New Har-Bal User: Can't Get Good RMS Levels!

Having problems using the greatest Visual Mastering software of the century? Use this area of the Forum to post your technical questions to Earle and Paavo regarding Har-Bal or ask questions regarding how to work on a certain area of the software? Post away!
Post Reply
elsongs1
Posts: 4
Joined: Sun Aug 14, 2011 2:18 am

New Har-Bal User: Can't Get Good RMS Levels!

Post by elsongs1 »

Hi, I'm a REAL late bloomer to Har-Bal, just got it last week, but I think it's awesome...much more practical than any EQ I've seen.

I'm trying to master songs for my upcoming album release, and although I've gotten some good mixes with Har-Bal already, I have a couple songs that, even after Har-Bal'izing, I'm still unable to get a decent RMS level (my goal is around -11) without nasty awful squishing (Using Waves L2 here). I've re-shaped any spikes, made the peak levels more smoother (and give them more headroom), even shelved everything below 40 or 50Hz...I still can't get a good-sounding RMS reading over -15 or -14 without sounding awful (Once the drums have lost their punch, it's no good). I've even re-done my mixes, trying to cut overlapping frequencies between the kick drum and bass guitar/bass synth...doesn't seem to work. Anything I should be doing differently?
har-bal
Site Admin
Posts: 647
Joined: Wed Dec 31, 1969 8:00 pm
Location: Atlanta/Australia
Contact:

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by har-bal »

Can you do me a favor and send me one of your tracks. Go to www.hdqtrz.com and look at the bottom left and you will see a link that allows you to upload tracks to us. Can you send me two versions both as either wave or aiff files

Send one mastered with your L2 and the other without any compression on your master buss.

I will analyze the track and determine what the problem is.

Cheers

Earle Holder
www.hdqtrz.com
www.har-bal.com
Jrel
Posts: 19
Joined: Mon Jul 31, 2006 2:50 pm

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by Jrel »

I think I'm having the same issue. When using dynamic compression on my current song in Har-bal 3.0, if there are multiple kicks within a span of few seconds (doublets, triplets, etc.), their punch becomes squashed. I've been able to negate this by only using the Limiter, but I'd still like to use the dynamic compression.

What do you recommend I can do to retain the punch of those kicks, or will the dynamic compression always squish their punch? Maybe all I need is the Limiter and I don't need to use the dynamic compression?

If I just use the limiter to get a -23 db to -14 db song (for example), the rightmost yellow edge of the peak waveform on the Histogram goes farther than what is recommended in the tutorial (before the compression step). The song still sounds fine when I do that so maybe that is not an issue.
har-bal
Site Admin
Posts: 647
Joined: Wed Dec 31, 1969 8:00 pm
Location: Atlanta/Australia
Contact:

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by har-bal »

Jrel wrote:I think I'm having the same issue. When using dynamic compression on my current song in Har-bal 3.0, if there are multiple kicks within a span of few seconds (doublets, triplets, etc.), their punch becomes squashed. I've been able to negate this by only using the Limiter, but I'd still like to use the dynamic compression.

What do you recommend I can do to retain the punch of those kicks, or will the dynamic compression always squish their punch? Maybe all I need is the Limiter and I don't need to use the dynamic compression?

If I just use the limiter to get a -23 db to -14 db song (for example), the rightmost yellow edge of the peak waveform on the Histogram goes farther than what is recommended in the tutorial (before the compression step). The song still sounds fine when I do that so maybe that is not an issue.
As you probably know any type of dynamic compression will compromise the loudness of a song even if the RMS level states it is louder. You can actually have a song at -13db sound louder than a song at -10db.The more compression you apply the softer your song will sound even if you make up the gain. Our perception of loudness is better judged between the loudest and softest part of a song.

Once we can make that distinction the song will sound better as well as louder. I generally don't use anymore than 2-3db of compression because I only want to attenuate the peaks. Most folks use more compression than necessary because they believe this is how you glue a song together.

Try this......Take any song you have and EQ it and then only apply a limiter (forget the compressor for a minute). Use the limiter to barely kiss the peaks. Save the file.

Now take the same file and apply a generous amount of compression and then apply make up gain and add the limiter. Take both tracks and burn them to CD. Which one sounds the fullest and liveliest. I bet it is the first one..without the compressor.

So what is the rule?

Only use the compressor sparingly and add the limiter to increase the volume without hitting too many peaks and you will have a loud master every time.

