Before posting your measurement results
Old 1st December 2012
  #31
Lives for gear
Quote:
Originally Posted by jim1961 View Post
Between what and what?
Between your preferred range (450 - 5.6K) and going up to 10-20K.
Old 1st December 2012
  #32
Lives for gear
 
jim1961's Avatar
 

Before posting your measurement results-12-01-etc-left-450-5.6k-14k.jpg

Blue = 450 - 14k
Black = 450 - 5.6k

I used 200us smoothing in order to better see things.

------------ ------------- --------------- ---------------

Perhaps I am interpreting the data wrong, but extending the ETC response to 14k gives me a false sense to how strong the reflections are where they count most. On average, the 5db drop makes one think your reflections are more tamed than they actually are.

Whats at work here is that lower frequency reflection energy in most rooms is louder than higher ones. So the more of the higher frequencies you include in your ETC, the lower the overall average will look. This can be illustrated using 1 octave bands. (see next post)
Old 1st December 2012
  #33
Lives for gear
 
jim1961's Avatar
 

Before posting your measurement results-12-01-band-study.gif

Purple = 11-22k
Green = 5.5-11k
Yellow = 2.8-5.6k
Orange - 1.4k-2.8k

Although a band study of this kind may give very different results depending on ones specific room, what one has keep in mind is the broader your freq range for the ETC, the more it averages the content.

But the above shows why in MY ROOM, the ETC magnitudes drop as you include more high freq data. Much of this is due to it being a carpeted room. Some of its due to the fact that high frequency is stopped in its tracks much easier than low frequencies.

So in the end, it seems including those bands that include where most of the music is seems a good compromise for running one graph. But I also think 1 octave band studies are revealing to give one a sense of not only the time and amplitude domain, but the frequency domain as well.
Old 5th December 2012
  #34
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Old 5th December 2012
  #35
Gear Guru
 
DanDan's Avatar
FuzzMeasure

A Lynch Mob?

I can find no FFT setting in FM, and the Manual is freely available.

FM is a 32 Bit app.

I would assume Record resolution is up to the interface.

Bit depth is selectable during IR Export. 16/24/32

Incidentally, REW is a 16 Bit app, and when queried JohnPM has presented a cogent case as to why this resolution is totally acceptable for IR work.

I imported the IR's into FM and REW.

In FM the Peak level on the Magnitude scale is -5dB while reading -15dB on the ETC. This looks quite usual if not normal to me. I would rather see dBFS than Magnitude, or misleading terms like SPL.

Quote:
The Magnitude Response graph displays the result of a Fourier transform contained within the analysis window for the selected impulse response(s). The window’s shape (i.e. Hamming, Half- Hamming, etc.) is taken into account for this calculation.
REW confirms -15dB on the ETC. I rather see them a bit higher, but with FM's 32 Bit resolution and normalise, perhaps these old level analogue or 16 Bit level setting habits should probably be relegated to history.

I always recommend driving the room to ear protection level, and Four 10Second sweeps.

Again I would welcome meters which show dBFS while monitoring input with the usual Over indicators. Just as in most DAW software.
The Graphs should obviously read the same dBFS levels with other Y choices being secondary.

I have not seen such a short and defined decay before. This issue is particular to this measurement only.

FM has two strange behaviours. Visual anomalies can occur at the decay tail. Making the Window as long as the graph range fixes the visual.
The other is the default Watefall settings, these are the two well known issues.

Conversely, FuzzMeasure, has welcome features not seen elsewhere. e.g. Sophisticated combining of measures. Multichannel recording capability.
FM works well with all interfaces, so it is the better tool for taking measurements for Mac users.

EDIT, I am very appreciative of the great piece of work Chris has done in providing us with a wonderful tool. This in the face of very unbalanced competition, i.e. Zero cost.
We may recall when REW didn't function on Mac at all, so FM was the only game in town.
I encourage purchase of his product and support of his ongoing activities.
I also encourage Reading of the Manual.



DD
Old 5th December 2012
  #36
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
A Lynch Mob?

I can find no FFT setting in FM, and the Manual is freely available.

FM is a 32 Bit app.

I would assume Record resolution is up to the interface.

