gray and the whites a lighter gray. The brightness range is now very narrow, and the image is both faint and flat. Using Photoshop, though, it is not difficult to take that faded, low-contrast image and make the blacks black again and the whites white.

I understand but what you have is only one parameter, brightness; in
sound you are dealing with two, frequency and intensity.

That's very much what I want to do to the sound recordings.

Best,

-Chris

Phil has given a suggestion for a plug-in that might do what you want;
give it a try.

Chuck

_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

--
Charles D Jonah
Building 200
Chemical Sciences and Engineering Division
9700 S. Cass Avenue
Argonne, IL 60439
630-252-3471 CDJonah@anl.gov

_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

[cegreen]

OK, thanks. I'll give it a shot. -Chris

[Sonic Purity]

I've been working on digitizing and cleaning-up some mono classical music LPs from the early 1950's. Not only is the dynamic range compressed, but the frequency range is significantly compressed as well.

I understand about RIAA equalization, and about using expanders and/or equalizers to adjust the dynamic range and the relative loudness of various portions of the sound. I think I'm OK with this part.
…
Update: After doing a little more experimenting with the RIAA and LP/78 equalizers in AP, it appears I didn't really understand how they work after all. I'll play with them some more, and see if I can do what I want.

I’m in agreement with Chuck that data has been lost. Not being an imaging expert, i’m having trouble coming up with a better analogy. Maybe gamut? Old recordings: small gamut. One can map the data to a bigger gamut, yet original data not captured by the small gamut recording cannot magically reappear… only a guess as to what might have been there. It may sound nice, and possibly better than the original (perhaps similar to stretching limited visual data into a wider area of a larger gamut setup), yet it is a form of synthesis rather than unleashing masked data.

Anyway… moving on to something i actually know about: intrinsic equalization.

First, if you are dealing with recordings from before something like April 1954 (sometime in 1954), you are dealing with pre-RIAA EQ. You’ll want to start by finding out the actual intrinsic equalization used for the particular recording, then figure out how to best be sure that it is being reproduced with that EQ.

If your vinyl digitization setup has a standard phono preamp (separate, inside an analog amplifier or receiver, or part of a USB turntable or the like), RIAA EQ has already been applied on the signal’s way off the record and into your Mac. So, there is no reason to apply any EQ via Amadeus… at least not for actual RIAA recordings.

Now, if you know that you used no preamplifier (!) with a standard analog turntable (turntable directly connected to Mac) or you know you used one without EQ, such as a “flat” preamp that would be used with a microphone, then for an RIAA recording, one would want to apply the RIAA EQ curve in AP. If the recording is pre-RIAA, you’d need to apply a different curve, probably one not (yet) in AP.

Going back to the first and more likely case, if your recording is pre-RIAA yet had RIAA applied to it during digitization, you have two options:

• 1) Apply EQ that is the difference between the original EQ and RIAA (may not be fun to calculate).
  2) Apply the Inverse RIAA EQ in AP (that may not be its name… i did not check), then apply the original EQ.

I strongly recommend researching the EQ used for each of your recordings, and being sure that it is correctly applied. Right there, i think you’ll find an improvement in apparent frequency coverage. Once you’re sure the EQ is correct, you’re in a much better position to try expander software if you still feel it is needed.

[cegreen]

Hi Sonic,

Many thanks for your detailed reply--it is much appreciated.

Your gamut analogy works fine for me. I understand that high- and low-frequency data was lost during the recording and subsequent transfer to vinyl. I have always understood that. I'm not hoping to recreate missing data--believe me, I know that cannot be done well, if at all.

What I am hoping to do, as you describe in your analogy, is to map from the smaller gamut of the original into a larger gamut. Yes, I know this is distorting the data. Yes, I know it won't sound like the original. In this case, I don't care.

Please understand that my goal is NOT to do a true restoration where I attempt to get the recording to sound the way it did when it was released and played on suitable equipment of that period.

I'm interested in the music, not in the recording as an historical artifact or in technical perfection.

Instead, I want to try to use modern tools and technologies to make the recording sound better (to my ears).

So the question remains, regardless of whether or not I should want to do such a thing, are there tools that will let me stretch the gamut of the sound?

How can I take low bass notes that are currently at 200Hz and push them down to 60-80Hz, and make corresponding adjustments to the mid and high frequencies?

Again, many thanks to everyone for your helpful replies.

BTW, I downloaded the Excitifier plugin, and after playing with it for awhile, it DID make a significant improvement in the sound.

My adjustments were rather crude, I'm sure, compared to what's possible, and it will take me awhile to become more familiar with the tool, but it is definitely a step in the right direction, so thank you, Phil, for the suggestion.