Cheers

Earle
HarBal
Site Admin
Posts: 761
Joined: Mon Apr 19, 2004 8:18 pm
Contact:

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by HarBal »

I concur with Earle. Over maximisation of loudness figures might make for "loud figures" but the track usually doesn't sound as loud as one mastered to a more conservative level because it loses too much dynamic range.

Har-Bal was never designed with loudness maximisation in mind. It was design for maximising clarity. As such the gain envelope speed is optimised for musicality and not volume maximisation. Anything beyond -13 dB and you are really pushing it way beyond what it was designed for. A level of -11dB is too high in my view. Remember that the figures reported in HB3 are relative to full scale square (for consistency reasons with respect to the histogram view) so the figure of -11dB you're aiming for is -8dB when view in HB 2.3. That's only 2 dB shy of the worst of the worst so I'd argue you should lower it at least 2 dB to -13dB re full scale square, though my personal preference would be for around -15dB.

The other issue that may well limit what you can do has to do with how the material was recorded. Percussive instruments are quite often poorly compressed in the tracking stage resulting in the introduction of strong short duration transients (when the attack time of the compressor is too long or there is not enough look ahead). If your track suffers from this problem it will limit what you can do at the mastering stage. The best course of action is to fix the transients prior to mixing and then master as per usual.

cheers,


Paavo.
Jrel
Posts: 19
Joined: Mon Jul 31, 2006 2:50 pm

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by Jrel »

Thank you very much for the helpful and informative replies. I did want you to know that I'm definitely not trying to get my track above -14 dB.

I've been trying to get my music as "loud" as possible to compare it with some favorite trance dance songs of mine (from Heatbeat and Matt Zo) that are at -9.6 to -10.6 dB RMS, but personally, I found that -14 dB was the max that I could get to without overly squashing everything. The goal of using the dynamic compression was that it made the instruments within the song "thicker" in a more pleasant way than using my old favorite Waves L2. My song is -22.36 dB when I first put it into Har-Bal 3 (-20.3 dB in Sound Forge), and by following the tutorial, I get it to -18.76 dB by using the Limiter, and then used Dynamic Compression to get it to -15.90 dB (which comes out to -14 dB RMS in Sound Forge). That was when I found the kick squashing issue. So instead, I only used Har-bal's Limiter to get it to -15.90 dB, and that seemed to work by getting the loudness I wanted. However, I'm going to try just squashing the peaks by 2-3 dB only. When I used to use Waves L2, I only limited the songs by up to 3 dB anyway.

Would I be better off using Har-Bal 2.3 for the Limiter or does Har-Bal use a better limiting algorithm?

I also noticed Har-Bal 3 doesn't show a peak range like 2.3 does; so would that mean a peak of 2 dB difference between the Average and Peak spectrum plots isn't exactly 2 dB because they are "relative to full scale square"?
HarBal
Site Admin
Posts: 761
Joined: Mon Apr 19, 2004 8:18 pm
Contact:

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by HarBal »

The limiter in HB2.3 introduces distortions not present in HB3 limiter. The HB3 limiter has an extra stage to avoid clipping and to smoothly bring down transient material. HB2.3 limiter would just clip it which gives it a harder sound. On certain material HB2.3 would introduce clicking sounds because of this whereas HB3.0 does not. Also, HB3 has a dc blocker in it to avoid gain control signal feed through shoud you source material have any dc in it. All in all I would recommend that you use the HB3 limiter over HB2.3.

The peak signal figure of merits were removed from HarBal because they were, (1) not particularly useful and (2) misleading (you can get a peak value greater than 0dB with it). The peak histogram gives you a better idea of what is going on anway. The "average" figure of merit is the same as the average one in HB2.3 apart from being relative to full scale square instead of full scale sine. The "processed" figure is the estimated loudness post processing.

cheers,


Paavo.
EugeneB
Posts: 13
Joined: Sat Apr 20, 2013 8:27 pm

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by EugeneB »

har-bal wrote:
Jrel wrote:As you probably know any type of dynamic compression will compromise the loudness of a song even if the RMS level states it is louder. You can actually have a song at -13db sound louder than a song at -10db.The more compression you apply the softer your song will sound even if you make up the gain. Our perception of loudness is better judged between the loudest and softest part of a song.

Once we can make that distinction the song will sound better as well as louder. I generally don't use anymore than 2-3db of compression because I only want to attenuate the peaks. Most folks use more compression than necessary because they believe this is how you glue a song together.

Try this......Take any song you have and EQ it and then only apply a limiter (forget the compressor for a minute). Use the limiter to barely kiss the peaks. Save the file.

