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
REW - When Are The Results Good Enough for Pro?
Old 6th November 2016
  #121
Lives for gear
 
jim1961's Avatar
 

Quote:
Originally Posted by Northward View Post
Not really. Broadband means every single filtered ETC is within the set boundaries.

While the unfiltered may show you a room with -60dB in 'x' ms, the filtered ones will show you a different story depending on the selected band that the unfiltered version won't show.

Usually it turns out the 63Hz band will be much 'longer' than the 500Hz one for example.

But an unfiltered response of -60dB in 70ms is somehow showing an already quite short room. The question is if it's balanced or not over the spectrum.
+1

I would ask if bands above 1K are allowed/preferred to roll off a bit. Additionally, if you have a sloped FR curve, which most prefer, I think its ok to let the decay follow. Meaning, if your FR is -3db at 1K compared to 125hz, them its ok that the 1K filtered response is 3db down as well.
Old 12th November 2016
  #122
Quote:
Originally Posted by RJHollins View Post
Great question/Topic.

Hopefully we get better answers than, 'You're done ... When You run out of money'.

Marked and following this thread.

thx
This made me laugh so much, so true. and if I may add when you run out of time or patience, and just accept you've done your best.

But sincerely, may I add some success criteria, I have conceived, from my best experience with some sound systems. It is a bit of a Holy Grail, but with a bit of luck and a lot of effort, and probably money, it can be attained.

When you can hear well produced music sounding stunning and lifelike, and anything with a lower level of definition or sonic quality is startlingly apparent, without any listening effort overhead on your part.

When you can immediately discern the "texture" or layers of processing (read - distortion) that were applied to the mix, especially in any music produced in the recent 30 years, cos you can almost touch the audio, like cloth, to feel its fabric - silk, velvet, cotton, brushed cotton.

When you obtain a 3 dimensional picture and can point at every instrument with absolute pin point accuracy in the 3D audio image in front of you. With the foreground and background absolutely unforgivingly presented, where background singers truly appear to be by the side or behind the lead singer.

When every inflection is exposed, breaths, finger snapping on guitars, or acoustic basses, and you can "hear" the compression applied to the audio -.

Sounds like nirvana, but its achievable, with a bit of effort, and some trial and error.

The challenge with most of us is when we get there we do not know it, cos most of the music we listen to unfortunately is not as well produced as our flawed reproduction systems have deceived us to believe, so we fight this impression of poor quality, that our perfect systems tell us.

What I would say is pick a few absolutely well produced tracks especially of acoustic music, and your best effort will tell you by the quality of life like image presented to your ears. There is a learning curve to retrain the ears, cos at first you may not believe your ears any more, when the audio takes on a 3 dimensional image. How do you describe what cannot be measured!!!
Old 12th November 2016
  #123
Quote:
Originally Posted by sage691 View Post
Yeah, I hear ya !

But a train wreck of a CR room (99.999% of DIY mix rooms) is the equivalent of having 50 lb cement shoes on, whereas the average big wig pro in a pro room has a pair of the best Nike running shoes.

So who (regardless of "skill", "talent") is going to win the 100 yard dash ??

To me this is the elephant in the room that goes largely unseen. The one thing that all of the big wig, hot shot mixing engineers have in common is a GREAT mix room that they fully understand in terms of its' strengths and weaknesses. If you don't know what you are hearing then you are doomed from the starting gate. Your growth will be stunted because point A is really point C,Y,T,E,F and G combined, and point B is somewhere in another universe smoking DMT with Einstien's alter ego .

Everything is a big schizophrenic blur of unreality in 99.9999% of rooms peeps are using for production/mixing. So the shot was fired starting the race, and every time you get up you're in quicksand and you fall right back down on your face -- on and on to infinity !!!

