Cool Dude Ted

REW software

Recommended Posts

Thanks.

Due to my inexperience, I hadn't realised that something as simple as a door opening could make a significant difference.

  • Like 1

Share this post


Link to post
Share on other sites

Not easy is it Stuart?

It’s great software but there are choices to be made at every stage which will affect the results.

Just a quick thought on your recent posts, and I’d welcome Martin’s input. You appear to be trying to create filters using a single measurement with no smoothing of the measured response. There is a problem with this as any filter will be particular to that precise listening position. As you can see, moving the measuring mic a bit will give a different result.

A couple of solutions come to mind. The first is a bit time consuming and involves taking an extra number of measurements around the area of the listening position and average them, then average them with the measurement taken at the listening position. A simpler method would be to alter the smoothing of the measured plot in the EQ page. By smoothing more (say going from 1/48 octave smoothing to 1/6) you will see that the filters are not so extreme and fewer in number. You could also change the setting that adjusts the adherence to the target curve. If set to 1dB the software will try to correct every variation in the measured curve. By increasing this setting you might get a more useful, in listening terms, set of less extreme filters.

These will give a broader brush approach that might well be enough to achieve a good listening experience which isn’t just suitable for the precise position of the measuring mic.

The important thing is to try various options and then listen carefully to find out which works best in practice.

Hope this helps rather than confuses.

  • Upvote 1

Share this post


Link to post
Share on other sites
19 minutes ago, Camverton said:

Not easy is it Stuart?

It’s great software but there are choices to be made at every stage which will affect the results.

Just a quick thought on your recent posts, and I’d welcome Martin’s input. You appear to be trying to create filters using a single measurement with no smoothing of the measured response. There is a problem with this as any filter will be particular to that precise listening position. As you can see, moving the measuring mic a bit will give a different result.

A couple of solutions come to mind. The first is a bit time consuming and involves taking an extra number of measurements around the area of the listening position and average them, then average them with the measurement taken at the listening position. A simpler method would be to alter the smoothing of the measured plot in the EQ page. By smoothing more (say going from 1/48 octave smoothing to 1/6) you will see that the filters are not so extreme and fewer in number. You could also change the setting that adjusts the adherence to the target curve. If set to 1dB the software will try to correct every variation in the measured curve. By increasing this setting you might get a more useful, in listening terms, set of less extreme filters.

These will give a broader brush approach that might well be enough to achieve a good listening experience which isn’t just suitable for the precise position of the measuring mic.

The important thing is to try various options and then listen carefully to find out which works best in practice.

Hope this helps rather than confuses.

Hi Malcolm

Thank you for taking the time to reply.

No, I am certainly not finding this easy and this comes from someone who has a modicum of IT knowledge.

Your first comment is spot on. Choices all over the place and every one seems to impact on the final results. All I want to do is to remove a couple of bass humps below 100Hz which shake the whole house, even at "normal" volume levels.

If I could just pick up on some of your comments and if you have time to reply that would be great;

1) I am reading your comments about smoothing as a recommendation to employ some smoothing rather than having none. Is this fair?

2) Can you point me in the direction of the setting that adjusts the adherence to the target curve.

3) I am still struggling with the concept of target curve/level. Should I ever set this manually, or let the SW calculate it for me by clicking "set target level" under "Target settings" tab?

4) Am I correct in now concluding that the volume level which I set my amplifier to (to take measurements) is/should be irrelevenat to the final outcome/EQ filters?

5) would purchasing a calibrated mic make the whole process easier - i.e does loading a preconfigured calibration file take away any of the other "complexities" or would it mean that (say) the preferences screens can be ignored?

Thank you

Stuart

Share this post


Link to post
Share on other sites

 You didn't ask me but I'll respond to a couple of these questions:

4 hours ago, Cool Dude Ted said:

2) Can you point me in the direction of the setting that adjusts the adherence to the target curve.

In the Filter Tasks section on the righthand side of the EQ window it is the setting labelled Flatness Target. Set this to 3 dB and it will try to get all points over the frequency range you choose to be within 3 dB of the target level/curve. 