Bit depth is selectable during IR Export. 16/24/32

Incidentally, REW is a 16 Bit app, and when queried JohnPM has presented a cogent case as to why this resolution is totally acceptable for IR work.

I imported the IR's into FM and REW.

In FM the Peak level on the Magnitude scale is -5dB while reading -15dB on the ETC. This looks quite usual if not normal to me. I would rather see dBFS than Magnitude, or misleading terms like SPL.



REW confirms -15dB on the ETC. I rather see them a bit higher, but with FM's 32 Bit resolution and normalise, perhaps these old level analogue or 16 Bit level setting habits should probably be relegated to history.

I always recommend driving the room to ear protection level, and Four 10Second sweeps.

Again I would welcome meters which show dBFS while monitoring input with the usual Over indicators. Just as in most DAW software.
The Graphs should obviously read the same dBFS levels with other Y choices being secondary.

I have not seen such a short and defined decay before. This issue is particular to this measurement only.

FM has two strange behaviours. Visual anomalies can occur at the decay tail. Making the Window as long as the graph range fixes the visual.
The other is the default Watefall settings, these are the two well known issues.

Conversely, FuzzMeasure, has welcome features not seen elsewhere. e.g. Sophisticated combining of measures. Multichannel recording capability.
FM works well with all interfaces, so it is the better tool for taking measurements for Mac users.



DD
Chris is going to look at the issues regarding FM levels: Finished treating room - graphs good enough?

And regarding resolution:
Before posting your measurement results
Finished treating room - graphs good enough?
Did my first measurement! Need help to analyse it.



EDIT:

And REW is not a "16 bit app": Before posting your measurement results
Old 5th December 2012
  #37
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
A Lynch Mob?
There is no software that is completely bug free. Finding and addressing issues (assuming there is one here) is what makes good software great.
Old 5th December 2012
  #38
Gear Guru
 
DanDan's Avatar
Tests

No doubt, as an early adopter and user of FM when there was no other option, I have always provided feedback to Chris on an ongoing basis.

I see the same Chris has offered to do tests specified by yourself to illuminate any suspected flaws. In his own forum.

I would more than happy to do similar tests right here for you. This diffusion over several threads and now a different forum is confusing to me. I can't distill what particular issue or issues are being referred to within all those other linked threads , but will gladly try to explain and test any particular one if raised directly here.

IF there is an issue, let's find it. In the meantime I suggest an innocent assumption.
This is a commercial product. A one man company. His living.

To be clear.

The 250mS issue is particular to this measurement here in this thread.
I have not seen anything like it before.

The lack of an absolute level meter in FM makes level setting difficult. REW does meter dBFS but since the meter can be Calibrated I am unsure if it is absolute.
Perhaps the Clip/Over indicator remains absolute?

I recommend driving the room fully, then adjusting the mic pre gain until the graphs show sensible levels. e.g. ETC within 10dB or so of Full scale.
If the audio interface has meters, none of this arises.

DD
Old 5th December 2012
  #39
Gear Guru
 
Glenn Kuras's Avatar
 

For Dan,



Old 5th December 2012
  #40
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
I would more than happy to do similar tests right here for you. This diffusion over several threads and now a different forum is confusing to me. I can't distill what particular issue or issues are being referred to within all those other linked threads , but will gladly try to explain and test any particular one if raised directly here.

IF there is an issue, let's find it. In the meantime I suggest an innocent assumption.
This is a commercial product. A one man company. His living.

To be clear.

The 250mS issue is particular to this measurement here in this thread.
I have not seen anything like it before.
Here:

Did my first measurement! Need help to analyse it.

Finished treating room - graphs good enough?


EDIT:
And: FuzzmeasurePro3: Analyze Results ...
Old 5th December 2012
  #41
Lives for gear
Quote:
Originally Posted by DanDan View Post
I recommend driving the room fully
Like this?

Quote
1
Old 5th December 2012
  #42
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Old 6th December 2012
  #43
Gear Guru
 
DanDan's Avatar
FM is Fine

Sorry Jens but I can't identify anything in particular that I could test in those links. If you see some particular issue you would like tested, I will be happy to help you out.