I tried to explain this recently to a very gifted visual artist friend of mine. He understood virtually nothing until i explained it to him in terms he could grasp. I said "imagine that every painting, every sculpting that you work on vigorously for countless hours becomes something COMPLETELY DIFFERENT the moment you cross the threshold of your door with it ?" His reply was that would be absolute madness and would probably make him suicidal !! So you thought you had a masterpiece on the level of Picasso, but outside of your room it's just an unrecognizable blob of madness.

Get your CR/mix room straight, or just throw in the towel. That is my advice.
Spot on. Until the room is optimised to a certain extent, no point in attempting to use DSP to fix early reflections and speaker anomalies. Also from your point the top end studios start off with the best speakers money can buy, so they are already ahead, and we listen to recordings from these studios and wonder - what am I doing wrong !!
Old 15th November 2016
  #124
Lives for gear
I hear ya -- but I honestly believe that the quality of speakers (above the "pro" level threshold) is MUCH less of a factor than CR design/acoustics ! There are speakers in the $2K to $3K range that, in a well designed/treated CR room, will far exceed a pair of $10K-$15K speakers in a haphazardly designed room. The smart people looking to make professional level commercial records/mixes will place all of their resources into getting the CR acoustics right first, and then will think about the gear after that.

I call this putting the cart before the horse. How many times have you seen photos on GS of a 10' X 10' X 10' bedroom with a super expensive pair of Barefoots and just the basic, bare bones room treatment ? I mean, it's as common as a McDonalds hamburger ! They think the super expensive speakers will somehow enable them to overcome the laws of physics. Aint gonna happen.

Then you see some dude really taking room design/treatment, speaker placement, measurements, etc. seriously and he's got a $1000 pair of Mackie monitors -- and it makes you smile that he's one of the intelligent few that will not have to learn the hard way !


Quote:
Originally Posted by kodebode View Post
Spot on. Until the room is optimised to a certain extent, no point in attempting to use DSP to fix early reflections and speaker anomalies. Also from your point the top end studios start off with the best speakers money can buy, so they are already ahead, and we listen to recordings from these studios and wonder - what am I doing wrong !!

Last edited by sage691; 16th November 2016 at 07:49 AM..
Old 7th December 2016
  #125
Gear Nut
 

Thread Starter
Peeling A Giant Union

Quote:
Originally Posted by Northward View Post
Not really. Broadband means every single filtered ETC is within the set boundaries.

While the unfiltered may show you a room with -60dB in 'x' ms, the filtered ones will show you a different story depending on the selected band that the unfiltered version won't show.

Usually it turns out the 63Hz band will be much 'longer' than the 500Hz one for example.

But an unfiltered response of -60dB in 70ms is somehow showing an already quite short room. The question is if it's balanced or not over the spectrum.
Thanks.

Tweaking Room Acoustics = The biggest union you've ever seen, and finding out how many layers you can peel off without Crying.

Old 8th December 2016
  #126
Lives for gear
Quote:
Originally Posted by Sonic Joe View Post
Thanks.

Tweaking Room Acoustics = The biggest union you've ever seen, and finding out how many layers you can peel off without Crying.

This is why IMO 99.9% would be much better off hiring a competent designer if their goal is to be on a level playing field with those who mix commercial records. The onion has so many layers and is so deep that even the most ardent DIY enthusiast's chances are essentially slim and none of ending up with a mix position whose level of neutrality and truthfulness can compare to the big wigs mixing in custom designed CRs.

You can wallow around in this swamp for endless days and hours, only to discover that you are sinking further and may even be doing more harm than good. I had absolutely no idea just how many "deal breaking" problems I had with my CR/mix position; it was nothing short of shocking to realize how far off the mark I was, and also how unlikely that I would have ever peeled back the onion that far on my own to even realize it ! There's a lot of back patting that goes on on these forums for results that no pro mixing commercial records would accept; and also a lot of BS fake Houdini mysticism surrounding big wig gurus who supposedly can mix hit records in a cardboard box, but ironically choose to work in $200,000 custom designed CRs instead !

