Jump to content

The Bass EQ for Movies Thread


Kvalsvoll

Recommended Posts

Tried a couple quick listening sessions after porting in a couple of these BEQ filter sets. Certainly more gravitas where it belongs.

 

Regarding this implementation in the nanoAVR: with the matrixing and bass management convention, I'm not following why LCR and surrounds are provided PEQ filters. Is this provide for other DSP solutions where the bass management is handled later?

 

Personally, I have a bit of an issue with my AVR forcing a steep (maybe 6th order) 80-120Hz LPF on ALL output to the sub due to the main channels needing to be set to "full range".

 

Just want to make sure I'm not missing any important details...

Link to comment
Share on other sites

Note on filter coefficients:

 

You should verify that the q-values entered gives the desired response.

For the MiniDSP, there seems to be differences between models, so that the q values are not consistent.

However, it seems the biquad-coefficients are consistent and can be used to verify the response.

 

You can use the spreadsheet for biquad coefficient calculation and then enter the calculated biquad instead of entering parametric values like f, q and gain.

 

More specific, the q-values for the MiniDSP 2x4 with 2-way advanced plug-in has a much higher numeric value than those in the spreadsheet. 

I do not know how other models match up.

The solution is to provide values that is correct for the spreadsheet, and then verify by calculating the biquad and enter the biquad into the dsp software.

 

 

 

The difference between old vs new is 2:

 

MiniDSP old q = MiniDSP new * 2

Link to comment
Share on other sites

Alien Anthology - The first: Alien


"Mommy always told me there are no real low frequencies in the alien films.. "

BUT THERE ARE...



The first of the Alien films has a rather simple sound design compared to many modern releases.
There is not much low bass, or bass at all, to find in this one, but overall the sound is quite good, no clipping or excessive compression.

I am looking forward to see this Director's Cut version, with Bass-EQ.
Now, you may be surprised how I can present BEQ for a film I have not seen, but the process of making the filters does not necessitate watching the complete movie.
Initial filter settings are done by measurement and signal analysis alone, then later I find some scenes I use as reference, and re-watch those as I adjust the filters.
This of course also means that there may be surprises in the resulting sound experience, even though levels have been monitored for the complete soundtrack.

Throughout most of the film there is a 13Hz rumble, seems to occur always when they are in the spacecraft.
It is easy to amplify this effect so that it really shakes everything, but that is not a good idea, as it will become tiresome and people can get sick.
It is perceived as a pressurising feeling, and then everything shakes when it gets loud enough.
I chose to rise the level to make it barely noticeable, with the subwoofers set at +3dB from flat, my usual setting.
Now, keep in mind flat does not mean it measures flat, there is a lift towards lower frequencies, so that lower frequencies have the right feel and impact when playing balanced, full frequency range content at reference level.
Played like this, at reference level, the pressurising is barely noticeable, and there is a very subtle shake.

Overall I did not lift the ulf up to very high levels, and the mid and upper bass is reduced some.
The goal is to get a balanced sound, and like this the low bass is now present making for a much better overall sound experience, without sounding too strange or bloated.


Bass-EQ filters:

NOTE: Q-values is incorrect for MiniDSP 2x4, use the biquad.


LFE:

low-shelf 30hz q=1.1 gain=10dB
low-shelf 30hz q=1.1 gain=10dB

MiniDSP biquad for ONE 30Hz Q=1.1 Gain=10dB:
a1=1.997108418699020,
a2=-0.997117078212166,
b0=1.001126540402070,
b1=-1.997099056563100,
b2=0.995999899946011

low-shelf 22hz q=0.86 gain=12dB

MiniDSP biquad for 22Hz Q=0.86 gain=12dB:
a1=1.997628030893830,
a2=-0.997632182424995,
b0=1.001181393803590,
b1=-1.997621842887770,
b2=0.996456976627460

Original SFM parameters:
low-shelf 30hz slope=1.8 20dB
low-shelf 22hz slope=1.4 12dB