If I may point out again, I use FM frequently and have never seen anything like the 250mS issue.
So it is not a behaviour or 'bug'. Unless one particular to that app on that computer. i.e. Corruption.

I suggest the best way to figure out that 250mS would be to ask that OP to test again. Let's use increasing sweep levels until the ETC gets within say 6dB of Full Scale, just in case level has something to do with it. I would be quite happy to guide this process, there are settings options in Audio Mid Setup which could restrict the action to 16 Bit.
I am quite familiar with the appearance of bit truncation and such from earlier 16 Bit times, but I haven't seen any of it since the major change to 32 Bit float.

Using FM, as I do, driving the graph levels up, I am not finding issues, problems, or bugs.
So Glenn, I think we are safe to continue recommending FM.

It might be worth noting that Apple and Java are not vbfs so it seems possible that REW might cease working on Macs again.
So perhaps FM should be cut some slack or given some respect for being one of the few good acoustic applications for Mac.

We all see plenty of posts of flawed testing and presentation.
When the test is redone correctly or the data presented better, both REW and FM seem all good to me.

EDIT, chasing around the matrix of linked threads, I see level anomalies on the different graphs. This is probably a mathematical thing.
But it certainly doesn't encourage full level sampling, which currently would appear as +10 or so on the FR graph for instance.
I will ask Chris about it.

DD
Old 7th December 2012
  #44
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
Sorry Jens but I can't identify anything in particular that I could test in those links. If you see some particular issue you would like tested, I will be happy to help you out.

If I may point out again, I use FM frequently and have never seen anything like the 250mS issue.
So it is not a behaviour or 'bug'. Unless one particular to that app on that computer. i.e. Corruption.

I suggest the best way to figure out that 250mS would be to ask that OP to test again. Let's use increasing sweep levels until the ETC gets within say 6dB of Full Scale, just in case level has something to do with it. I would be quite happy to guide this process, there are settings options in Audio Mid Setup which could restrict the action to 16 Bit.
I am quite familiar with the appearance of bit truncation and such from earlier 16 Bit times, but I haven't seen any of it since the major change to 32 Bit float.

Using FM, as I do, I am not finding issues, problems, or bugs.
So Glenn, I think we are safe to continue recommending FM.

It might be worth noting that Apple and Java are not vbfs so it seems possible that REW might cease working on Macs again.
So perhaps FM should be cut some slack or given some respect for being one of the few good acoustic applications for Mac.

We all see plenty of posts of flawed testing and presentation.
When the test is redone correctly or the data presented better, both REW and FM seem all good to me.

DD
I decided to do a quick search (in GS) for IR files exported from Fuzz to see if these latest files I´ve seen, are just unlucky exceptions. This is what I found (the IRs discussed recently not included):


Peak level (in dBFS) and bit depth of IR files:

-23,92, 16 bit
-21,34, 16 bit
-23,84, 16 bit
-25,36, 16 bit
-20,15, 16 bit
-20,42, 16 bit
-18,73, 16 bit
-16,39, 16 bit
-13,01, 16 bit
-23,13, 16 bit
-27,16, 16 bit
-27,19, 16 bit
-29,64, 16 bit
-28,18, 16 bit
-21,01, 24 bit
-15,65, 16 bit
0,00, 16 bit (possibly clipping or normalized)
-18,00, 16 bit (very noisy in lows)


So either your saying that the average FuzzMeasure user is terrible att setting a proper gain structure to get a good recording level, or there´s something not right here. Even if including the IR with a 0,00 dBFS peak level (possibly clipping); the average peak level of all these files end up at -20,73 dBFS. Levels this low might not be a problem if recording and exporting in 24 bit, but if 16 bits (as the majority of users are recording/exporting in), the files gets truncated (or something else is wrong) and this is the result:

Before posting your measurement results-left-arc-.gif
Before posting your measurement results-left-no-arc.gif
Before posting your measurement results-richs-exported-left-right.gif
Before posting your measurement results-right-arc-.gif
Before posting your measurement results-right-no-arc.gif
Before posting your measurement results-1.gif
Before posting your measurement results-2.gif
Before posting your measurement results-3.gif
Before posting your measurement results-4.gif
Before posting your measurement results-5.gif
Before posting your measurement results-6.gif
Before posting your measurement results-7.gif
Before posting your measurement results-8.gif
Before posting your measurement results-9.gif
Before posting your measurement results-10.gif

As you can see, these IR files are “cut off”, low level information is lost, either due to truncation or some other issue, since the decay rate suddenly increases drastically: Not a normal behavior in any room.

