Can someone with more experience of reading REW generated spectrograms help me decipher what is going on here? Thank you!
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Help with spectrogram
- Thread starter caught gesture
- Start date
Can someone with more experience of reading REW generated spectrograms help me decipher what is going on here? Thank you!
ernestcarl
Major Contributor
Just another way of looking at the IR... REW's online help:
When viewing sweep measurements in Fourier or Wavelet mode the time starts before the peak of the impulse so that the onset of the response can be seen. The areas where the response is decaying more slowly show up as streaks along the time axis. The dashed line is the Peak energy time trace which shows the peak level in the plot at each frequency. This can highlight variations in peak energy arrival versus frequency - an ideal peak energy time trace would be a straight line with the same time value for all frequencies.
Some of the modal/reflection effects show as delayed energy. The theoretical ideal is uniformly flat with the same amplitude level and starting time, which doesn't exist in the real normal rooms.
Some explanations
Gearspace.com - View Single Post - My small production room
Post 13652424 -Forum for professional and amateur recording engineers to share techniques and advice.
gearspace.com
Lesson learned using Wavelets & Group Delay for sub alignment (REW) - Gearspace.com
A quick thread to share some things I've learned whilst using the Group Delay (GD) and Wavelet tools in REW for aligning a subwoofer with main speakers. Here's a shot of what my speakers look like with the subwoofer engaged: Easy to notice the big spike i
gearspace.com
- Joined
- Feb 23, 2016
- Messages
- 20,771
- Likes
- 37,635
Try looking at the same measurement as a waterfall plot. Showing the same thing really. Does that make any more sense?
Basically you are seeing the start of the signal at each frequency at the bottom and how many miliseconds up the chart with the level at those times past the initial signal.
Basically you are seeing the start of the signal at each frequency at the bottom and how many miliseconds up the chart with the level at those times past the initial signal.
If I take a look at this graph I would say that frequencies <25Hz are ringing significantly (>1 sec) compared to >25Hz
Also there is some ringing around 30-40 and 50-70Hz (around 400-600ms which I would not consider bad at these frequencies in a non-treated room)
Just my 2 cents, I might be wrong
Also there is some ringing around 30-40 and 50-70Hz (around 400-600ms which I would not consider bad at these frequencies in a non-treated room)
Just my 2 cents, I might be wrong
OP
- Thread Starter
- #6
I guess my confusion comes from the relationship between the Peak energy time trace and the decay time. The peak at 68Hz and the corresponding decay time, for example. Why in some cases there is a delay. I’ll delve deeper and educate myself. Thanks, everyone.
OP
- Thread Starter
- #7
Thanks for your input. In layman’s terms it is easier to understand specific to my room. I’m new to REW and am trying to dial in my subs better. I’ve already been able to make big improvements by adjusting the cutoff frequency and the phase on the subs.
ernestcarl
Major Contributor
Some of it could be noise from fast 128k length sweep. I prefer using a slower sweep measurement for viewing a waterfall or spectrogram graph.
OP
- Thread Starter
- #9
Useful tip. Thanks!
OP
- Thread Starter
- #10
I used the slower sweep, 512 instead of 256 used in the first test. The volume was different as I downloaded the latest version of REW and recalibrated the settings. The microphone was in the same position, but this time I opened the lounge window as well as a door into another room leading off from the lounge area. I’m finding it fascinating trying different combinations of doors and windows being open or closed, basotec panel behind the microphone or not, etc. The room is not treated except for a skyline diffuser I made a few years ago on the front wall between the speakers. I do have a lot of books covering the side walls, with floor to ceiling bookshelves plus large rug on the floor and a chunky two seater sofa. On the Voxengo plugin in Audirvāna I have eqd a hpf at 30Hz based on what I am seeing from these spectrograms. Am I heading in the right direction with my reading of these charts and making changes via eq?
dasdoing
Major Contributor
the sweep doesn't realy reject the noise for waterfall/spectrogram.
the only thing that works is loudness.
what i do is go into the appearance setting and bring the scale-range down until the modes stop to melt into each other.
also i highly recomend using "normalize to peak". thats where you realy see what is going on, as it will flat out the FR
the only thing that works is loudness.
what i do is go into the appearance setting and bring the scale-range down until the modes stop to melt into each other.
also i highly recomend using "normalize to peak". thats where you realy see what is going on, as it will flat out the FR
dasdoing
Major Contributor
is your room like 490 X 330cm?
OP
- Thread Starter
- #13
Close, 3,30 X 5,30. There are some features that make the room dimensions confusing. Here is a quick sketch. The dimensions are not 100% accurate. There is a door on the left of the diagram that goes into another room. The room on the right is the kitchen which is approx 6,00m in length. The corridor to the top left leads to a further two other rooms.
The grey boxes indicate my speakers and subs (the larger being the subs). Also, the spectrogram with the changes in appearance as you suggested.
Last edited:
dasdoing
Major Contributor
see that black line? how it is delayed around the modes. I think a phase correction (software) can cure the worst here.
OP
- Thread Starter
- #15
The black line on the leading edge of the nodes? I have manual phase correction ability on the individual subs. Would that be something I could further make fine adjustments of to make improvements? I have already made some improvement by doing this. If a software approach via plugin or other software would be a better choice, do you have any recommendations?
edit: Would Dirac Live Room Correction Suite Stereo be something I should consider? My main speakers are running full range due to everything coming from an integrated amplifier. I could host it in Audirvāna.
Last edited:
dasdoing
Major Contributor
these wont do it for these...as theye act on the whole. You don't use any correction software with phase correction abilities?
dasdoing
Major Contributor
look at the "tail" at 50Hz for example. if phase corrected the ringing would be reduced by 100ms-ish
OP
- Thread Starter
- #18
No, I don’t use any correction software other than the Voxengo PrimeEQ as a plug-in within Audirvāna. I’m willing to buy some software.
dasdoing
Major Contributor
I only have experience with free software (http://drc-fir.sourceforge.net/)...it requires a lot of time invested.
Dirac sure does correct phase, too. there is a topic somewhere from @mitchco comparing diferent software
ernestcarl
Major Contributor
I agree with you mostly. But the length does still have a small effect that show up as an unwanted variability in the waterfall and spect -- at least with my own measurements. It would be better if one could turn off the source of the noise if coming from within the house like a machine running with a motor e.g. an electric fan or HVAC system.
Similar threads