99.9% of acoustics DIY folks are in cement shoes trying to run a race with peeps that have the best pair of Nike's on their feet. Hire a pro and when they break down the REW data for you then you will see why -- and cry !!

IMO this is the one corner you just cannot cut in dis game.
Old 3rd January 2017
  #127
Gear Nut
 

Thread Starter
Happy New Year !

Reflecting back on 2016, I spent a considerable amount of time working ON my room. Ultimately, I'm a tweeker and will never be satisfied, but I'm going to spend 2017 working IN my room.

I started this thread hoping to set some attainable Pro Level targets for my one room, multi use (tracking, mixing, mastering) studio. I understand all rooms are different, and ever person/Budget has different goals.

Anyone who has spent time working on their room acoustics knows, you get to a point where nearly Anything you do to the room makes one thing better and another thing worse. That is where the real world "choices" are made. Maybe you give a little in SPL/FR to get a tighter response on the ETC for better imaging, etc., etc., etc.

That said, I still believe there are some Common metrics/targets that are reasonable for any Pro Level room.

Here are the targets I ended up focusing on in 2016, that IMHO are good enough for Pro.

1. All SPL (frequency response) +/- 6dB at 1/24 smoothing (I personally have a fairly Flat Target, which I know some debate, but ideally with a subtle tilt from upper left low end to lower right high end but still within the +/-6dB range)

2. IR / ETC no reflections above -20 and no reflections above-30 beyond 5ms with an even decay until down -60dBFS by 70ms non-filtered/broadband. Of course you can start peeling the onion with the filtered IR, but IMO, unless the room was purpose built from the ground up it would be very difficult to attain this target at all filtered bands.

3. Topt or T20 below .15 at 500Hz (neither the most reliable in small rooms, but still useful to support you personal Felling of how dry/tight the room sounds)

4. 320ms Maximum decay time on waterfall graph (over a 40db range, focusing in on 20Hz - 400Hz)

All the best to Making Music in 2017!

Joe
Old 7th January 2017
  #128
Lives for gear
 

Quote:
Originally Posted by Sonic Joe View Post
Reflecting back on 2016, I spent a considerable amount of time working ON my room. Ultimately, I'm a tweeker and will never be satisfied, but I'm going to spend 2017 working IN my room.

I started this thread hoping to set some attainable Pro Level targets for my one room, multi use (tracking, mixing, mastering) studio. I understand all rooms are different, and ever person/Budget has different goals.

Anyone who has spent time working on their room acoustics knows, you get to a point where nearly Anything you do to the room makes one thing better and another thing worse. That is where the real world "choices" are made. Maybe you give a little in SPL/FR to get a tighter response on the ETC for better imaging, etc., etc., etc.

That said, I still believe there are some Common metrics/targets that are reasonable for any Pro Level room.

Here are the targets I ended up focusing on in 2016, that IMHO are good enough for Pro.

1. All SPL (frequency response) +/- 6dB at 1/24 smoothing (I personally have a fairly Flat Target, which I know some debate, but ideally with a subtle tilt from upper left low end to lower right high end but still within the +/-6dB range)

2. IR / ETC no reflections above -20 and no reflections above-30 beyond 5ms with an even decay until down -60dBFS by 70ms non-filtered/broadband. Of course you can start peeling the onion with the filtered IR, but IMO, unless the room was purpose built from the ground up it would be very difficult to attain this target at all filtered bands.

3. Topt or T20 below .15 at 500Hz (neither the most reliable in small rooms, but still useful to support you personal Felling of how dry/tight the room sounds)

4. 320ms Maximum decay time on waterfall graph (over a 40db range, focusing in on 20Hz - 400Hz)

All the best to Making Music in 2017!