4 hours ago, Cool Dude Ted said:

3) I am still struggling with the concept of target curve/level. Should I ever set this manually, or let the SW calculate it for me by clicking "set target level" under "Target settings" tab?

I don't think I have ever used a value calculated by the software xD. It's not a bad idea for you to use this as a guide to begin with. Your graphs above only went up to 500 Hz: is this as high as you've measured to just what you chose to show? To help set the target level I'd always measure up to 20 kHz.

Here's an old example of a single point correction I used, shown using Var smoothing that hides details but shows the main trends. This has a deliberate slight downward slope from left to right but note that the the level at higher frequencies gives a guide to pretty much what you want the low frequency target to be. 

770549502_Inroomexample.jpg.1256b789e6130a38f51f0d122e43ed51.jpg

4 hours ago, Cool Dude Ted said:

4) Am I correct in now concluding that the volume level which I set my amplifier to (to take measurements) is/should be irrelevenat to the final outcome/EQ filters?

Yes. If it wasn't you'd need a different correction every time you changed the volume when listening to music too which I'm pleased to say you don't :)

4 hours ago, Cool Dude Ted said:

5) would purchasing a calibrated mic make the whole process easier - i.e does loading a preconfigured calibration file take away any of the other "complexities" or would it mean that (say) the preferences screens can be ignored?

A calibrated microphone would make your measurements more accurate and give you greater confidence that the curve you see is what is actually going on. It wouldn't make as of the process easier though.

  • Like 1
  • Upvote 1

Share this post


Link to post
Share on other sites
5 hours ago, Camverton said:

Just a quick thought on your recent posts, and I’d welcome Martin’s input. You appear to be trying to create filters using a single measurement with no smoothing of the measured response. There is a problem with this as any filter will be particular to that precise listening position. As you can see, moving the measuring mic a bit will give a different result.

It's my fault that just a single point is being used right now as I suggested it to keep things simple to begin with. Whilst someone is basically bound to go through multiple loops of measuring and trying things I tend to think this saves time. Purely for bass corrections I also think it works pretty well to be honest and it's all I ever did before Dirac. I do agree that some from of multi-point average is better though, and the moving microphone technique @tuga suggested is one way to do this although not one I've tried.

  • Like 1

Share this post


Link to post
Share on other sites

@Cool Dude Ted I remember the other reason I suggested a single point measurement is because you wanted to be able to check whether what you saw in REW would be replicated when transferring values to Moode. The way to do this is to make a measurement, generate a simple set of filters (you could even limit yourself to one), enter the values in Moode and then make a second measurement to see if the change is what you expected. For this to be a fair comparison you want to not move the microphone or change anything else in between the two measurements.

  • Upvote 1

Share this post


Link to post
Share on other sites
5 hours ago, Cool Dude Ted said:

Hi Malcolm

Thank you for taking the time to reply.

No, I am certainly not finding this easy and this comes from someone who has a modicum of IT knowledge.

Your first comment is spot on. Choices all over the place and every one seems to impact on the final results. All I want to do is to remove a couple of bass humps below 100Hz which shake the whole house, even at "normal" volume levels.

If I could just pick up on some of your comments and if you have time to reply that would be great;

1) I am reading your comments about smoothing as a recommendation to employ some smoothing rather than having none. Is this fair?

2) Can you point me in the direction of the setting that adjusts the adherence to the target curve.

3) I am still struggling with the concept of target curve/level. Should I ever set this manually, or let the SW calculate it for me by clicking "set target level" under "Target settings" tab?

4) Am I correct in now concluding that the volume level which I set my amplifier to (to take measurements) is/should be irrelevenat to the final outcome/EQ filters?

5) would purchasing a calibrated mic make the whole process easier - i.e does loading a preconfigured calibration file take away any of the other "complexities" or would it mean that (say) the preferences screens can be ignored?

