First time with REW measurements

[yelloguy]

Hi,

I did a very quick measure with a umik-1 last night. This is the first time I've done this sort of a thing so I am sure I am missing something. I need to do more reading and learning. But I do have some quick questions on what I am seeing (attached).

I did a SPL calibration at 75dB so I am assuming that the high output I am seeing before 150hz or so is undesirable? Since after 150 the output falls down to about 75dB. For clarification the two measurements are for B&W 683 floorstanding fronts with an SVS PB 1000 sub. Audyssey is set to L/R Bypass and sub has been boosted to +1.5 (from Audyssey recommended -6). The speaker crossover was set to 40hz. For the second line, I changed the xo to 80 and knocked down the sub to 0.

I am assuming the bass from the speakers is adding to the sub output and giving me these peaks and dips.

I like bass heavy music and I enjoy it with these settings but my room gets some resonance issues in that my shelves vibrate. I can clear out my shelves, heh, or I can bring down the dB on those lower frequencies.

Thanks

 

[OCA]

Welcome to the rabbit hole

The boost you see below 150Hz is probably because you left Dynamic EQ on while measuring your system (it still adds a lot of boost to bass at 75dB) or Audyssey didn't get the subwoofer response and correction correctly. Althoguh this will not have much effect on the bass response, instead of 1/6 smoothing try to apply frequency dependent windowing of ie 15 cycles to the responses to see better what needs to be corrected.

 

[yelloguy]

Thank you for your help. And you are spot on - Dynamic EQ is ON.

I need to read up on the FDW option and what it means. Or watch some YT videos. But in the meantime, I have attached a screenshot of left speaker with FDW windows applied below.

 

[OCA]

Seems ok to me. Needs two cutoff filters at the small peaks at 45 and 65Hz.

 

[yelloguy]

Oh wow! Thank you so much.

But I don't have DSP to introduce those filters. Or do you mean I can do that with Audyssey? I have seen one of your YT videos where you talked about using an Excel file to build a new curve based on REW measurements and then feeding it to Audyssey. I will be watching it again later to see if that allows me to do it. Do you have any other pointers?

Thanks very much!

Edit: I just realized that my MultiEQ phone app will let me edit the curve directly. I guess that is what you meant. I will try to play with it in a bit

 

[yelloguy]

I did more measurements with Dyn EQ off this time. I ws able to steadily bring the bass areas down by further reducing the sub compensation from +1.5 to 0 to -3. At which point the only hitch was the two peaks at 45 and 70. I decided to try to address those in the sub curve using the app. Just dragging the two points down to -2.7 or so did the trick.

Ignore the differences 1K and up. It was measured at Audyssey Reference which adds a boost there.

I listened to some bass heavy tracks after increasing the sub to +1.5 again and I got no resonance this time! And I got floor shaking bass. Thanks for your help.

 

[Chrispy]

I did more measurements with Dyn EQ off this time. I ws able to steadily bring the bass areas down by further reducing the sub compensation from +1.5 to 0 to -3. At which point the only hitch was the two peaks at 45 and 70. I decided to try to address those in the sub curve using the app. Just dragging the two points down to -2.7 or so did the trick.

Ignore the differences 1K and up. It was measured at Audyssey Reference which adds a boost there.

I listened to some bass heavy tracks after increasing the sub to +1.5 again and I got no resonance this time! And I got floor shaking bass. Thanks for your help.

How does Audyssey provide a boost with Reference mode? It's natural mode is to roll off highest frequencies a bit. In any case glad you're making progress, it takes some time....