Joe
Great discussion topic and thanks for sharing your conclusions!
Old 11th January 2017
  #129
Gear Maniac
 
bukarin's Avatar
 

Very interesting thread!! Wish I'd read this before building my new studio!! Filtering the impulse responses in REW has opened up a lot more info for me to process. One thing that has gone un-noticed in this thread is something Thomas from Northward Acoustics said regarding speaker response and subsequently removing from your measurements. I've got a pair of ATC SCM100a flush mounted on heavy dummy loaded concrete plinths on springs in a NE style room with huge rear bass traps (7.5m3 of low density mineral wool on the back wall) and full coverage of all surfaces with at least 200mm low density mineral wool.

The speakers are ported and I believe the port is tuned to 30Hz approx. How will that effect the decay of the room in that area? The room is very controlled with a noticeably longer decay at around 28Hz (approx 125ms longer than all other frequencies). I don't want to post data or anything yet as I'm still tweaking but how do you really remove the speaker response from the room measurements? Can't find any useful waterfalls or graphs for the ATCs but I know even a well designed ported system can add a longer decay centred around its tuned frequency. If you've got a seriously well treated room can you in theory see the effect of a bass port in the room measurements?
Old 11th January 2017
  #130
Gear Guru
Minus

Block the bass ports temporarily and measure again. You will need to raise the LF level to equalise the room stimulus to make a comparison. A simple LF boost using Eq.
DD
Old 11th January 2017
  #131
Lives for gear
 
Jens Eklund's Avatar
Quote:
Originally Posted by bukarin View Post
Very interesting thread!! Wish I'd read this before building my new studio!! Filtering the impulse responses in REW has opened up a lot more info for me to process. One thing that has gone un-noticed in this thread is something Thomas from Northward Acoustics said regarding speaker response and subsequently removing from your measurements. I've got a pair of ATC SCM100a flush mounted on heavy dummy loaded concrete plinths on springs in a NE style room with huge rear bass traps (7.5m3 of low density mineral wool on the back wall) and full coverage of all surfaces with at least 200mm low density mineral wool.

The speakers are ported and I believe the port is tuned to 30Hz approx. How will that effect the decay of the room in that area? The room is very controlled with a noticeably longer decay at around 28Hz (approx 125ms longer than all other frequencies). I don't want to post data or anything yet as I'm still tweaking but how do you really remove the speaker response from the room measurements? Can't find any useful waterfalls or graphs for the ATCs but I know even a well designed ported system can add a longer decay centred around its tuned frequency. If you've got a seriously well treated room can you in theory see the effect of a bass port in the room measurements?
If ATC´s, I wouldn’t worry too much about decay due to ported design: https://www.gearslutz.com/board/12081225-post56.html
Old 11th January 2017
  #132
Gear Maniac
 
bukarin's Avatar
 

So I just did a very quick test by blocking the port with a fleece and it made absolutely no difference! Not 100% sure if that is fully conclusive but its put my inquisitive mind at rest regarding that particular small issue for me.

I'm still curious about this issue of removing the speaker response from the room response to fully appreciate whats going on acoustically? I'm thinking that for a great deal of room/speaker systems acoustic/modal issues are so much more significant than pure speaker issues that its not top of the priority list for many.

In a super high end room can you measure anomalies caused by crossovers and ports etc?
Old 11th January 2017
  #133
Gear Guru
Isolate

You can measure speaker anomalies outdoors. You can be a tad more conclusive. I presume the fleece was fairly packed, i.e. impermeable? If not try a cork or something like that. The 28Hz extension is most likely a room mode or anomaly. A simple sine wave should let you know. What are the dimensions of your room, before treatment?
DD
Old 11th January 2017
  #134
Lives for gear
 

Start filtering

Quote:
Originally Posted by Sonic Joe View Post
2.... Of course you can start peeling the onion with the filtered IR, but IMO, unless the room was purpose built from the ground up it would be very difficult to attain this target at all filtered bands...
Happy new year! I like this thread

Frequency Response
We routinely show filtered versions of our Frequency Response.
Like 1/48 octave or 1/24 or wider BW. Why?
-Because unfiltered FR is impossible to read (pic 1) if we haven’t attenuated reflections by absorption, gating or filtering (smoothing) pic 2

As Time domain is the inverse of the frequency domain...