If inspecting the IR using lower resolution and a smaller vertical (dB) scale, an inexperienced user might be fooled to believe that the modal decay is good even if it´s not.


So what´s the fuzz about here DanDan? I´m already in contact with Chris (that also thinks that there might be something not quite right) and he´s going to have a look at this when time permits.
Old 7th December 2012
  #45
Lives for gear
It looks sort of like the background is barrel-shaped. Is that what you mean by truncated? At first I was wondering if I was seeing an artifact in the loudness vector but I guess it could be a time artifact. I'd think the latter less likely unless maybe a calculation artifact? I admit I'm grasping at straws.

Edit: or is this a gaussian decay-shaped artifact? (I'm struggling to remember if that's what an inverted "S" shaped decay is called) Is this showing up because the level during the test was low and the artifact magnified?
Old 7th December 2012
  #46
Gear Guru
 
DanDan's Avatar
Export

When files are Exported by FM, there is a choice of 16, 24, 32 Bit.
Perhaps dither is not included, which might explain the patterns normally seen in truncation. Even at 16 Bit, we are dealing with a 96dB range if I remember correctly. Including these 'conservative' -20dB levels, lets call it 76dB.
That is plenty IMO, particularly in a typical prosumer scenario with ambient at 40dB and sweep levels of 90-94dB.
At the risk of further repetition, REW, and the Studio Six (Terrasonde), software are 16 Bit, and again, John PM has shown in detail how this is plenty for the job at hand.

On a quick scan through half a dozen REW ETC's from clients and GS I have advised, I found none higher than -20dBFS. This did not hinder the work at hand in any sense.

So the ETC levels commonly seen in actual files from real people, using both REW and FM appear to be coming in the same, around -20dB.

I have noticed and been curious about the lowish levels over time, but have not found them to be at all a hinderance in using the software to get the job done.
There are other operational issues which are definitely a hinderance. But considering that REW doesn't function at all with many interfaces, well....

I have pointed to the Magnitude level difference between the FR graph and ETC and so on elsewhere and have informed Chris. There is either an explanation, probably mathematical, or else there will be changes in future versions. ETC and Third octave Decay times, were not in the early versions.
Chris responds. He's a good guy, his product works well. As does REW.

Houston, I think we do not have a problem.


DD
Old 7th December 2012
  #47
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
When files are Exported by FM, there is a choice of 16, 24, 32 Bit.
Perhaps dither is not included, which might explain the patterns normally seen in truncation. Even at 16 Bit, we are dealing with a 96dB range if I remember correctly. Including these 'conservative' -20dB levels, lets call it 76dB.
That is plenty IMO, particularly in a typical prosumer scenario with ambient at 40dB and sweep levels of 90-94dB.
At the risk of further repetition, REW, and the Studio Six (Terrasonde), software are 16 Bit, and again, John PM has shown in detail how this is plenty for the job at hand.

On a quick scan through half a dozen REW ETC's from clients and GS I have advised, I found none higher than -20dBFS. This did not hinder the work at hand in any sense.

So the ETC levels commonly seen in actual files from real people, using both REW and FM appear to be coming in the same, around -20dB.

I have noticed and been curious about the lowish levels over time, but have not found them to be at all a hinderance in using the software to get the job done.
There are other operational issues which are definitely a hinderance. But considering that REW doesn't function at all with many interfaces, well....

I have pointed to the Magnitude level difference between the FR graph and ETC and so on elsewhere and have informed Chris. There is either an explanation, probably mathematical, or else there will be changes in future versions. ETC and Third octave Decay times, were not in the early versions.
Chris responds. He's a good guy, his product works well. As does REW.

Houston, I think we do not have a problem.


DD

Did you see the waterfalls above?


