The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Some thoughts on "high resolution" audio processing Dynamics Plugins
Old 16th December 2014
  #181
Gear Nut
 
dangerousben's Avatar
 

Whoa, I'll check that filter tomorrow!
Old 16th December 2014
  #182
Lives for gear
 
inlinenl's Avatar
 

Verified Member
Thnx Fabian ! I'll play around if work/time permits ....
Old 16th December 2014
  #183
Lives for gear
 
FabienTDR's Avatar
 

Thread Starter
Thanks guys, any input would be greatly appreciated. I personally don't work much at super high rates, maybe someone in touch with DXD or similar biz could give it a try.
Old 16th December 2014
  #184
Gear Maniac
 

Old 16th December 2014
  #185
Lives for gear
 

interesting !

will look to test this.
thx
Old 16th December 2014
  #186
Lives for gear
 
FabienTDR's Avatar
 

Thread Starter
Quote:
Originally Posted by gomfist View Post
ran a short mix hot 4 db into a limiter. with and without TDR @ 88khz

Zippyshare.com - without TDR.wav
Zippyshare.com - with TDR.wav
This is a very interesting example. Am I the only one to hear the mid range opening up with the filter? (transients are minimally softer however)
Old 16th December 2014
  #187
Gear Addict
 
vladg's Avatar
Quote:
Originally Posted by FabienTDR View Post
This is a very interesting example. Am I the only one to hear the mid range opening up with the filter? (transients are minimally softer however)
Yeah, it's really interesting. On my mixcubes (it's too late to turn on "big" sound) "with TDR" has darker sound but with crystal clean mids, while "without TDR" is very bright but with dirty mids. So IMO there's a kind of compromise between 2 different types of sound. I suppose "Amount" knob allows to choose the right one.
Old 16th December 2014
  #188
Quote:
Originally Posted by FabienTDR View Post
Am I the only one to hear the mid range opening up with the filter?
To me the biggest difference seems to appear on the high frequencies.
Old 16th December 2014
  #189
Gear Addict
 
vladg's Avatar
Old 16th December 2014
  #190
Gear Maniac
 

yep i changed a bit but couldnt hear an audible difference.
Old 16th December 2014
  #191
Lives for gear
 
Analogue Mastering's Avatar
 

Verified Member
It seems to sound a bit more liquid/fluid in the midrange, bit more "hifi" if you will, very subbtle, but interesting for sure.
Old 17th December 2014
  #192
Lives for gear
 
Hermetech Mastering's Avatar
 

Verified Member
I definitely prefer the version with it on, but isn't it just a case of slightly rolled off highs? I know that's the point though...
Old 17th December 2014
  #193
Motown legend
 
Bob Olhsson's Avatar
 

Verified Member
Treating ultrasonic response eliminates garbage at lower frequencies. One approach is raising the frequency of the filter and the other is using better filters. The cost of better filters is typically more latency and/or CPU overhead.
Old 17th December 2014
  #194
Lives for gear
Fabien, if we have a clipper before the final limiter, would you put the filter before or after the clipper?
Old 17th December 2014
  #195
Gear maniac
 

Quote:
Originally Posted by FabienTDR View Post
Am I the only one to hear the mid range opening up with the filter? (transients are minimally softer however)
With the TDR filter I do listen to a clearer mid range also!

I remember using the brick-wall lpf in Ozone 5 right before the limiter with similar results.
Old 17th December 2014
  #196
Lives for gear
 
FabienTDR's Avatar
 

Thread Starter
Quote:
Originally Posted by scraggs View Post
Fabien, if we have a clipper before the final limiter, would you put the filter before or after the clipper?
I'd try both. Maybe also before AND after the clipper?

The great strength of this little tool is the bypass button. It's easy to try out and see what sounds better in whatever situation.

This whole concept of IMD reduction in super wide bandwidth system is still very new to me, that's why I'm interested in your own workflows and experiences with the tool.
Old 17th December 2014
  #197
Lives for gear
 
FabienTDR's Avatar
 

Thread Starter
Quote:
Originally Posted by Hermetech Mastering View Post
I definitely prefer the version with it on, but isn't it just a case of slightly rolled off highs? I know that's the point though...
It's certainly a factor, but not really the intent here.

I find it interesting how it cleans up the mid-freq region in busy mixes. The filter's 20-20k passband is very flat (below 0.1dB variance), so this can't relate to masking/unmasking effects. It seems to be the direct result of reduced IMD.