Thank you

Stuart

1) I’ve found that it is best not to overdo it when it comes to EQ. If you EQ precisely for one listening position then any correction will be wrong for another position, or in my case, when I find myself reclining more as the evening goes on. Using smoothing on the measurement plot will make the EQ less precise but still take care of broad peaks that are clearly audible. As Martin points out, taking readings at multiple mic positions is best but you then have to decide how to balance the readings around the main position to the main listening position. If starting out it is probably easier to measure at the listening position and use smoothing to get usable filters.

2) See Martin’s reply.

3) First off, before taking measurements I set the output to 75dB by measuring the sound coming from the speaker (s) using the “check for level” full range in preferences. Consistency is more important than absolute value here. This is one advantage of using a calibrated mic such as a umik but you can also use an SPL meter or app on a smart phone. 

Clicking on the “set target level” will automatically compensate if your measurements are at too high or low a value. I usually click on it to see what it gives but also adjust to give me the filters I am looking for. How do you know what you are looking for? A lot of measuring and listening to the results helps here! I tend to look at the frequency response and decide what sort of filters I want and then use EQ to calculate the values for me. (This bit is for another day, perhaps, but having got the frequency response right and measured again to confirm I then look at spectrogram and waterfall plots etc to look for any oddities in delay, although by and large reducing a bump in the frequency response will reduce a delay - it can get quite involved so keep to the basics first!)

4) See Martins reply but it is best if your measurements are at 75dB to get a good result. If measuring at too low a level you get nearer the noise floor - usually 35 - 45dB in many rooms and if you measure at too high a level you could damage your equipment or hearing.

5) I started off playing with a radioshack SPL meter then moved on to an iPad app before buying a umik mic. I certainly wouldn’t want to do this sort of thing now without a calibrated mic. I guess it depends on whether the outlay is worth it to you and how much use might get from it. Be warned though, that having peered down the rabbit hole you might find all this stuff fascinating and find yourself in ever expanding burrows! It is an interesting hobby in its own right and can give results that greatly improve ones enjoyment of the dear old music.

  • Like 1
  • Upvote 1

Share this post


Link to post
Share on other sites

@MartinC

@Camverton

thank you both for your detailed replies. I really do appreciate your help.

I’ll have a look in greater depth in the morning.

Share this post


Link to post
Share on other sites
27 minutes ago, Camverton said:

4) See Martins reply but it is best if your measurements are at 75dB to get a good result. If measuring at too low a level you get nearer the noise floor - usually 35 - 45dB in many rooms and if you measure at too high a level you could damage your equipment or hearing.

As the OP doesn't have a calibrated microphone his 75 dB could be anything really of course. So just using a volume that doesn't sound really loud or really quiet is probably the best advice.

(I think 75 dB is a rather low threshold to be concerned about system or hearing damage for what it's worth, but it's certainly a safe value.)

Share this post


Link to post
Share on other sites
22 minutes ago, MartinC said:

As the OP doesn't have a calibrated microphone his 75 dB could be anything really of course. So just using a volume that doesn't sound really loud or really quiet is probably the best advice.

(I think 75 dB is a rather low threshold to be concerned about system or hearing damage for what it's worth, but it's certainly a safe value.)

Indeed, I think the problem is that without checking the volume of the measurement sweeps it would be possible to use way above 75dB by accident. Without a meter most people wouldn’t have any idea how loud 75dB actually is. A simple phone or tablet app will suffice for getting the level near enough. 

As ever, when setting volumes for measurements start at a very low volume and slowly increase until 75dB, as an example, is reached.

Share this post


Link to post
Share on other sites

@MartinC

@Camverton

this must be the most pronounced case of 3 steps forward, 2 steps back that I’ve ever encountered anywhere at any any time in my life.

What I think I have learned taking inspiration from the above posts;

1 I now have a basic SPL meter on my IPad and now have a vague idea of “what does 75DB sound like”. If I now understand this correctly, I’ll be able to set the volume on my analogue amp so that the average sound (pressure?) level coming out of the speakers is around 75Db. If I’m honest, it so happens that this is the sort of volume level that I normally listen to music at.