ETC looks quite OK pic 3. Reflections at least 15 dB down but...
ETC ”favours” high frequencies. E.g.In the unfiltered ETC plot all low frequencies are hidden. huh

Start carefully with 250 Hz 1/1 octave where desks, racks, computer-monitors begin ”to be large” compared to wave length. So why refuse some filtering?

Give it a try and you’ll be surprised . Pic 4
Let 2017 be the first year in GS history where ETC plots are routinely filtered.
Attached Thumbnails
REW - When Are The Results Good Enough for Pro?-1.-unsmoothed.jpg   REW - When Are The Results Good Enough for Pro?-2.-1-48-smooth.jpg   REW - When Are The Results Good Enough for Pro?-3.-etc-no-filter.jpg   REW - When Are The Results Good Enough for Pro?-4.-etc-250-1-1-smooth.jpg  
Old 3 weeks ago
  #135
Gear Nut
Quote:
Originally Posted by akebrake View Post
Start carefully with 250 Hz 1/1 octave where desks, racks, computer-monitors begin ”to be large” compared to wave length. So why refuse some filtering?

Give it a try and you’ll be surprised . Pic 4
An old thread, but...
The lower ETC bands have inherently very poor time resolution because of the filtering itself. You can try importing a .wav containing just a unit impulse, and looking at the filtered ETC: 63 Hz goes below -20 dB at 16 ms, 125 Hz at 8 ms and 250 Hz at 4 ms.
Old 3 weeks ago
  #136
Lives for gear
Quote:
Originally Posted by DanDan View Post
+1 The published Broadcast Standards are well worth a look, including any BBC ones you might find in their R&D archives.
Anyone got a link to the BBC specs ? Couldn't find them anywhere.
Old 2 weeks ago
  #137
Gear Guru
Links

Old 2 weeks ago
  #138
Lives for gear
 

Time resolution

Quote:
Originally Posted by Antti H View Post
An old thread, but...
The lower ETC bands have inherently very poor time resolution because of the filtering itself. You can try importing a .wav containing just a unit impulse, and looking at the filtered ETC: 63 Hz goes below -20 dB at 16 ms, 125 Hz at 8 ms and 250 Hz at 4 ms.
Thanks for your additional comment. Yes, filter response time is important.
But isn't the "zero phase filtering" (REW filtered IR graph panel) a method to "limit" the filter time influence when checking Path Length Difference for example?

Best

Ake

BTW I am not sure everybody is familiar with "a unit impulse" (me included)
Would you mind expanding on that and show an example? (Graph?)
Old 2 weeks ago
  #139
Gear Nut
Quote:
Originally Posted by akebrake View Post
Thanks for your additional comment. Yes, filter response time is important.
But isn't the "zero phase filtering" (REW filtered IR graph panel) a method to "limit" the filter time influence when checking Path Length Difference for example?

Best

Ake

BTW I am not sure everybody is familiar with "a unit impulse" (me included)
Would you mind expanding on that and show an example? (Graph?)
Zero phase filtering means the center of the impulse doesn't shift (the spreading of the IR is symmetrical) and also the decay is shorter than with minimum phase, but the width of the transition band (aka steepness of the low and highpass filters at band edges) sets a fundamental limit on how little time domain spreading there can be. Unfortunately REW doesn't allow viewing multiple bands at once, so there's no easy way to focus on the low frequency IR without significant spreading (0-350 Hz filter has less than half the spreading of 180-350 Hz filter).