L, C, R:

low-shelf 34hz q=0.86 gain=16dB
low-shelf 34hz q=0.86 gain=16dB

MiniDSP biquad for ONE 34Hz Q=0.86 gain=16dB:
a1=1.996732180595880,
a2=-0.996740053359426,
b0=1.002474784800730,
b1=-1.996711280087760,
b2=0.994286169066821

Original SFM parameters:
low-shelf 34hz slope=1.4 16dB
low-shelf 34hz slope=1.4 16dB
 

 

Spectrum graphs:

Original left channel:
post-181-0-97501800-1420425861_thumb.jpg

Left channel with Bass EQ:
post-181-0-98696200-1420425944_thumb.jpg

Original LFE:
post-181-0-21865400-1420426017_thumb.jpg

LFE with Bass EQ:
post-181-0-59674900-1420426070_thumb.jpg

Link to comment
Share on other sites

Note on filter coefficients:

 

You should verify that the q-values entered gives the desired response.

For the MiniDSP, there seems to be differences between models, so that the q values are not consistent.

However, it seems the biquad-coefficients are consistent and can be used to verify the response.

 

You can use the spreadsheet for biquad coefficient calculation and then enter the calculated biquad instead of entering parametric values like f, q and gain.

 

More specific, the q-values for the MiniDSP 2x4 with 2-way advanced plug-in has a much higher numeric value than those in the spreadsheet. 

I do not know how other models match up.

The solution is to provide values that is correct for the spreadsheet, and then verify by calculating the biquad and enter the biquad into the dsp software.

 

nanoAVR biquads will be different due to nanoAVR's 96kHz sampling rate, it may also explain the differences in Q values.

 

JSS

Link to comment
Share on other sites

nanoAVR biquads will be different due to nanoAVR's 96kHz sampling rate, it may also explain the differences in Q values.

 

JSS

 

Yes, but for same sample rate (48K) and same biquad, they give the same frequency response, but the q in filter settings is different.

 

If I use the spreadsheet you sent me, I can calculate q values which fit right into the MiniDSP-provided biquad-spreadsheet, and the response is a match.

If I use the same q in parametric filters on my MiniDSP 2x4 with advanced 2-way plug-in the response does not match at all.

If I use the biquad calculated in the spreadsheet, the response is a match.

Link to comment
Share on other sites

Yes, it means to use x amount of cascaded filters.

 

If you are using biquad filter(s), like MiniDSP, sox, Audacity, and other software, I have found the filters are smoother and more well-behaved when several are cascaded at lower boost/cut than one larger one, especially at Q's greater than 0.7.  You can try just one big filter, but I cannot guarantee the presentation will be the same.

 

What do you mean by well-behaved?  Are there oscillations or clipping or something like that with one high gain shelf filter engaged?  Have you compared the 2 in a loopthru using REW?  Or are you using the cascading filters to increase the order steepness of the shelves? 

Link to comment
Share on other sites

I mean no hysteresis/overshoot before and after the shelving points. When you increase the Q above 0.7, some hysteresis is inevitable, and it is 'corrected' by the 0.5 Q filters at the end of a filter chain.  Using less amounts of filters with larger boosts when Qs are above 0.7 makes the problem worse.  Ask me how I know....

 

I have played around with BEQ for a long time, and I will bet I have seen more individual channel and whole-mix PvAs than anyone except maybe nube or Bosso, and I know that seeing a graph, I have about a 75% idea of what the film will sound like in the LF department, so I usually have a goal PvA to BEQ to when I see a set of PvAs that look pretty obviously filtered.  While others may disagree with my methods and the amount of filters, I find they give some of the cleanest recovery of LF with the least loss of midbass in the process, given the limitations of IIR/biquad/DSP filters.

 

That being said, anyone is free to contribute their BEQ solutions to Hollywood's occasional LF-filtering predisposition that has been seen and cataloged here on this site...just note what you normally run subwoofer trim at and any built in house curve so that people who run flat and with no subwoofer boost will know what to expect...

 

JSS

Link to comment
Share on other sites

Aliens Special Edition Bass EQ

And now we enter some more serious low frequency effects, as well as better overall sound for this second of the Alien films.
It is a delight to watch at reference because the sound level is right, and still there are occasional peaks here that reaches full digital 0dB.