The effect disappears for simple, monophonic kind of material. This again supports the idea.
Old 17th December 2014
  #198
Lives for gear
 

Quote:
Originally Posted by gomfist View Post
ran a short mix hot 4 db into a limiter. with and without TDR @ 88khz

Zippyshare.com - without TDR.wav
Zippyshare.com - with TDR.wav
Can anyone here differentiate these two files under ABX? I've only got headphones and a laptop DAC here atm, and thought I could hear a difference to start with as well ... but the difference between these two files peaks at approx -74 dBFS. I have no doubt that filtering in this way in conjunction with other processing could be more obvious, but this particular example is very, very subtle (IMHO).
Old 17th December 2014
  #199
Lives for gear
 
Analogue Mastering's Avatar
 

Verified Member
listen on range limited speakers (iPhone /iPad / laptop) it gets more clear.
Old 18th December 2014
  #200
Lives for gear
 
stinkyfingers's Avatar
 

Quote:
Originally Posted by -tc- View Post
Can anyone here differentiate these two files under ABX? I've only got headphones and a laptop DAC here atm, and thought I could hear a difference to start with as well ... but the difference between these two files peaks at approx -74 dBFS. I have no doubt that filtering in this way in conjunction with other processing could be more obvious, but this particular example is very, very subtle (IMHO).
also worth noting the difference is >19kHz...
there's one spot in the file (3.724 secs +/-) where something seems to clip in one but maybe not the other ?...but other than that, the residual looks like dither...

imo this whole concept of worrying about a little extra imd when there's more than enough to choke a horse already in the audible band with 'normal' sample rates is just...
it's like people who drink diet soda because they are concerned about their health...
Attached Thumbnails
Some thoughts on "high resolution" audio processing-residual.jpg   Some thoughts on "high resolution" audio processing-residual2.jpg  
Old 18th December 2014
  #201
Lives for gear
 

Quote:
Originally Posted by stinkyfingers View Post
there's one spot in the file (3.724 secs +/-) where something seems to clip in one but maybe not the other ?...but other than that, the residual looks like dither...
It's correlated to the signal; gain it up and have a listen. This is consistent with one file containing more IMD than the other.

However -- a spectrogram shows events indicative of truncation distortion (most obviously between approximately 1.5 and 3 seconds). We can even see this in the time domain by zooming in to the area just before the hit at ~ 2.57 seconds (for example).

This suggests that the diff does not contain dither (which we would expect it to), but that it does contain truncation distortion. There are various ways this could have happened in the processing chain. So how much more *IMD* is there in the unfiltered file? I'm still interested in whether anyone can ABX them

The event at ~ 3.724 seconds could just be the limiter responding differently to the filtered signal, due to the difference in peak level. This is something to keep in mind when slapping this filter in front of another plug-in. If this kind of pre-processing is 'necessary', it really should be incorporated into the processor's design for best results, rather than added separately.

Quote:
Originally Posted by stinkyfingers View Post
it's like people who drink diet soda because they are concerned about their health...
Based on this example alone, it is a bit! I will not have time to experiment further with this TDR filter for about a week ... so will stop nit-picking now.
Old 19th December 2014
  #202
Lives for gear
 

I experimented with the Ultrasonic filter on a just-for-fun master at 96kHz. I had about 10 plugins in a chain (all producing harmonic distortion of various degrees, mainly Acustica Audio Aqua and Nebula instances) and smashed into Voxengo Elephant ridiculously hot (limiting up to 8dB at best). I inserted Ultrasonic filter first in the chain and then after (and thus before) each and every plugin, including after elephant (so actually after the limiting.. just before final dither). The difference is ASTONISHING!! It is so ridiculously obvious and much better with the filters than without. A truly mind-blowing "WTF???!!" experience for me. It basically removed a lot of problems I've identified in mixes and previous masters that had heavy processing with a lot of harmonics.. mainly that nasty build up of "ringing" around 2.3 to 4kHz area. It also cleared out the congested mids and lower mids.

This is a game changer for me. I will not do any more high resolution mastering without this filtering. Period. Heck, I will never mix at high resolution again without this plugin inserted all over the place.
Old 19th December 2014
  #203
Lives for gear
 

Quote:
Originally Posted by bmanic View Post
The difference is ASTONISHING!! It is so ridiculously obvious and much better with the filters than without.
Awesome! Are you able to upload preprocessed, and with/without filter clips?
Old 20th December 2014
  #204
Lives for gear
 
FabienTDR's Avatar
 

Thread Starter
Quote:
Originally Posted by bmanic View Post
I experimented with the Ultrasonic filter on a just-for-fun master at 96kHz. I had about 10 plugins in a chain (all producing harmonic distortion of various degrees, mainly Acustica Audio Aqua and Nebula instances) and smashed into Voxengo Elephant ridiculously hot (limiting up to 8dB at best). I inserted Ultrasonic filter first in the chain and then after (and thus before) each and every plugin, including after elephant (so actually after the limiting.. just before final dither). The difference is ASTONISHING!! It is so ridiculously obvious and much better with the filters than without. A truly mind-blowing "WTF???!!" experience for me. It basically removed a lot of problems I've identified in mixes and previous masters that had heavy processing with a lot of harmonics.. mainly that nasty build up of "ringing" around 2.3 to 4kHz area. It also cleared out the congested mids and lower mids.