 

[yelloguy]

Chris, the highest frequencies you are talking about are 10-15k and up. I was talking about 1k and up as shown in the attached graph

This is one of the first things I checked after I got my umik. I've listened to music in Pure Direct mode (2.0) on my B&W speakers and I absolutely love that sound. After calibration, I tried the Reference mode and it made them sound like... KEF :vomit:

Jokes aside, I liked the subwoofer part. (have I mentioned bass heavy anywhere in this thread before?) So I found that switching to LR Bypass mode gives me the best of both worlds. I had no way to verify what my ears were telling me. So I tried mapping these modes with REW. Sure enough, you can see the LRB line coincides with the OFF line and the REF line flattens the 1K and up areas. It makes the speakers sound very generic. I have always called LRBypass the "B&W mode"

 

[OCA]

I have seen one of your YT videos where you talked about using an Excel file to build a new curve based on REW measurements and then feeding it to Audyssey

REW received many updates and added many new features lately. The latest tutorial has scripts that automates everything and works from your browser. You can use Audyssey's own measurments or measure everything manually in REW:

 

[yelloguy]

No way! I was on the supreme calibration video and copy pasting the Excel values in ady file using the json editor. Somehow the save and move back to the MultiEQ app is telling me that the curve is invalid. So I am trying to compare the files as we speak. Let me check watch the new one!
Edit: I said X app which is incorrect. I am using the phone app which is called MultiEQ

 

[OCA]

No way! I was on the supreme calibration video and copy pasting the Excel values in ady file using the json editor. Somehow the save and move back to the X app is telling me that the curve is invalid. So I am trying to compare the files as we speak. Let me check watch the new one!

The new one can also use MultEQ-X measurements again with just a script.

 

[yelloguy]

lol dude! what kind of loser has json editor! I have it on my bookmarks. ROFLOL

 

[yelloguy]

I got the ady file generated with your new script. It imported back to the Multi EQ app correctly this time.

The video has a lot of details on using the Audyssey measurements in REW. I am skipping over those because I will be doing my own measurements using umik-1.

I am not entirely sure where I would connect my own measurements with the "perfect speaker" ady file I prepared. (I have not sent it to the receiver yet).

I am yet to finish watching the full video.

(also for your future videos, can you increase the font size or reduce the resolution on your computer - at least while making these videos. it is very hard to read anything in the video)

 

[OCA]

I got the ady file generated with your new script. It imported back to the Multi EQ app correctly this time.

The video has a lot of details on using the Audyssey measurements in REW. I am skipping over those because I will be doing my own measurements using umik-1.

I am not entirely sure where I would connect my own measurements with the "perfect speaker" ady file I prepared. (I have not sent it to the receiver yet).

I am yet to finish watching the full video.

(also for your future videos, can you increase the font size or reduce the resolution on your computer - at least while making these videos. it is very hard to read anything in the video)

You should send the perfect speaker ady to the receiver, select Audyssey reference, turn on Dynamic EQ and then measure all your speakers with the supplied sweeps (use the recently published "lossless" TrueHD Atmos sweeps for even better resolution). The filters generated, speaker distances, volumes and crossovers will then be added to the perfect speaker ady with XL and REW scripts. FInally, send this ady to the receiver and enjoy.

The video has 1440p resolution but it's shot in ultra wide format so it may be hard to read in a 16:9 monitor. If I use wide screen recording, only a limited number of windows can be displayed simultaneously and it becomes harder to understand

 

[yelloguy]

Mac cannot play ATMOS sweeps! So I don't have the option to use those. I'll be using REW sweeps.

I watched the video twice and took notes the second time. I've got the part with speaker volume level calculations with Excel. But the EQ part is something I am still working on understanding and writing down the steps.

A couple of questions - is there a write up for this? Even if there is a partial post I can read and follow that. But the 1 hour long video gets hard to follow especially when loaded this this much info. I tried searching YT comments but those are scattered as well.

You are showing a subwoofer sweep of up to 20kHz. Is that not going to damage the sub or the AVR? I thought I was not supposed to play anything over 200Hz on the sub?

So far I have-
1. Upload the perfect curve, set dyn eq=on
2. Do REW measurements (I want to do 5 positions for 5 speakers + the sub)
3. Cross Corr, Vector Avg graphs
4. Align SPL 1kHz 5 octaves (no subwoofer)
5. In Excel - start with center distance, pick up align offset #, Receiver number is generated, adjust offsets
6. Add subwoofer. Put center on top, Cross corr align. Send IR shift in mm to Excel

 

[yelloguy]

@OCA After various problems I was able to make some progress this morning. But I can't decide what to do with the extra rows in the Excel sheet. I can enter values for the FL C FR SLA and SRA but I don't have the four extra speakers you have. Should I leave those values as empty? Enter 0? Delete rows? Thoughts?

