Strategies for reducing treble harshness on 1930s tracks

It's all about the equipment

Moderators: Mr Awesomer, JesseMiner, CafeSavoy

Message
Author
SoundInMotionDJ
Posts: 154
Joined: Wed May 30, 2007 8:27 pm

Re: Strategies for reducing treble harshness on 1930s tracks

#16 Post by SoundInMotionDJ » Wed Apr 08, 2009 6:31 am

Cyrano de Maniac wrote:I take it this spectrogram was run against the entire song?
No. As pointed out following your post, Audacity will only profile about 20-ish seconds of a song. In the case of that graph, I selected about 16 beats of bass heavy material from a modern bass heavy pop song (the graph is labeled as such in my original post). The graph was posted as an example of what RTA output looks like.

This is NOT a graph of a 1930's song.
Cyrano de Maniac wrote:Unfortunately that won't likely show you what you're really interested in, as it's sort of the average across every moment of the song. The 48Hz peak might be from a kick drum or bass or just a recording technology artifact.
Actually the 48Hz peak is a bass guitar.
Cyrano de Maniac wrote:Even then, you run into the problem that loudness as perceived by human ears is not constant across the audio spectrum.
True. See the Equal Loudness Contour for more on this.

Cyrano de Maniac wrote:The higher frequencies require a lot less energy to be perceived as loud.
Semi-True. Human hearing is much more sensitive in the spoken voice frequency range than either low or high frequencies. And, in general, the power density of music has a -3dB/oct slope as frequency increases. So, human hearing is not as sensitive to high frequencies, but it does takes a lot less energy to produce sounds in that range. This means that high frequencies need to be louder than mid frequencies to be perceived to be "the same" volume level.

As an example, at 40phon, 10kHz needs to be 15dB louder than 1kHz for both to be perceived to be "equally" loud. But, each tone could be produced with "the same" amount of energy.

Psychoacoustics can be fun.

Haydn
Posts: 1277
Joined: Tue Nov 09, 2004 5:36 am
Location: London

#17 Post by Haydn » Thu Apr 23, 2009 12:08 pm

I've been playing around with ways of reducing treble harshness on late 30s/early 40s tracks recently. I think the problem is usually with high-pitched horns on tracks recorded between 1936 and 1941. It's also only a problem with SOME tracks - some sound just fine, others sound really bad in some places.

Just going through my songs, here are some examples of songs which suffer from harsh trebles, and some which don't.


Suffer from harsh trebles in some places

A-Tisket, A-Tasket, Chick Webb and Ella Fitzgerald 1937
Doin' The Jive, Glenn Miller 1937
Gotta Pebble In My Shoe, Chick Webb and Ella Fitzgerald 1938
Free Wheeling, Artie Shaw 1937
Ya Got Me, Tommy Dorsey 1938


Don't have a problem with harsh trebles

Bei Mir Bist Du Schoen, Chick Webb and Ella Fitzgerald 1937
Betcha Nickel, Chick Webb and Ella Fitzgerald 1939
Jeepers Creepers, Gene Krupa 1938
Undecided, Chick Webb and Ella Fitzgerald 1939
You've Got Me Voodoo'd, Charlie Barnet 1940


Solutions

I know that some of the posters on this thread have offered technical insight into the problem, but I just don't understand enough of this yet (I have found that the treble harshness is usually around 2.5Khz, 5Khz and 10Khz frequencies). I am very interested in solutions though ...

I've found one simple way to make things sound better is to turn up the bass and turn down the treble. But it's best not to turn the treble down too much, or the songs lose some of their treble 'sparkle'. I sometimes have to have my hand constantly on the treble knob when DJing.

Another method I've tried with some success is to use an equaliser to fine-tune the sound for particular tracks. You can do with programs like Audacity to permanently save an altered version of a song. I use a program called Amadeus Pro on my Mac.

But I think the best solution I've found so far is to use specialist sound enhancement software for playback. I've never used these for DJing, but played around with them on my Mac at home and PCs in internet cafes, and they certainly make late 30s tracks sound better and reduce a lot of the treble harshness.