This is a game changer for me. I will not do any more high resolution mastering without this filtering. Period. Heck, I will never mix at high resolution again without this plugin inserted all over the place.
Glad that the filter helped! The benefits/effects you are describing are very similar to what I experienced: A surprisingly obvious clean up of the low-mid/mid region with certain types of material.

I haven't tried really aggressive mixing scenarios yet, but the pile of "mud" appearing around 2-4kHz in long chains can certainly be attributed to the standard IMD effect.

@-tc-: I had a very busy month and haven't found the time yet to run proper listening tests and document the whole thing. I'll do that as soon as possible, but until then more uploads and examples would be really helpful indeed!
Old 20th December 2014
  #205
Lives for gear
 

Interest peaked from bmanic post !

Simple question ... If working at 44.1k, would this filter plugin be appropriate to still test ??

as soon as I can, I'm trying to find time to test this.

Very interested.

thx
Old 20th December 2014
  #206
Lives for gear
 
FabienTDR's Avatar
 

Thread Starter
It's meant for high sample-rates. Higher than "needed". The filter automatically disables itself below 88.2kHz.
Old 20th December 2014
  #207
Gear Head
 

My experience is similar to bmanic. Initially tried it in the monitor path, which made a very noticeable difference. Then worked up to what bmanic describes, as well as immediately before and after an analog loop. Seems nuts to use so many filters, but it sounds great. The clarity improvement is really impressive. Is there any potential downside to stacking all these filters?

BTW, thank you Fabien and Vlad for your fantastic work. The Kotelnikov GE is simply excellent.
Old 20th December 2014
  #208
Lives for gear
 
FabienTDR's Avatar
 

Thread Starter
Quote:
Originally Posted by craigb View Post
Is there any potential downside to stacking all these filters?
Yes, there is:

1. They introduce latency. Not much, but it stacks up and could probably trouble the operator if it goes too far.
2. If the input signal contains discontinuities (HF essentially), the removal of these will provoke the Gibbs' effect, and as a result, increase peak level. Not the RMS level however. Because of the filtering, the RMS will even be slightly lower than the original signal.
3. There's a "transient weakening" effect often associated with very steep linear-phase audio filters, especially as soon the cut off frequency is set somewhere within the audible band. This effect is very real, but in this case of our very high cut-off frequencies and relatively shallow slopes, absolutely negligible.


Generally, you can imagine this "loss" as a single, very gentle, high quality AD/DA or SRC stage. The steepest slope in the ultrasonic filter is 2 or 3 times less steep than a typical 44.1kHz AD/DA/SCR steps ask for. Also, we're offering two options meant to precisely control and balance any sort of negative effect.

Make sure to bypass often and check how it interacts with the chain.
Old 20th December 2014
  #209
Lives for gear
 

Quote:
Originally Posted by scraggs View Post
Fabien, if we have a clipper before the final limiter, would you put the filter before or after the clipper?
That would depend if the clipper makes alias tones. Putting a linear phase filter after a hard clipper causes Gibbs effect.
Old 20th December 2014
  #210
Gear interested
 

Fabien, thanks a lot for this investigation! It was quite educational.

However, I'm still not sure if introducing even more audio processing is the right way to fix the IMD issue. Basically, what you're suggesting is cleaning up a mess after someone's bad job. But, if I know that a plug-in A does a dirty job, I won't simply buy it. As a result its developer will start thinking of how to improve it.
Or, if processing at 44/48 gives a cleaner output, then I'll work at that SR. That's all. Otherwise, what's the point of extending freq. range (going to 96 and higher), first, and then cutting it back to the audible one by the pre/post filter? If, after the plug-in A's application we've got something unwanted in the audible range, it's there to stay, no matter whether we did pre/post filtering or not.

I think the important question is do I know this side (IMD/aliasing) of my plug-ins or I just believe they are good? And that is when tests like yours are becoming very important. I think creating testing tools (like VST analyzer) is what we need. And this is something you could easily sell because everybody wants to know a plug-in better before paying for it.
New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+  Submit Thread to Reddit Reddit 
 
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump