ultra-sharp linear phase filter?
kosmokrator
Thread Starter
#1
7th May 2013
Old 7th May 2013
  #1
mymixisbetterthanyours!
 

Thread Starter
ultra-sharp linear phase filter?

Hey,
for a forensic job I need a linear-phase filter with ultra-small bandwidth (think 2-5Hz). For offline processing.

Normally this is accomplished with the filtfilt-function in Mathlab, but I'd like to stay in my DAW of choice.

Any ideas? I'm unsing WL and/or PTHD on Mac.

Thanks!
#2
7th May 2013
Old 7th May 2013
  #2
are you looking for a liner phase EQ ?
or a way to process w/EQ linearly ?
i have not found any LP EQ i am really happy with, so for offline processing, i use a 'regular' digital EQ and process the file forwards and backwards.
you could use something like Mbandpass and get up to 240dB slope...

edit: maybe something like ReaFir would work
or EQuilibrium and stack filters (many options for type)
also, i thought that said 2-5 kHz in the op, not 2-5 Hz...that's tight
#3
8th May 2013
Old 8th May 2013
  #3
Gear maniac
 
Miiika's Avatar
 

Quote:
Originally Posted by stinkyfingers View Post
are you looking for a liner phase EQ ?
or a way to process w/EQ linearly ?
i have not found any LP EQ i am really happy with, so for offline processing, i use a 'regular' digital EQ and process the file forwards and backwards.
you could use something like Mbandpass and get up to 240dB slope...
You process the file forwards and backwards? Could you please explain? Thank you.
#4
8th May 2013
Old 8th May 2013
  #4
Quote:
Originally Posted by Miiika View Post
You process the file forwards and backwards? Could you please explain? Thank you.
well, i lied about the order...actually i process it backwards and then forwards, but i don't think it makes a difference, it's just ocd i guess.
anyway, you just reverse the audio track with an audio editor, process it with an EQ, reverse the track back to normal and process again with the same EQ / settings.

here's a link to a more 'technical' explanation.
Phase Response
Quote
1
#5
8th May 2013
Old 8th May 2013
  #5
Gear maniac
 
Miiika's Avatar
 

Thanks, stinkyfingers, interesting. So I assume if you wanted to say have ab EQ boost of 3dB, you would boost 1.5dB backwards and 1.5dB forwards. Gotta be a king at math for that stuff ;-)
#6
8th May 2013
Old 8th May 2013
  #6
yes, boost / attenuation will be x2, and so will slope for HPF / LPF.
DAWs should have a calculator on the master bus...

edit:
maybe something like S.P.E.A.R. would work for the op. or Expert Sleepers' Spectral Conquest. (some type of spectral editor maybe)
#7
8th May 2013
Old 8th May 2013
  #7
Gear addict
 
NotchontheRocks's Avatar
 

Quote:
Originally Posted by stinkyfingers View Post
well, i lied about the order...actually i process it backwards and then forwards, but i don't think it makes a difference, it's just ocd i guess.
anyway, you just reverse the audio track with an audio editor, process it with an EQ, reverse the track back to normal and process again with the same EQ / settings.

here's a link to a more 'technical' explanation.
Phase Response
Interesting. Does this really make a difference in how it sounds, or is the "OCD" claim accurate? The theory sounds solid, but it may be a waste of my time if it does not produce any noticeable results. Also, I'd like to know how many MEs actually do this type of processing. I'm assuming the sonic "color" of the EQ plays a major role as well since you are essentially doubling it.

Sent from my DROID BIONIC
#8
8th May 2013
Old 8th May 2013
  #8
Quote:
Originally Posted by kosmokrator View Post
Hey,
for a forensic job I need a linear-phase filter with ultra-small bandwidth (think 2-5Hz). For offline processing.

Normally this is accomplished with the filtfilt-function in Mathlab, but I'd like to stay in my DAW of choice.

Any ideas? I'm unsing WL and/or PTHD on Mac.

Thanks!
Have you looked at SplineEQ (Photosounder)?
huejahfink
Verified Member
#9
8th May 2013
Old 8th May 2013
  #9
Lives for gear
 
huejahfink's Avatar
 

Verified Member
The hum-remover tool in Izotope RX offers extremely narrow cuts, can be linear or regular min phase.
#10
8th May 2013
Old 8th May 2013
  #10
Quote:
Originally Posted by NotchontheRocks View Post
Interesting. Does this really make a difference in how it sounds, or is the "OCD" claim accurate? The theory sounds solid, but it may be a waste of my time if it does not produce any noticeable results. Also, I'd like to know how many MEs actually do this type of processing. I'm assuming the sonic "color" of the EQ plays a major role as well since you are essentially doubling it.
i doubt there is an audible difference. it might somewhat depend on the EQ / filter 'quality'. but i would think it would work the same either way or maybe it's broken?
i'm speaking of (itb) 'digital' EQ and nothing with harmonic distortion.
it is probably is a waste of time if you have a LP EQ that gives you the desired results.
but if a situation arises where your EQ can't 'get it done', then this is an alternate to try.

ex...i don't know too many LP EQ's with a HPF you can set down @ 1Hz (most are 20 Hz, maybe 10 Hz the lowest), but you can do that with SonEQ if you wanted too..etc.
or if you just have an EQ you like the filters on but it's not LP, you can make it LP.
#11
8th May 2013
Old 8th May 2013
  #11
Gear addict
 
NotchontheRocks's Avatar
 

Quote:
Originally Posted by stinkyfingers View Post
i doubt there is an audible difference. it might somewhat depend on the EQ / filter 'quality'. but i would think it would work the same either way or maybe it's broken?
i'm speaking of (itb) 'digital' EQ and nothing with harmonic distortion.
it is probably is a waste of time if you have a LP EQ that gives you the desired results.
but if a situation arises where your EQ can't 'get it done', then this is an alternate to try.

ex...i don't know too many LP EQ's with a HPF you can set down @ 1Hz (most are 20 Hz, maybe 10 Hz the lowest), but you can do that with SonEQ if you wanted too..etc.
or if you just have an EQ you like the filters on but it's not LP, you can make it LP.
Thanks. I may give this a try later and compare for any audible differences. Probably just going to use Waves REQ or HEQ. Hopefully one of those is sufficient.

Sent from my DROID BIONIC
Hermetech Mastering
Verified Member
#12
8th May 2013
Old 8th May 2013
  #12
...

Last edited by Hermetech Mastering; 8th May 2013 at 01:40 PM.. Reason: OT
kosmokrator
Thread Starter
#13
8th May 2013
Old 8th May 2013
  #13
mymixisbetterthanyours!
 

Thread Starter
Could we keep this on-topic please guys?

Thanks for the suggestions, but this seems all like a workaround. I need a solution that is quick and reliable. Just a linear/zero-phase bandbass-filter that is VERY steep.

Nothing out there?
#14
8th May 2013
Old 8th May 2013
  #14
what do you consider 'very steep' ?
and
do you have any idea what the ringing will be like from LP filters that steep ?

i would say EQuilibrium is your best bet.
what you want to do is 'unconventional', so it may require unconventional methods...
kosmokrator
Thread Starter
#15
8th May 2013
Old 8th May 2013
  #15
mymixisbetterthanyours!
 

Thread Starter
Quote:
Originally Posted by stinkyfingers View Post
what do you consider 'very steep' ?
and
do you have any idea what the ringing will be like from LP filters that steep ?

i would say EQuilibrium is your best bet.
what you want to do is 'unconventional', so it may require unconventional methods...
How steep?
As I said, in the vicinity of 3-10Hz.
Ringing or anything 'sound'-related' is irrelevant, since this is purely for forensic purposes. The only requirements are linear/zero-phase, bandpass, and VERY high Q.
#16
8th May 2013
Old 8th May 2013
  #16
Gear maniac
 
Miiika's Avatar
 

DDMF's LP10 eq has a very steep bandpass called BPF2. It goes much steeper than in my picture.
Attached Thumbnails
ultra-sharp linear phase filter?-lpeq.jpg  
Hermetech Mastering
Verified Member
#17
8th May 2013
Old 8th May 2013
  #17
You're using the older version of DDMF LP10, any reason why? Having said that, the new version doesn't have BPF2 so...
#18
8th May 2013
Old 8th May 2013
  #18
Gear maniac
 
Miiika's Avatar
 

Quote:
Originally Posted by Babaluma View Post
You're using the older version of DDMF LP10, any reason why? Having said that, the new version doesn't have BPF2 so...
I haven't used it (or any other LP eq) for years, so I didn't get the new version when it came out.
Quote
1
huejahfink
Verified Member
#19
8th May 2013
Old 8th May 2013
  #19
Lives for gear
 
huejahfink's Avatar
 

Verified Member
Quote:
Originally Posted by kosmokrator View Post
Thanks for the suggestions, but this seems all like a workaround. I need a solution that is quick and reliable. Just a linear/zero-phase bandbass-filter that is VERY steep.