Breakaway for Windows

http://www.claessonedwards.com/winmore.htm

This one sounded great when I tried it. It's by one of the people who produced Volume Logic. If they did 'Breakaway for Mac' I would buy it and try DJing with it.


Volume Logic for iTunes on Mac

http://www.swingdjs.com/phpbb2/viewtopic.php?t=2194

This is great, but unfortunately now discontinued, and I can't get it to work consistently with my Mac.


I've also played around with 'Hear' and 'iWow' on the Mac, but so far they don't sound as good as Breakaway or Volume Logic.

I think these programs somehow 're-process' the sound to make the song sound better. I think if I could find the right sound enhancement program, it might well be the best (and simplest) solution to this problem.

mellifluent
Posts: 4
Joined: Wed Apr 15, 2009 7:39 pm
Location: Los Angeles, CA

Using an RTA and Possible Tools

#18 Post by mellifluent » Sun Apr 26, 2009 1:05 am

You can use an RTA to identify what frequencies might be creating the shrill sounds you're trying to eliminate by looking for relative peaks in the upper mid to high frequency range. They way you would be using an RTA here is similar to how Cyrano de Maniac was describing using a parametric EQ and your ears rather than a piece of software.

Once you have a possible problem frequency from the RTA, You can use a number of tools to try to address the frequency ranges the RTA suggested might be problem areas.

Tools that could be useful to address these problems are:
-Parametric EQ (either in a digital audio workstation (DAW) like Audacity or a hardware EQ in line after your sound card's output)
-Compression
-Multiband compression (more likely to be a DAW software plugin, hardware versions of these tend to be quite pricey)

Parametric EQ can be useful if the problem is, in fact, a single problem frequency range. You want to be careful using this because you can really kill the sound quality if you start taking sweeping hacks at your song's frequency content. Also, if you want to get picky, there are issues with how an EQ affects the phase cohesion of your song. If you start taking deep stabs (I would consider more than 3dB a deep stab) with an EQ, you will begin to adversely affect parts of your song that you don't want to.

Compressors could be quite useful to help fix this problem, but there is more of an art to the use of compressors than EQ. A standard compressor will actively turn down the volume of a song (adjusting at the rate of fractions of a second) based on a single threshold that is set for the entire frequency range. However, you if this harshness that you're hearing isnt creating the loudest peak in the RTA readout, your compressor will start to trigger before the source of the harshness can trigger the compressor (so you would be compressing the part of the song you want to preserve). To avoid this problem you can use a multiband compressor, which has a threshold that can be set differently across selected frequency bands.

I am not an expert on restoration mastering, but I wouldn't be surprised if someone at "studioforums" is. Throw a post up on the mastering forum there and maybe someone would know. That place tends to get frequented by a few people who really know what they're talking about.
http://studioforums.com/eve/forums

-Nathan

Haydn
Posts: 1277
Joined: Tue Nov 09, 2004 5:36 am
Location: London

#19 Post by Haydn » Fri May 08, 2009 8:21 pm

In London, part of the vintage dance scene specialises in playing from 78s (and occasionally vinyl). For example, at the monthly Black Cotton Club http://www.myspace.com/blackcottonclub as far as I know, they play mainly from 78s. A lot of what they play is rare early 30s stuff, but some is more mainstream. But it's almost all played from 78s. I was there tonight, and I was listening to the sound quality of the music and comparing it to CD. They played quite a few tracks that I have on CD, so I could compare the two. I noticed that on the 78s the treble was a lot richer, and there was accompanying old record treble noise.

Also, I heard no treble harshness from the 78s at all during the whole night. This makes me think that the treble harshness I've been talking about is something that's created when the music is transferred to modern formats like CD.