The Special Edition adds some scenes of value for those of us who have already seen the film originally, and thus we know how the story goes -
very, very bad..

There is this scene with the vehicle driving around, the ulf is quite well done here, you just need to dig it out.
Now you can feel the ground move as the vehicle passes by, and I chose to preserve the growling engine sound.
Come to think about it, I may be the first person to actually experience this..

This film was quite challenging, but now I feel the result is just like i want it - restored ulf, preserved engine-growl and upper bass impact, while reduced some of the worst booming around 50Hz.
As the measurements show, this one does certainly not lack in the low bass department, and the low bass is mostly quite well done and matches what is going on in the picture.

One of the problems doing eq is to ensure the individual channels sum up correctly.
The filtering shifts the phase, and then the channels will not add correctly and much of the bass can be lost.
If the filters are the same on all channels, all will be well, but here it is necessary to treat lcr different from lfe.
To verify that the end result is as expected, I checked the sum levels digitally, and also by playing scenes and recording them, comparing original to Bass-EQ.
When it comes to accuracy and objectivity, a measurement always trumps my ears - there is no "palpable bass with authority" about the graphs.

Aliens Special Edition l channel original:
post-181-0-10082200-1420685440_thumb.jpg

Aliens Special Edition l channel BEQ:
post-181-0-63493100-1420685556_thumb.jpg

Aliens Special Edition lfe channel original:
post-181-0-77438300-1420685580_thumb.jpg

Aliens Special Edition lfe channel BEQ:
post-181-0-40166800-1420685650_thumb.jpg

Aliens Special Edition vehicle scene original (Recorded from playback, full range):
post-181-0-44502700-1420685678_thumb.jpg

Aliens Special Edition vehicle scene BEQ (Recorded from playback, full range):
post-181-0-58725700-1420685686_thumb.jpg
 

SL graph of recorded scenes repeat because they are run continuously; necessary to ensure the whole time interval shows.


MinDSP new filter parameters (use the minidsp spreadsheet to find biquads):

lfe:
gain -6dB
low-shelf 12hz 1.03 -12dB
low-shelf 32hz 1.07 8dB
low-shelf 32hz 1.07 8dB
low-shelf 26hz 0.90 9dB
low-shelf 24hz 1.07 9dB
high-shelf 40hz 0.77 2dB
gain +6dB


lcr:
gain -6dB
low-shelf 12hz 1.03 -12dB
low-shelf 26hz 0.90 8dB
low-shelf 22hz 1.07 8dB
low-shelf 22hz 0.90 8dB
low-shelf 22hz 0.90 6dB
gain +6dB


Original SFM filters:

lfe:
gain -6dB
low-shelf 12hz 2.0 -12dB
low-shelf 32hz 2.2 8dB
low-shelf 32hz 2.2 8dB
low-shelf 26hz 1.6 9dB
low-shelf 24hz 2.2 9dB
high-shelf 40hz 1.2 2dB
gain +6dB


lcr:
gain -6dB
low-shelf 12hz 2.0 -12dB
low-shelf 26hz 1.6 8dB
low-shelf 22hz 2.2 8dB
low-shelf 22hz 1.6 8dB
low-shelf 22hz 1.6 6dB
gain +6dB
 

  • Like 1
Link to comment
Share on other sites

Terrific results!

 

Very interesting approach, adding the gain back after the correction, as it allows no change to the MV control on the playback system.  Do you find you lose some midbass by using such steep filters?  Or do you prefer a little less midbass in the correction, as most filtered films add more midbass as the  ULF is reduced?

 

JSS

Link to comment
Share on other sites

Depends, as you say, many filtered movies have this mid-bass "boom" lift, and this is automically reduced by the filter, because it overshoots and reduces the level.

Here I have used several filters, just like you have done, to prevent too much overshoot.

None of the alien-films have much "boom" actually, it was basically just this one that had this quite heavy 50hz, but you can't remove it, because that 50hz is the sound design.

