bass build-up when using parallel single-band compression

[citrusburst]

Hi There,

I've recently begun experimenting with parallel single-band compression, as per Bob Katz's instructions in "Mastering Audio." In general, the sonic improvement is pretty impressive. One thing I've noticed, though, is that this technique can lead to a noticeable build-up in the bass frequencies. Obviously, I can use an EQ on my compressed signal to correct this, but I'm wondering if this kind of build-up is "normal". In other words, is bass build-up an anticipated effect of parallel compression? Or, does this speak to a deficiency in my technique, or, alternately, the source material?

Best Regards,
Jed

[Ged Leitch]

I havent used it in ages but as far as I remember It's often a good idea to Hi pass the Parallel comp'd channel maybe around 60 - 70hz .

But Take the above as a general way of using it.

Hopefully Lucey can jump in and correct me - I think he uses it occasionally???

Jon Scrip too I think, these guys are the best to help you out.

[bob katz]

Originally Posted by citrusburst

Hi There,

I've recently begun experimenting with parallel single-band compression, as per Bob Katz's instructions in "Mastering Audio." In general, the sonic improvement is pretty impressive. One thing I've noticed, though, is that this technique can lead to a noticeable build-up in the bass frequencies. Obviously, I can use an EQ on my compressed signal to correct this, but I'm wondering if this kind of build-up is "normal". In other words, is bass build-up an anticipated effect of parallel compression? Or, does this speak to a deficiency in my technique, or, alternately, the source material?

Best Regards,
Jed

Dear Jed: There's always something I left out of the first edition! In the second edition I cover: Sidechain equalization, sidechain linking, parallel multiband and other topics.

But I find that the usual problem with single band compression (parallel or "normal downward") is that the bass is "over controlled" rather than "under controlled" so I'm a bit puzzled. The most common equalization in compressors is a high pass in the sidechain, to allow the bass to "speak up more" and be less compressed. Which is the opposite of your issue.

What tools are you using?

If for the particular music you are working you are getting too "fat" bass while parallel compressing with a single band comp, I'd suspect you are going to have to filter the compressor (split bands), not the sideband, and someone would have to listen to see what could be your problem.

For most of the parallel compression I do, I've been using split band techniques (multiband) rather than sidechain, and if I do use a single band compressor, a high pass in the sidechain helps give a cleaner, "louder", more transient bottom end rather than overcontrolled. But you seem to be encountering the opposite situation!

[MattGray]

I think your bass problem may be source dependent. I often use parallel compression & also prefer to use single band compression with a HPF in the sidechain to let the bottom end breath & punch more.

Obviously if you are attempting to do this in your DAW with plug-ins be very careful of latency as this can cause all sorts of phase related issues which could also be the cause of your increased bottom end.

With regards to bass problems in the original source material, it's possible that parallel compression can fatten the bass in certain ways & can cause your mix to make the bass sound bloated, in these cases I normally EQ the bottom end listening to the original source & then insert a low shelf cut before the parallel compressor but on the parallel send to help clean it up more. There is many ways you can use parallel compression to achieve the result you're after which way you choose is source dependent.

For most of the parallel compression I do, I've been using split band techniques (multiband) rather than sidechain, and if I do use a single band compressor, a high pass in the sidechain helps give a cleaner, "louder", more transient bottom end rather than overcontrolled. But you seem to be encountering the opposite situation!

Hey Bob I'm assuming you are talking about the MD4 in parallel mode? Do you still use the Weiss DS-1 in parallel from time to time? Out of curiosity why do you now prefer MB over SB with parallel? I've just got myself a DS-1 & haven't really had the chance to use it in parallel yet, but some of the brief testing I've done so far indicate that it is very suitable for this purpose. Is there a way of applying a HPF to the sidechain of the parallel compressor with the DS-1? I see there is a filter option but does that cut off the low end that is feeding the compressor or is it actually a filter to the sidechain of the compressor allowing the low end to pass through without being compressed?

Matt

[bob katz]

Originally Posted by MattGray

Hey Bob I'm assuming you are talking about the MD4 in parallel mode? Do you still use the Weiss DS-1 in parallel from time to time?

I'm using the Weiss in parallel mode less often now that I've found the "secrets" of the MD4. I go for the Weiss when I'm looking for the MOST transparent (invisible) parallel compression, and for the TC when I'm looking for some "attitude" or "tonalization" or coloration. It is possible to adjust the Weiss to do some "tonalization" but it is not as successful as the TC, and vice versa, it is possible to adjust the TC to be very invisible, so invisible that I tend to forget that the Weiss is a percent superior.