EDIT:

Again:
Quote:
Originally Posted by Jens Eklund View Post
Levels this low might not be a problem if recording and exporting in 24 bit, but if 16 bits (as the majority of users are recording/exporting in), the files gets truncated (or something else is wrong) ...
Old 7th December 2012
  #48
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
I made a simple test:


Using the second set of IRs from member Fetachin that are 24 bit files (unlike the 16 bits previously posted) found here:
Finished treating room - graphs good enough?

I changed the peak level of the IR to -20 dBFS and exported as a 16 bit and a 24 bit file. I also made one file with a 0,1 dBFS peak level and exported this as a 16 bit file.

This is what you get when looking at the waterfall plots:

Before posting your measurement results-ir2-left-sub-20dbfs-16bit.gif
Before posting your measurement results-ir2-left-sub-20dbfs-24bit.gif
Before posting your measurement results-ir2-left-sub-0-1dbfs-16bit.gif


This presumably proves a number of things:

A -20 dBFS peak level of a 16 bit IR file is not going to show the decay properly due to truncation of low level signals.

REW is not a “16 bit app” since if it was; the 24 bit file with a -20 dBFS peak level would look similar to the 16 bit file with the same peak level.
Old 7th December 2012
  #49
Gear Guru
 
Glenn Kuras's Avatar
 

Quote:
So Glenn, I think we are safe to continue recommending FM.
I never stopped. I do think this is a good thread to make sure.
Old 7th December 2012
  #50
Gear Guru
 
DanDan's Avatar
16 Bit

REW Manual
Quote:
16 bit resolution is the same as used on CDs, and is the resolution REW supports.
We both have found -20dB Levels to be quite the norm in measurements posted from FM and REW equally.
So if there was an issue here it would not be confined to FM.

I tried but failed to replicate the Waterfall behaviour seen in Jens software, ARTA I presume.

Here are two Waterfalls, one was recorded at 16 Bit, the other 24 Bit.
Name:  16 Bit.jpg
Views: 508
Size:  82.5 KBName:  24 Bit.jpg
Views: 470
Size:  81.7 KB

I really do not see a problem here.

DD
Old 7th December 2012
  #51
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
REW Manual
"16 bit resolution is the same as used on CDs, and is the resolution REW supports."

We both have found -20dB Levels to be quite the norm in measurements posted from FM and REW equally.
So if there was an issue here it would not be confined to FM.

I tried but failed to replicate the Waterfall behaviour seen in Jens software, ARTA I presume.

Here are two Waterfalls, one was recorded at 16 Bit, the other 24 Bit.


I really do not see a problem here.

DD
DanDan:


Just because an application records in 16 bits does not make it a 16 bit app. If you don´t believe me; why not have another look at the waterfalls posted (from REW b t w …) demonstrating the benefit of 24 bit resolution compared to 16 bit.

I´m really struggling to understand why you choose to ignore data presented here and in other threads. Why the eager to jump to the conclusion that there is nothing wrong here? Either there is something not right in FM or the majority of users (that I´ve seen so far) are struggling to use it properly and if so, this is also an issue.

I don´t understand the reluctance to try and locate the possible issue here since this would presumably also be beneficial to you as an FM user, right?
Old 7th December 2012
  #52
Gear Guru
 
DanDan's Avatar
Problem

The internal resolution of audio software is usually much greater than the capture or output resolution, e.g. 48 Bit fixed point, and more recently 32 or even 64 floating.

But REW's capture is 16 Bit. That is a fact, according to REW author JohnPM.

Whether 16 Bit is enough or not has been debated and John has shown it to be plenty for the job at hand. I have had similar discussion with Studio Six/Terrasonde.

I just created two Waterfalls, trying to replicate Jens skilled, perhaps creative, work in ARTA.

Both were taken with identical levels, as loud as I could do. Both still hit only
-20dB Magnitude on the ETC graph.
Both Waterfalls look very similar to me. I see no sign of truncation.

Once again, the -20 Levels have been commonly seen in REW as well as FM
I won't bore by presenting them here, you all have them at hand.

So yet again, I see absolutely no reason or logic in suggesting that there is something wrong with FM in particular.


DD
Old 7th December 2012
  #53
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Old 7th December 2012
  #54
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
The internal resolution of audio software is usually much greater than the capture or output resolution, e.g. 48 Bit fixed point, and more recently 32 or even 64 floating.