 

Just remember that the high-shelf boost-correction can only be applied to the lfe channel, not necessarily obvious to see that, when you only look at the signals from 100hz down..

 

And it did turn out very well, and just because the potential was so good, I continued to work on it until I got it right, it took looong time.

Now, how do you think the next ones are.. because, they are definitely not lesser than this one.

Link to comment
Share on other sites

I am glad that this film you found to be the best, as it is my favorite of the series, including Prometheus.  Funnier than the first (Bill Paxton and 'Sarge' steal the show for me), but just as frightening at times.  Serious sense of doom moments, almost as bad as the "these guys are SCREWED" feeling in Black Hawk Down when I first saw it.

 

JSS

Link to comment
Share on other sites

Yes, I like the 2. best as well, and to find out that the sound was quite good and then could be improved to a full-range experience was even better.

This level of improvement is rare to find in audio today, it's like going from using 2 larger fronts only, with the accompanying sketchy bass reponse, to a well integrated full-range system with subwoofers.

 

The -6dB/+6dB gain is to avoid clipping when processing the signals, the level is a bit high initially on some few peaks. 

 

Alien3 is up next, but I think I will revise the ew for the lcr for this one first, it was the easiest of them to do, but I think there is some improvement to gain by changing the eq a little.

Link to comment
Share on other sites

Alien3 Special Edition Bass EQ

This was the easiest of the Alien films to fix, because it is less severely filtered than the others.
Still, it required several rounds to get it right, and it certainly deserves the attention, as the sound is quite good.

Overall sound kind of improves for the newer releases, at least partly due to improvements in technical performance of sound production equipment.
Less noise, more extended frequency response, and more flexibility for creation of sound design and effects.

The problem with this one was to avoid clipping in the LCR-channels.
There will be some clipping when the last +6dB gain is applied to make the levels match the original, but it is not really noticeable, because the signals are already distorted on those transient peaks, either due to clipping somewhere earlier in the chain, or simply because this is how the sound design was made.
Still, far from todays clipped, compressed and harsh sounding releases.

As the waveforms reveal, there is plenty of ulf in this film, so go on and enter the Bass-EQ filters and turn it up.

Scenecaps were accidentally made with subwoofer level +3dB compared to the others.
Also, the scenecaps include the emergency pod crash in the beginning, and I had this one increased in amplitude by 6-9dB.


Alien3 Special Edition l channel original:

post-181-0-46370100-1420936595_thumb.jpg

Alien3 Special Edition l channel BEQ:

post-181-0-97841000-1420936609_thumb.jpg

Alien3 Special Edition lfe channel original:

post-181-0-34480400-1420936642_thumb.jpg

Alien3 Special Edition lfe channel BEQ:

post-181-0-80431400-1420936662_thumb.jpg

Alien3 Special Edition Scenecap start->0:06:31 original:

post-181-0-22560000-1420937208_thumb.jpg

Alien3 Special Edition Scenecap start->0:06:31 BEQ:
post-181-0-20783300-1420937237_thumb.jpg




MinDSP new filter parameters (use the minidsp spreadsheet to find biquads):

LFE:
low-shelf 24hz q=0.78 gain=12dB
low-shelf 24hz q=0.78 gain=12dB

LCR:
gain -6dB
low-shelf 10hz q=0.78 gain=-6dB
low-shelf 24hz q=0.92 gain=14dB
low-shelf 26hz q=0.96 gain=8dB
low-shelf 40hz q=1.23 gain=-1dB
low-shelf 100hz q=0.71 gain=1dB
gain +6dB


Original SFM parameters:

LFE:
low-shelf 24hz 1.2 12dB
low-shelf 24hz 1.2 12dB

LCR:
gain -6dB
low-shelf 10hz 1.2 -6dB
low-shelf 24hz 1.6 14dB
low-shelf 26hz 1.8 8dB
low-shelf 40hz 3.0 -1dB
low-shelf 100hz 1.0 1dB
gain +6dB
 

Link to comment
Share on other sites

SFM filters only apply to a specific plug-in for audacity, you recognize it if you use it.