Unit impulse is simply a signal that's 1.0 at a single time instant and zero everywhere else. So a wav file that's digital silence except a single sample is set to 1.0 (or some other amplitude, it doesn't matter for this example).
Old 2 weeks ago
  #140
Lives for gear
 

Quote:
Originally Posted by akebrake View Post
BTW I am not sure everybody is familiar with "a unit impulse" (me included)
Would you mind expanding on that and show an example? (Graph?)
https://en.wikipedia.org/wiki/Kronecker_delta
fun stuff, man
Old 2 weeks ago
  #141
Gear Guru
Aunty

Old 1 week ago
  #142
Lives for gear
 

Good enough decay

Quote:
Originally Posted by Antti H View Post
... Unfortunately REW doesn't allow viewing multiple bands at once, so there's no easy way to focus on the low frequency IR without significant spreading (0-350 Hz filter has less than half the spreading of 180-350 Hz filter)...
Thanks for the lesson!

But are you saying that 1/1 oct bandpass filtering is not a valid measure of decay time in the low end?

I think it shows a lot more detail than a waterfall. e.g. the smoothness of decay plus the Schroeder plot

Plus we have an excellent referens (LINK) of really pro room. Thanks Northward.

Antti, what is your prefered method of decay measurement?

Best
Old 1 week ago
  #143
Lives for gear
 

Quote:
Originally Posted by OpusOfTrolls View Post
Thanks. I recall a lecture by the late Richard Heyser many moons ago. He explained Dirac Delta functions. Unfortunately I still don't have the math required. Infinite magnitude, infinitely narrow pulses...

Best
Old 1 week ago
  #144
Gear Nut
Quote:
Originally Posted by akebrake View Post
But are you saying that 1/1 oct bandpass filtering is not a valid measure of decay time in the low end?
1/1 oct bandpass at or below 500 Hz as currently implemented in REW has too poor time resolution to pick up individual reflections or indeed say that much about the first 20 ms. It can tell you that you that you have problems, but it doesn't let you correlate them to anything. IOW, it's mostly useless when trying to solve first reflections issues.

It's perfectly fine for viewing the overall shape of the decay, of course. IOW, as an RT60 replacement or complement to waterfall.

I've found that manually exporting the IR, applying a roughly 3 dB/oct lowpass filter starting from 150 Hz or so (one or two shelving EQs really) and importing that back to REW is great for showing the effect of low frequency attenuation while keeping the ETC clear enough to see which reflection is which. The normal unfiltered ETC plot overemphasizes high frequencies and the 3 dB/oct tilt fixes that.
Old 1 week ago
  #145
Lives for gear
 

Pink Filter

Quote:
Originally Posted by Antti H View Post
1/1 oct bandpass at or below 500 Hz as currently implemented in REW has too poor time resolution to pick up individual reflections or indeed say that much about the first 20 ms. It can tell you that you that you have problems, but it doesn't let you correlate them to anything. IOW, it's mostly useless when trying to solve first reflections issues.
I suspected that...

Quote:
It's perfectly fine for viewing the overall shape of the decay, of course. IOW, as an RT60 replacement or complement to waterfall.
Good

Quote:
I've found that manually exporting the IR, applying a roughly 3 dB/oct lowpass filter starting from 150 Hz or so (one or two shelving EQs really) and importing that back to REW is great for showing the effect of low frequency attenuation while keeping the ETC clear enough to see which reflection is which. The normal unfiltered ETC plot overemphasizes high frequencies and the 3 dB/oct tilt fixes that.
Interesting! Thanks for that tip, Antti!
I guess we can call it ”partly Pink Filtered ETC”.
Have tested built in LP filtering on other platforms long time ago but have not tried that with REW.
Learning is fun!

Cheers
Old 1 week ago
  #146
Lives for gear
 

Quote:
Originally Posted by akebrake View Post
Thanks. I recall a lecture by the late Richard Heyser many moons ago. He explained Dirac Delta functions. Unfortunately I still don't have the math required. Infinite magnitude, infinitely narrow pulses...

Best
A continuous time system has "infinite" as an attribute, but the measurements of audio are not, it is discrete time and bound by those lovely nyquist-shannon theories. The Kronecker delta is bandlimited by the sampling rate.
Loading mentioned products ...
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 
 
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