Haydn
Posts: 1277
Joined: Tue Nov 09, 2004 5:36 am
Location: London

#20 Post by Haydn » Thu Jun 11, 2009 12:41 pm

mellifluent wrote:I am not an expert on restoration mastering, but I wouldn't be surprised if someone at "studioforums" is. Throw a post up on the mastering forum there and maybe someone would know. That place tends to get frequented by a few people who really know what they're talking about.
http://studioforums.com/eve/forums

-Nathan
I was just going to post a question on 'Studio Forums' and I got this message:

"This page is not available. The service for this site has been suspended."

Does anyone know of another similar forum?

User avatar
trev
Posts: 736
Joined: Tue Sep 14, 2004 8:20 pm
Location: Perth, Australia

#21 Post by trev » Thu Jun 11, 2009 11:58 pm

Haydn wrote:Also, I heard no treble harshness from the 78s at all during the whole night. This makes me think that the treble harshness I've been talking about is something that's created when the music is transferred to modern formats like CD.
Yes, more often than not, I've found that these sorts of problems are more to do with the transfer, rather than the actual recordings. I'm sure most of us now seek out particular labels for their reliable transfers.

Obviously the sound system/venue makes a big difference too.

Haydn
Posts: 1277
Joined: Tue Nov 09, 2004 5:36 am
Location: London

#22 Post by Haydn » Fri Jun 12, 2009 12:20 pm

trev wrote:Obviously the sound system/venue makes a big difference too.
I haven't found the sound system or venue really affect the harsh treble sounds. If the trumpets sound horrible, they just sound horrible no matter what. Obviously the louder the treble, the worse it is, and if you turn the treble down, you lessen the impact on your ears, but if the treble sounds bad, it will still sound bad irrespective of the sound system or room - that's my experience anyway.

User avatar
trev
Posts: 736
Joined: Tue Sep 14, 2004 8:20 pm
Location: Perth, Australia

#23 Post by trev » Tue Jun 16, 2009 1:36 am

Haydn wrote:...if the treble sounds bad, it will still sound bad irrespective of the sound system or room - that's my experience anyway.
Yes, if the trebles have been mucked up somewhere along the line (excessive noise reduction for example - which is really common on budget priced CDs), there's not a lot you can do other than turning the volume of the annoying frequencies down. But the problem, in theory, can be at any point.

I think of things like this:

Recording session -> masters -> 78 record -> CD -> MP3 -> computer -> mixer -> sound system -> room

If someone has botched things at any one of these stages, the sound passed along can be crap and impossible to recover: A recording session may have used an inferior setup or technology; Good reissues skip over the 78 stage, so that is one less thing to get wrong; Poor CD remastering (noise compression etc) messes things up at the CD stage; A cheap soundcard can add noise thru the computer; A boxy room with lots of hard surfaces will lose the clarity and warmth of the sound.

I spent time with a sound engineer at a live gig recently it was really interesting to see him EQ the band for the venue: he'd listen for frequencies that clashed and adjust things to eliminate the problems.

User avatar
J-h:n
Posts: 193
Joined: Wed Mar 14, 2007 6:09 am
Location: Gothenburg, Sweden

#24 Post by J-h:n » Sun Jun 21, 2009 3:12 pm

Good reissues skip over the 78 stage
It ain't necessarily so. John RT Davies generally used 78s, and his remasterings often sound better to me than remasterings done from original metal parts - especially when it comes to harsh treble sounds. They may not be quite as clear, but they're much more pleasant to the ear.

If you can, compare the remasterings on Strictly Jive to those on the Ella/Chick Swingsation CD. Or try any JRTD remastering compared to any official Decca version of the same song.

User avatar
trev
Posts: 736
Joined: Tue Sep 14, 2004 8:20 pm
Location: Perth, Australia

#25 Post by trev » Sun Jun 21, 2009 7:13 pm

J-h:n wrote:
It ain't necessarily so. John RT Davies generally used 78s, and his remasterings often sound better to me than remasterings done from original metal parts - especially when it comes to harsh treble sounds. They may not be quite as clear, but they're much more pleasant to the ear.
Yes, I completely agree - perhaps what I should have written was that generally a better result can be obtained by directly sourcing the master recordings.

Locked