.

For MiniDSP, use the MiniDSP new parameters.

For older MiniDSP plug-ins, use the MiniDSP supplied spreadsheet to calculate biquads and enter those.

Old vs new, it is the q-value that does not match up.

 

For JRiver, we should verify that the MiniDSP new parameters are valid.

This can be done by entering a filter in JRriver and compare the predicted frequency response to the one that is calculated in the MiniDSP supplied spreadsheet.

I don't remember if this has been discussed in this thread already.

 

You can always use this spreadsheet to verify the response, by entering the filters in the sheet and compare the calculated response to the predicted JRiver response.

This way you can get it right even if the filter parameters does not match up, becauase you can adjust according to the calculated reference.

 

The gain + - is to avoid clipping when processing the filters.

There may be peaks higher in level earlier in the filter chain, and peaks often will occur at different places.

This is because filters affect the phase of the signal as well as the amplitude, the phase shift can cause large differences in peak amplitude.

Also, I want to see exactly where it clips, if it clips, when doing the final gain adjustment to bring the level back up.

Now you also have the option to leave out the last gain increase, and increase the mastyer volume level instread, just remember to do the gain reduction on all channels first.

Link to comment
Share on other sites

For comparison of the levels in the recorded scene sl plots for the Alien films, here is the Oblivion from start to coming-in-hot:

post-181-0-46181700-1421448123_thumb.jpg

 

The levels are around 10dB lower until the final landing, where the levels are actually higher than really possible, because the waveforms are severely clipped to push the level up as loud as possible, or, rather, louder than possible.

It still sounds reasonably good, because the harmonics are masked - this is mostly a wide-band noise signal.

 

Link to comment
Share on other sites

Kubrick films Bass-EQ

 

While we wait for the final Alien Bass-EQ - the best one for sound-quality - I have some filters for some of the movies from the Kubrick Masterpiece Collection.

 

You do not watch these for the sound, but they can still be improved, with Bass-EQ the overall tonal balance is improved some, and there are occasional moments where low frequencies are noticeable.

 

I have watched Shining and Eyes Wide Shut.

Neither has what you can call good sound, especially dialogue spectral balance sounds very movie-like, this can be improved some by tilting down the upper frequencies during playback (not part of the Bass-EQ), many AVR/processors have a Home-Theater EQ function for this.

Note the scene in Eyes Wide Shut when he enters the Jazz bar - suddenly there is music with great mid-bass punch, it really sounds like a liv performance in a Jazz club.

 

 

Shining:
 
lcr+lfe:
low-shelf 36hz q=0.91 12dB
low-shelf 36hz q=0.91 12dB


A space odyssey:

lcr+lfe:
low-shelf 28hz q=1.07 8dB
low-shelf 28hz q=1.07 8dB


Eyes wide shut:

lcr+lfe:
low-shelf 26hz q=0.92 16dB


Full Metal Jacket:

lcr:
low-shelf 36hz q=0.92 16dB
low-shelf 26hz q=0.90 8dB

lfe:
low-shelf 42hz q=0.91 12dB
low-shelf 36hz q=0.95 6dB
 

Link to comment
Share on other sites

  • 4 weeks later...

Since Avengers 2 is coming up, now would be a good time to start reviewing the lead-up films with BEQ!

Let's start with Iron Man:

Before:

post-20-0-61950100-1384299587.jpg

After:

post-20-0-29739500-1423868857.jpg

The corrections:

LFE:

1. Gain -7dB

2. Low Shelf 16.5Hz, Q of 0.707, +5dB  (4 filters stacked for a total of 20dB)

 

LCR:

1. Gain -7dB

2. Low Shelf 16.5Hz, Q of 0.707, +3dB 

 

Surrounds:

1. Gain -7dB

2. Low Shelf 20Hz, Q of 0.707, +6dB

 

The film had terrific extension present in the LCRS already, hence its terrific reputation.  Adding that kind of extension to the LFE puts this one over the top.

 

JSS 

post-20-0-29739500-1423868857_thumb.jpg

  • Thanks 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...