But REW's capture is 16 Bit. That is a fact, according to REW author JohnPM.

Whether 16 Bit is enough or not has been debated and John has shown it to be plenty for the job at hand. I have had similar discussion with Studio Six/Terrasonde.

I just created two Waterfalls, trying to replicate Jens skilled, perhaps creative, work in ARTA.

Both were taken with identical levels, as loud as I could do. Both still hit only
-20dB Magnitude on the ETC graph.
Both Waterfalls look very similar to me. I see no sign of truncation.

Once again, the -20 Levels have been commonly seen in REW as well as FM
I won't bore by presenting them here, you all have them at hand.

So yet again, I see absolutely no reason or logic in suggesting that there is something wrong with FM in particular.


DD
No one has ever argued against REW recording in 16 bits.

If you record at close to clipping level (indicated by Fuzz) and yet the actual peak level of the IR file is -20 dBFS, this would support my theories.

I don´t use REW for making measurements but I don´t recall having seen these issues with IRs posted from REW and I´ve seen a lot more of those on the forum that FM IRs.
Old 7th December 2012
  #55
Gear Guru
 
DanDan's Avatar
Correct

That suspicion is correct, I have stated such repeatedly. The input meter on FM is called Level Meter. By default it shows 94dB at Full Scale.
Unless calibrated to the real world using an SLM or Calibrator, this seems a bit arbitrary. It is. However 94dB happens to be a decent level to strive for and is a commonly used Calibration Level.

As ever, I would prefer my recording level meter to be dBFS.

I do believe the use of Magnitude on the FR Y axis discourages higher levels.
I have always mistakenly kept it kissing 0dB.

Even with this misassumption, my ETC's typically hit -10dB.

REW has a dBFS meter, but also an automatic level setting helper.
The results seems to be the same as FM.
i.e. the great unwashed achieve conservative -20dB levels, while brilliant sound engineers achieve highs of -10dB.....

I might try some more tests, but I suggest some help from Chris and JohnPM might be better. I am convinced the minutae under debate here are matters of design choice and mathematics. None of which render any of the software incompetent in any way. I wonder is it even possible to hit 0dBFS on an ETC in any of this software.

EDIT There are much bigger issues with posted measurements IMO. e.g. The persistent use of Two Speakers, the short Waterfalls ending in a sheer cliff. These are common to REW and FM.
I could post any number of examples of GS posted REW IR's, ETC -30 to -20. Again, I see no difference between it and FM and again see no reason to single out FM.
And again, I don't see any problem with any of this. JohnPM and Chris are talented dedicated people. As a user I am delighted with their inventions.

DD
Old 7th December 2012
  #56
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
That suspicion is correct, I have stated such repeatedly. The input meter on FM is called Level Meter. By default it shows 94dB at Full Scale.
Unless calibrated to the real world using an SLM or Calibrator, this seems a bit arbitrary. It is. However 94dB happens to be a decent level to strive for and is a commonly used Calibration Level.

As ever, I would prefer my recording level meter to be dBFS.

I do believe the use of Magnitude on the FR Y axis discourages higher levels.
I have always mistakenly kept it kissing 0dB.

Even with this misassumption, my ETC's typically hit -10dB.

REW has a dBFS meter, but also an automatic level setting helper.
The results seems to be the same as FM.
i.e. the great unwashed achieve conservative -20dB levels, while brilliant sound engineers achieve highs of -10dB.....

I might try some more tests, but I suggest some help from Chris and JohnPM might be better. I am convinced the minutae under debate here are matters of design choice and mathematics. None of which render any of the software incompetent in any way. I wonder is it even possible to hit 0dBFS on an ETC in any of this software.

EDIT I can post any number of examples of GS posted REW IR's, ETC -30 to -20.
DD
The level of the peak in the ETC is irrelevant since it´s normally normalized to 0 dB anyway. I´m naturally referring to the peak level of the actual IR file.