2 my analogue mic is not producing a high enough input signal to achieve results which could even remotely be regarded as accurate by even the most optimistic of users. REW says that analogue mics should be recording at -12 to -24 Db, -18 Db a good average - yet I am miles away from this figure.

What I have yet to learn/discover or simply get my head round;

3 whether I have any prospect whatsoever of reaching the stage where my analogue mic is sufficient for what I want to do - namely remove a couple of bass humps sub 100Hz. I had assumed that because my laptop had a mic in Jack that somewhere inside would be a (hardware) mic pre amp. Now I am not so sure. I accept it makes no sense for me to proceed down the route of a mic pre amp (to buy one), when a USB mic can be had for around £100

4 as an aside, but connected issue - even before starting out on this journey, I knew I have never been able to grasp the “Whole Db Thing”. I have always been confused by the seemingly interchangeability (not sure if this is the right word) between plus and minus decibel figures. This has been truly  exacerbated today by the SPL meter on my IPad. The graph, plotted frequency on the X axis and Db on the Y axis, shows sound level as a minus figure on the axis and a higher volume level is represented by a decrease in the negative value. I.E -40Db is louder than -60Db. I still cannot get my head round why these figures are minus. Against this, the figure for overall SPL - showing as DB-A - is a positive figure. If anyone knows of an idiots guide/video to help me sort out this confusion, then I would be delighted to hear. Of course, this is not specific to REW.

5 Under preferences in REW there is an instruction for loop back and measuring the sound card. Is this only relevant for an analogue mic? Sorry if obvious, but I just can’t get my head around this. If a USB mic circumvents this requirement - an even better reason for buying one.
 

Thanks. Once again, I quite understand if you don’t have the time to reply to what seems like a never ending saga.

Edited by Cool Dude Ted
Spelling error

Share this post


Link to post
Share on other sites
Super Dealer

I use an analogue mic, but with a mic preamp, a focusrite ‘scarlet’, a usb mic is more straightforward.

You just don’t need to worry about calibration, simply adjust volume until the test tones are similar to your listening levels, 75dB is ‘quiet’, 85dB is becoming loud and 95dB is extremely loud!

REW tells you if the input is to quiet, ‘green’ is go.

This will give you an idea,

Keith

Edited by PuritéAudio
  • Upvote 1

Share this post


Link to post
Share on other sites
2 minutes ago, PuritéAudio said:

I use an analogue mic, but with a mic preamp, a focusrite ‘scarlet’, a usb mic is more straightforward.

You just don’t need to worry about calibration, simply adjust volume until the test tones are similar to your listening levels, 75dB is ‘quiet’, 85dB is becoming loud and 95dB is extremely loud!

REW tells you if the input is to quiet, ‘green’ is go.

Keith

Thanks Keith.

if I jump straight to the “Measure” screen and hit “check levels”, then REW reports “level is low” (in Blue) and I simply can’t find a way of increasing the input level to a point where REW reports it as being OK (which I think it shows in green with a minus Db figure).

Stuart

Share this post


Link to post
Share on other sites
Super Dealer

Are you hearing the test tone through the loudspeakers?

Turn up the volume until REW goes green would be my advice, you have set mic as your input and your dac as output?

Keith

Share this post


Link to post
Share on other sites
23 minutes ago, Cool Dude Ted said:

this must be the most pronounced case of 3 steps forward, 2 steps back that I’ve ever encountered anywhere at any any time in my life.

I'll have more time for detailed replies over the weekend but I'll just say you shouldn't feel you're doing badly at all. Pretty much everyone goes through a learning phase as you are. As I hope my posts earlier in this thread suggested I don't think anyone should expect to get a somehow perfect result first time. I certainly didn't.

I'll also just say that so far I'm trying to work out if you actually have the issue with two bass peaks that you think you might. 

  • Upvote 1

Share this post


Link to post
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.


  • Recently Browsing   0 members

    No registered users viewing this page.