Now take the same file and apply a generous amount of compression and then apply make up gain and add the limiter. Take both tracks and burn them to CD. Which one sounds the fullest and liveliest. I bet it is the first one..without the compressor.
I read this post and agree with it completely, I noticed that my tracks compressed with T-Racks have RMS of ~ -9.190 but sound soft, whereas my tracks mastered with Har-Bal's Volume Match have RMS of ~ -10.412 (supposedly softer) but in actuality sound much louder.

So is the Har-Bal Volume Match a limiter, a dynamic compressor, or some other kind of compressor? It seems to do a better job of raising volume than all the various compressors in T-Racks.
har-bal
Site Admin
Posts: 647
Joined: Wed Dec 31, 1969 8:00 pm
Location: Atlanta/Australia
Contact:

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by har-bal »

EugeneB wrote:
har-bal wrote:
Jrel wrote:As you probably know any type of dynamic compression will compromise the loudness of a song even if the RMS level states it is louder. You can actually have a song at -13db sound louder than a song at -10db.The more compression you apply the softer your song will sound even if you make up the gain. Our perception of loudness is better judged between the loudest and softest part of a song.

Once we can make that distinction the song will sound better as well as louder. I generally don't use anymore than 2-3db of compression because I only want to attenuate the peaks. Most folks use more compression than necessary because they believe this is how you glue a song together.

Try this......Take any song you have and EQ it and then only apply a limiter (forget the compressor for a minute). Use the limiter to barely kiss the peaks. Save the file.

Now take the same file and apply a generous amount of compression and then apply make up gain and add the limiter. Take both tracks and burn them to CD. Which one sounds the fullest and liveliest. I bet it is the first one..without the compressor.
I read this post and agree with it completely, I noticed that my tracks compressed with T-Racks have RMS of ~ -9.190 but sound soft, whereas my tracks mastered with Har-Bal's Volume Match have RMS of ~ -10.412 (supposedly softer) but in actuality sound much louder.

So is the Har-Bal Volume Match a limiter, a dynamic compressor, or some other kind of compressor? It seems to do a better job of raising volume than all the various compressors in T-Racks.
Please read the article on our main page http://www.har-bal.com/ titled "Dynamics Processing". Har-Bal determines what is needed to be done to the track upon initial analysis. It is pure genius from Paavo.

Cheers

Earle
HarBal
Site Admin
Posts: 761
Joined: Mon Apr 19, 2004 8:18 pm
Contact:

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by HarBal »

There's nothing particularly magical in the algorithm except that (1) it works slowly enough to preserve short term transients and (2) it works in the log RMS domain rather than the RMS domain (ie, the envelope detector works on a dB signal not an RMS level signal) which makes it better at tracking wide level changes without much sign of pumping. That said, if you overuse Har-Bal's dynamic processing to push for "contemporary" levels then it won't work well. It wasn't designed for that level of "abuse" so if that is what you're after then you'll need to go elsewhere to another product.

I think far too many people these days are obsessed with the figures and chase numbers rather than actually listening to what it sounds like. Our hearing perception isn't driven by numbers and the sorts of processing generally applied these days, whilst given better numbers just doesn't make much difference to the apparent loudness and usually comes at a high cost so you are far better off not going that far.

I find it interesting to say the least, that when I'm listening to the commercial CD's in my collection I find I have to play the ones mastered to -10dB with the same volume on my amp than the ones mastered at -13dB yet the -10dB ones sound like shit in comparison. To my ear at least there was no benefit going that loud only cost.

cheers,


Paavo.
EugeneB
Posts: 13
Joined: Sat Apr 20, 2013 8:27 pm

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by EugeneB »

So as I understand from this quote from 2012,
Percussive instruments are quite often poorly compressed in the tracking stage resulting in the introduction of strong short duration transients (when the attack time of the compressor is too long or there is not enough look ahead).
the drums/kicks need a Compressor on them in order to avoid strong peaks? Prior to processing through Har-Bal, we need to make sure that the Drum track (when recorded solo) doesn't oscillate too much and looks continuous without large peaks?
HarBal
Site Admin
Posts: 761
Joined: Mon Apr 19, 2004 8:18 pm
Contact:

Re: New Har-Bal User: Can't Get Good RMS Levels!

Post by HarBal »

Yes, that's correct.

Short duration transients (a few milliseconds or less) should not pose much of a problem for the Har-Bal limiter,but yes, to get a high RMS level you'll find it much easier if you compress the drum tracks before mixing.
Post Reply