Best,

-Chris

[CDJonah_alt]

What you _don't_ want to do is push something at 200 hz down. If there
is power at 200 hz, that is really the frequency it was at. What you
want to do is add some additional signal at lower frequencies. The
plug-in you downloaded does this for the higher frequencies basically by
adding overtones of the frequencies that are there. I am not sure there
is an obvious analog to go to lower frequencies, but one could make the
assumption that the frrequencies you hear from 100 to 200 hz are
overtones of the original lower frequencies of 50 to 100 hz (or
whatever, depending on your cutoff -- note this assumption is certainly
not true as there will be considerable material with fundamentals at 200
hz or so) You could try
1) Do a spectrum of different places in your recording to estimate
where you think the roll-off at low frequncies is.
2) Duplicate the original track to make two tracks (since we are
starting with a mono you should have only one track)
3) Use the change pitch and speed command with time at 100% and pitch
-12 (one octave factor of 2
4) Use a bandpass filter to cut all power above the cutoff you see in 1.
5) Use the volume settings for the two tracks and get a sound that you
like
6) Flatten to a new document.
You might also want to look at Tom Erbe's freeware program "Soundhack".
It does a lot of things for generating interesting sounds. I used as a
simple sound handling program before Amadeus (and before it was freeware).

Chuck

cegreen wrote:

Hi Sonic,

Many thanks for your detailed reply--it is much appreciated.

Your gamut analogy works fine for me. I understand that high- and low-frequency data was lost during the recording and subsequent transfer to vinyl. I have always understood that. I'm not hoping to recreate missing data--believe me, I know that cannot be done well, if at all.

What I am hoping to do, as you describe in your analogy, is to map from the smaller gamut of the original into a larger gamut. Yes, I know this is distorting the data. Yes, I know it won't sound like the original. In this case, I don't care.

Please understand that my goal is NOT to do a true restoration where I attempt to get the recording to sound the way it did when it was released and played on suitable equipment of that period.

I'm interested in the music, not in the recording as an historical artifact or in technical perfection.

Instead, I want to try to use modern tools and technologies to make the recording sound better (to my ears).

So the question remains, regardless of whether or not I should want to do such a thing, are there tools that will let me stretch the gamut of the sound?

How can I take low bass notes that are currently at 200Hz and push them down to 60-80Hz, and make corresponding adjustments to the mid and high frequencies?

Again, many thanks to everyone for your helpful replies.

BTW, I downloaded the Excitifier plugin, and after playing with it for awhile, it DID make a significant improvement in the sound.

My adjustments were rather crude, I'm sure, compared to what's possible, and it will take me awhile to become more familiar with the tool, but it is definitely a step in the right direction, so thank you, Phil, for the suggestion.

Best,

-Chris




_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

[Gerard Bik]

Excellent idea, Chuck. I doubt if the result is worth the trouble, so I'm definitely going to try it out sometime!

As for the analogy with Photoshop, the high and low frequencies are clipped, like when you cut off the shadows and highlights. That information is never coming back. In an image it is no problem to expand the remaining data to get a picture with true blacks and even the clipped highlights can be restored somewhat. (Love the Shadows/Highlights filter)
In sound editing this won't work because we are dealing with *MUSIC* ! It may technically be possible to remap your 200 Hz basses to 50 Hz and the 10.000 Hz to 20.000 Hz , but what will you have then? A bass guitar tuned way too low and a mickey mouse trumpet. Every note but the middle C will be out of tune! LOL

An idea worth trying out, but only for the sound experimentalist, not for music lovers.

Excitifier is your only hope as far as I know. And Chuck's bass restoration protocol, maybe.

Comparing sound editing to image editing is great fun, but you just have found the main difference: music.

Gerard (Editing more images than sounds)

What you _don't_ want to do is push something at 200 hz down. If there is power at 200 hz, that is really the frequency it was at. What you want to do is add some additional signal at lower frequencies. The plug-in you downloaded does this for the higher frequencies basically by adding overtones of the frequencies that are there. I am not sure there is an obvious analog to go to lower frequencies, but one could make the assumption that the frrequencies you hear from 100 to 200 hz are overtones of the original lower frequencies of 50 to 100 hz (or whatever, depending on your cutoff -- note this assumption is certainly not true as there will be considerable material with fundamentals at 200 hz or so) You could try
1) Do a spectrum of different places in your recording to estimate where you think the roll-off at low frequncies is.
2) Duplicate the original track to make two tracks (since we are starting with a mono you should have only one track)
3) Use the change pitch and speed command with time at 100% and pitch -12 (one octave factor of 2
4) Use a bandpass filter to cut all power above the cutoff you see in 1.
5) Use the volume settings for the two tracks and get a sound that you like
6) Flatten to a new document.
You might also want to look at Tom Erbe's freeware program "Soundhack". It does a lot of things for generating interesting sounds. I used as a simple sound handling program before Amadeus (and before it was freeware).