Out of curiosity why do you now prefer MB over SB with parallel? I've just got myself a DS-1 & haven't really had the chance to use it in parallel yet, but some of the brief testing I've done so far indicate that it is very suitable for this purpose. Is there a way of applying a HPF to the sidechain of the parallel compressor with the DS-1? I see there is a filter option but does that cut off the low end that is feeding the compressor or is it actually a filter to the sidechain of the compressor allowing the low end to pass through without being compressed?

Matt

You can't apply a HPF to the sidechain of the DS-1, but splitting the bands produces very nice results, less interaction, I prefer that to the traditional sidechain approach. The filter is actually a filter in the audio path, so it's a "one band" compressor.

[MASSIVE Master]

I tend to HPF (and LPF for that matter) the parallel often. Depends on the mix and what I'm trying to "bring out" vs. what I'm trying to minimize.

If I'm using parallel, I'm probably trying to add "density" to something... I probably don't want extra density in the ex-girlfriend frequencies (around 2.5k), I might not want the "blossom" in the low end or the stridency in the high end.

In any case, more times than not, there's probably going to be an EQ "working hard" on the parallel (IME, YMMV, etc., etc.).

[MattGray]

Originally Posted by bob katz

I'm using the Weiss in parallel mode less often now that I've found the "secrets" of the MD4. I go for the Weiss when I'm looking for the MOST transparent (invisible) parallel compression, and for the TC when I'm looking for some "attitude" or "tonalization" or coloration. It is possible to adjust the Weiss to do some "tonalization" but it is not as successful as the TC, and vice versa, it is possible to adjust the TC to be very invisible, so invisible that I tend to forget that the Weiss is a percent superior.

Fair enough, I can understand the tendency to lean towards Weiss products for transparency but I wouldn't have thought that the TC would be considered a 'colour' box. I would've thought that some good ol' transformers or vca style analog compression would've given you all you need in those areas? I've got a Shadow Hills mastering compressor on order for adding different flavours of colour & the Weiss stuff for transparency. Of course the HEDD's processing in moderation can also be a nice way to add some different spices of colour to the mix also.

You can't apply a HPF to the sidechain of the DS-1

I thought that was the case, seems a shame that the filter section can't be re-ordered to achieve this. Perhaps I'll mention it to Daniel for a future update.

, but splitting the bands produces very nice results, less interaction, I prefer that to the traditional sidechain approach. The filter is actually a filter in the audio path, so it's a "one band" compressor.

When you say splitting the bands I'm assuming you mean the frequency dependent compression using the filters to compress certain bandwidths? This can't be done in parallel mode though right?

[citrusburst]

Gentlemen,

Thanks very much for your thoughtful replies. I'm mastering within Pro Tools for electronic delivery (mp3, aac, etc).

The chain I was using was as follows:

Source -> Waves LinEQ (corrective equalization) -> Waves LinMB (Two busses running in parallel, corrected for latency) - > Waves SSL Comp (Two busses running in parallel, corrected for latency) -> Sonnox Inflator -> Waves L2

With the LinMB processing, I was including only the three bands between 90 Hz and 11 KHz.

I was not using sidechain equalization on the parallel Waves SSL Comp.

John's comment about using the parallel comp to increase "density" is right on. That's exactly the result I was getting. But the bass did have a noticeable build at points. Using a HPF on that signal definitely helps.

Would you gentlemen (and ladies, of course) be open to listening to an mp3 and offering commentary on the mastering?

[bob katz]

Originally Posted by MattGray

Fair enough, I can understand the tendency to lean towards Weiss products for transparency but I wouldn't have thought that the TC would be considered a 'colour' box. I would've thought that some good ol' transformers or vca style analog compression would've given you all you need in those areas? I've got a Shadow Hills mastering compressor on order for adding different flavours of colour & the Weiss stuff for transparency. Of course the HEDD's processing in moderation can also be a nice way to add some different spices of colour to the mix also.



I thought that was the case, seems a shame that the filter section can't be re-ordered to achieve this. Perhaps I'll mention it to Daniel for a future update.

I wouldn't bother. In my opinion, splitting the bands so that the high pass is the compressed band and with a flat band below it avoids the interactions that occur with a filtered sidechain approach.

When you say splitting the bands I'm assuming you mean the frequency dependent compression using the filters to compress certain bandwidths? This can't be done in parallel mode though right?

Oh yes, it can. You can run a filtered band in parallel mode.

As for the TC being a color machine, you haven't lived until you investigate using 5 bands of parallel compression. One tenth of a dB move in a band fader which is tonalizing the low- and mid- level material at low levels can selectively:

--warm up
--add presence to
--fatten
--etc.

It's a VERY powerful tool. VERY different from an equalizer.

BK