Nothing out there?
Perhaps you missed this post....

Quote:
Originally Posted by huejahfink View Post
The hum-remover tool in Izotope RX offers extremely narrow cuts, can be linear or regular min phase.
How is this a "workaround" exactly? I find it very quick and precise.
kosmokrator
Thread Starter
#20
8th May 2013
Old 8th May 2013
  #20
mymixisbetterthanyours!
 

Thread Starter
Quote:
Originally Posted by huejahfink View Post
Perhaps you missed this post....



How is this a "workaround" exactly? I find it very quick and precise.
Because I need a bandpass, not a notch filter.
Or am I missing something?
huejahfink
Verified Member
#21
8th May 2013
Old 8th May 2013
  #21
Lives for gear
 
huejahfink's Avatar
 

Verified Member
Ah ok I see.... I misread that.

In which case - Ozone 5 EQ. Two brick wall filters?
mastertone
Verified Member
#22
8th May 2013
Old 8th May 2013
  #22
Gear maniac
 
mastertone's Avatar
 

Verified Member
Check out toneboosters flx, lots of interesting filter types.

Sent from my GT-I9300
#23
8th May 2013
Old 8th May 2013
  #23
Lives for gear
 
thermos's Avatar
Equilibrium can do up to a q of 50 (!) notch in linear phase mode. I can't imagine you wanting anything tighter than that.
#24
9th May 2013
Old 9th May 2013
  #24
Gear addict
 

kosmokrator
Thread Starter
#25
9th May 2013
Old 9th May 2013
  #25
mymixisbetterthanyours!
 

Thread Starter
Quote:
Originally Posted by thermos View Post
Equilibrium can do up to a q of 50 (!) notch in linear phase mode. I can't imagine you wanting anything tighter than that.
Q of 50 or more is what I need, but notch is not useful for me.
I need a bandpass filter.
huejahfink
Verified Member
#26
9th May 2013
Old 9th May 2013
  #26
Lives for gear
 
huejahfink's Avatar
 

Verified Member
If the ozone 5 brick wall filters don't do it, then you could always try mixing a notched version back into the original with inverted polarity.
kosmokrator
Thread Starter
#27
9th May 2013
Old 9th May 2013
  #27
mymixisbetterthanyours!
 

Thread Starter
Quote:
Originally Posted by huejahfink View Post
If the ozone 5 brick wall filters don't do it, then you could always try mixing a notched version back into the original with inverted polarity.
This has to be a transparent process without any 'workarounds' since it will be used in court.

Thanks for all your answers though.
Seems I have to stay in the forensic DAW for this.
huejahfink
Verified Member
#28
9th May 2013
Old 9th May 2013
  #28
Lives for gear
 
huejahfink's Avatar
 

Verified Member
RX - use frequency selection tools to select what you DON'T want and then press delete.

RX will print out a process list as well that can be used as part of the evidence I believe.
#29
10th May 2013
Old 10th May 2013
  #29
Lives for gear
 
mdoelger's Avatar
Quote:
Originally Posted by kosmokrator View Post
This has to be a transparent process without any 'workarounds' since it will be used in court.

Thanks for all your answers though.
Seems I have to stay in the forensic DAW for this.
Just to remind you. If you introduce extreme ringing, it's not really transparent anymore.
kosmokrator
Thread Starter
#30
10th May 2013
Old 10th May 2013
  #30
mymixisbetterthanyours!
 

Thread Starter
Quote:
Originally Posted by mdoelger View Post
Just to remind you. If you introduce extreme ringing, it's not really transparent anymore.
I'm fully aware of the ringing problem of linear phase hi-Q filters.
Again, that's not relevant for this particular problem.
'Transparent' here means transparent in a chain of evidence-sense. Nothing audio-related.
New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+ 
 
Topic:
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
 
Register FAQ Search Today's Posts Mark Forums Read

SEO by vBSEO ©2011, Crawlability, Inc.