I´m using ARTA for recording, and if the level indicator in ARTA indicates -1,0 dBFS peak; the peak level of the IR file is -1,0 dBFS, as it should be.
Old 7th December 2012
  #57
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
EDIT There are much bigger issues with posted measurements IMO. e.g. The persistent use of Two Speakers, the short Waterfalls ending in a sheer cliff. These are common to REW and FM.
I could post any number of examples of GS posted REW IR's, ETC -30 to -20. Again, I see no difference between it and FM and again see no reason to single out FM.
And again, I don't see any problem with any of this. JohnPM and Chris are talented dedicated people. As a user I am delighted with their inventions.

DD
So you now agree that there might be an issue with FM (and perhaps also REW, who knows ... No one has singled out anything here. I´ll check the REW level meter once time permits)?
Old 7th December 2012
  #58
Gear Guru
 
DanDan's Avatar
Levels

Quote:
The level of the peak in the ETC is irrelevant since it´s normally normalized to 0 dB anyway. I´m naturally referring to the peak level of the actual IR file.
Normalisation in FM and REW is an option, not a norm, as is blatantly obvious from the posted levels of -30 to -20dB

Quote:
I´m using ARTA for recording, and if the level indicator in ARTA indicates -1,0 dBFS peak; the peak level of the IR file is -1,0 dBFS, as it should be.
Good for you. Seems to be useful in finding esoteric problems.
But many of us are using REW and FM, quite happily. For Acoustic purposes.

Again the Y Axis in FM reads Magnitude. From the Manual.

Quote:
Magnitude Response
The Magnitude Response graph displays the result of a Fourier transform contained within the analysis window for the selected impulse response(s). The window’s shape (i.e. Hamming, Half- Hamming, etc.) is taken into account for this calculation.
REW does read dBFS on the ETC graph, typically the seemingly low levels seen everywhere.
i.e. -30 to -20dB

Out of curiousity, I may try some tests to see if this ETC level can be driven up much higher, although, I am quite content with my usual -10 Levels or for that matter the more usual -20 levels. I have no intention of exporting all these files transformed to WAV's etc, as it may not be a 1:1 relationship. And it simply doesn't matter in my experience and in the tests I have done here.

EDIT, no I do not have an operational issue with either REW or FM. Both work fine in my hands.

I have sent John and Chris a link to this thread to see if they will throw some light on these detailed internal matters of their software.

DD
Old 7th December 2012
  #59
Gear Guru
 
DanDan's Avatar
Recall

Quote:
No one has singled out anything here.
Simply not true.

Quote:
a potential problem is being exposed with Fuzz Measure:

http://www.gearslutz.com/board/bass-...ml#post8509843

Does anyone using FuzzMeasure (I don´t since I´m on a PC) know what the problem is?

DD
Old 7th December 2012
  #60
Lives for gear
 
Jens Eklund's Avatar
 

Thread Starter
Quote:
Originally Posted by DanDan View Post
Normalisation in FM and REW is an option, not a norm, as is blatantly obvious from the posted levels of -30 to -20dB



Good for you. Seems to be useful in finding esoteric problems.
But many of us are using REW and FM, quite happily. For Acoustic purposes.

Again the Y Axis in FM reads Magnitude. From the Manual.



REW does read dBFS on the ETC graph, typically the seemingly low levels seen everywhere.
i.e. -30 to -20dB

Out of curiousity, I may try some tests to see if this ETC level can be driven up much higher, although, I am quite content with my usual -10 Levels or for that matter the more usual -20 levels. I have no intention of exporting all these files transformed to WAV's etc, as it may not be a 1:1 relationship. And it simply doesn't matter in my experience and in the tests I have done here.

EDIT, no I do not have an operational issue with either REW or FM. Both work fine in my hands.

I have sent John and Chris a link to this thread to see if they will throw some light on these detailed internal matters of their software.

DD
Why would you want to look at an un-normalized ETC?

Well, anyway; whatever levels are reported within the applications for various plots is not really relevant. It´s the low level of the recorded/exported IR file that is causing the problems seen.
New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+ 
 
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
Dannibal / Post Production forum!
9
LaChunkyStudio / Post Production forum!
1
Dissdnt_penguin / So many guitars, so little time!
19
GeorgeSkoufalos / Post Production forum!
2
evangelista / Post Production forum!
0

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

SEO by vBSEO ©2011, Crawlability, Inc.