Also, my Applied IR shift REW values are the following

17.0
18.0
18.0
1363.0
-229.0

I am assuming that SRA is messed up since SLA is 18. I will have to do another set of measurements and start from scratch anyway. I am just doing this one as a learning exercise.

Thanks for your help!

 

[yelloguy]

Answering my own question above, the script understands the missing channel. I left them as blank in Excel and everything worked out. Getting ready to transfer the ady file to the AVR next

 

[yelloguy]

So the ady file came out as corrupt. I made a few changes to fix that - but the one that most likely made the difference was the removal of comment from the filter text files. I noticed in your video that in Windows, when you check the box for not including comments in filters, it was indeed not including the comment as visible in the preview. Whereas on my mac, REW was still including the comment in the text file. I tried various things to lose that comment but decided I was imagining things. So in the first go round, the ady file was generated with comments in the filter text files.

When my phone app called it corrupt, I tried manually editing the filter files and removed all the lines at the top starting with a star *. When I regenerated the ady file the app accepted it. I made a few other small changes but I am pretty sure they were not breaking it.

Anyway, I don't like the sound I get with these filters. I can sort of see the filters designed by this process are pretty wild in my setup. I must have done something wrong somewhere. So I will start again and give it one more shot. Hopefully the process should be quicker this time.

One more thing to help someone else - watching the video full screen on a 4K monitor made a huge difference. Maybe my eyes are old but being able to read the small text helps. I understand your reason for doing it this way since you need to show everything together. To be honest I had to do my REW settings on a 4K monitor for the same reason as well.

PS - It was a hugely informative exercise and I cannot thank @OCA enough for sharing this

 

[OCA]

I am sorry @yelloguy, I was on a short holiday and totally missed your questions. Let me try and answer them:

• You can change all white boxes in the XL file and should replace the first column with your set of speakers. Make sure you have the latest version of the support files as I have uploaded them a couple of more times after revisions since the video launch.
• If your SLA and SRA are fairly equal distance from the LP, this difference must be a system delay of some sort during measurement and probably you should repeat the measurements.
• If REW Mac version is saving comment lines when the option to remove them is ticked, we should inform John about this bug. If you are certain, I will raise it in the AVNirvana forum.
• You wouldn't believe how expensive it is to record and edit videos in 4K in this day and age. Most free versions are still capped at 720p!
• You are right about the necessity for a write up but I didn't have time and there wasn't enough interest anyway

 

[yelloguy]

Thanks for the reply!

I got the support files last week but I will check [later] to see if it is updated now. But anyway, that all worked out - and it works very well to be honest.

My surround speakers are equidistant from MLP. The one that was sticking out was actually correct - as the Audyssey measurements were also closer to that one. But I will do the whole process again this weekend to see if I get better results. This is where the written series of steps might come in handy. Watching a 1 hour video and skipping parts of it can easily make me miss some steps - and I have to skip around because some parts don't apply to me, and I have watched the video... about 5 times now. But anyway...

I wasn't looking for 4k video quality. I was just talking about the small font size on my 16 inch macbook pro in full screen. I had to watch it on the 27 inch 4K monitor full screen to be able to read the text.

Comments are definitely a bug but John is dealing with a death in the family so I would leave him alone for a bit.

Meanwhile after measuring for a few days I tried to listen to the system yesterday instead. After comparing various calibrations, I came back to my original one as the most natural. The one where I reduced the bass 2dB at 40 and 70 hz also seems a little constricted. I listened to a whole album from DJ Khaled last night and it was very enjoyable due to soothing musical bass and accurate speaker imaging. It is my 8 measurement Audyssey with mid-range compensation turned off. Sounds great for 2.1 stereo in Audyssey Flat or LR Bypass (which is another way of doing low frequency adjustments only). I will try to do more REW measurements on this calibration and make incremental changes like distances and trim levels to see if I can improve it. Or I will add an external amp to my system to see if I can do better. EQ is getting me worse results so it might not be for me.