Chuck

cegreen wrote:

Hi Sonic,
Many thanks for your detailed reply--it is much appreciated.

Your gamut analogy works fine for me. I understand that high- and low-frequency data was lost during the recording and subsequent transfer to vinyl. I have always understood that. I'm not hoping to recreate missing data--believe me, I know that cannot be done well, if at all.

What I am hoping to do, as you describe in your analogy, is to map from the smaller gamut of the original into a larger gamut. Yes, I know this is distorting the data. Yes, I know it won't sound like the original. In this case, I don't care.

Please understand that my goal is NOT to do a true restoration where I attempt to get the recording to sound the way it did when it was released and played on suitable equipment of that period.
I'm interested in the music, not in the recording as an historical artifact or in technical perfection.

Instead, I want to try to use modern tools and technologies to make the recording sound better (to my ears).

So the question remains, regardless of whether or not I should want to do such a thing, are there tools that will let me stretch the gamut of the sound?
How can I take low bass notes that are currently at 200Hz and push them down to 60-80Hz, and make corresponding adjustments to the mid and high frequencies?

Again, many thanks to everyone for your helpful replies.
BTW, I downloaded the Excitifier plugin, and after playing with it for awhile, it DID make a significant improvement in the sound.
My adjustments were rather crude, I'm sure, compared to what's possible, and it will take me awhile to become more familiar with the tool, but it is definitely a step in the right direction, so thank you, Phil, for the suggestion.

Best,

-Chris




_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

--
______________________ _
Gerard Bik grafische vormgeving
Van Aerssenstraat 263
2582 JM Den Haag
070 3554081
_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

[rfwilmut]

The suggestions for reconstructing frequencies above the cutoff frequency simply amount to introducing substantial amounts of second harmonic distortion.

Remember that with a single note, the tone colour is carried by the relative levels of harmonics above it in progression - if the note is C, then it's C above, then going upwards, G, C, E, G, B (very approx) and then closer together. Shift them up an octave and audio chaos results.

[Gerard Bik]

Just tried the procedure Chuck described. It does work better than I
imagined, the impression of bass comes back, but it is (inevitably) a
muddy kind of bass.

Experiment:
I started with a good track (A), filtered out the low below 125 Hz
(B). Of that B track I lowered the pitch one octave and filtered out
everything above 125 Hz (C). Combined this bass replacement C with
track B and the result was better than just B but not very good
compared to A.

Conclusion: you won't get back what was lost, but the overal
impression of the sound is improved.

You can hear the experiment if you like. I put track A, B and C in a file at
http://dl.dropbox.com/u/3491123/ChuckBassTisserands.mp3

GB

ps. The quality of the pitch converter proves to be very important.

--
______________________ _
Gerard Bik grafische vormgeving
Van Aerssenstraat 263
2582 JM Den Haag
070 3554081
_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

[cegreen]

Hello Everyone (as Karl Haas used to say),

Again, many thanks to all of you for your very helpful comments, suggestions, and insights.

My take-away from this discussion is that my assumptions in likening music editing to image editing did not include one extremely important distinction: music has harmonics, something that images do not have. This makes them fundamentally different.

And because these complex, highly-interrelated harmonics permeate the music, and are crucial to our perception of it, the kinds of adjustments that are routine in image editing are essentially impossible with sound.

This now makes perfect sense to me, and I'm truly grateful to you for taking the time to provide this crucial insight, which I'm sure all of you sound editors learned on Day One.

Best,

-Chris

[Gerard Bik]

That should be track A, track B and track B+C.

Just tried the procedure Chuck described. It does work better than I imagined, the impression of bass comes back, but it is (inevitably) a muddy kind of bass.

Experiment:
I started with a good track (A), filtered out the low below 125 Hz (B). Of that B track I lowered the pitch one octave and filtered out everything above 125 Hz (C). Combined this bass replacement C with track B and the result was better than just B but not very good compared to A.

Conclusion: you won't get back what was lost, but the overal impression of the sound is improved.

You can hear the experiment if you like. I put track A, B and C in a file at
http://dl.dropbox.com/u/3491123/ChuckBassTisserands.mp3

GB

ps. The quality of the pitch converter proves to be very important.

--
______________________ _
Gerard Bik grafische vormgeving
Van Aerssenstraat 263
2582 JM Den Haag
070 3554081
_______________________________________________
Amadeus forum mailing list
Unsubscribe / change settings at http://two.pairlist.net/mailman/listinfo/forum_list

--

______________________ _
Gerard Bik grafische vormgeving
Van Aerssenstraat 263
2582 JM Den Haag
070 3554081