I think what you’re both commenting on is due to me having no preamp, running the Pi ar full output, and using Roon’s volume control.
1 Like
Are you though? It looks to me the volume is implemented in the DAC, after RAAT.
1 Like
Which is what both myself and @Suedkiez thought from the signal path.
Something is going on with the Pecan possibly?
Thanks all.The only volume attenuation is coming from Roon-Pi is outputting full volume.
I think my problem is the .cfg file. I need to make one OR just redo/combine the two REW output .wav files into a single stereo file. I didn’t because the new version of REW (unlike the older one used in the video) does not allow me to use the settings recommended in the video with a stereo output
Only using DSP volume when going direct from DAC out to power amp, it‘s normal to see higher attenuations in order to not blow speakers and/or ear drums.
Of course, the required attenuation for normal listening depends on the DAC’s nominal max output vs the amp‘s gain and speaker‘s efficiency, and seeing ˋonly´ -18dB in the OP‘s setup, it seems we‘re not having many reserves, but let‘s figure out stereo correction file generation, before we might have to get back to that later.
Roon‘s DSP volume control is executed in the end point itself, so that‘s expected behavior, actually.
I recently redid my corrections with current REW version 5.31.3, and here you’ll have to use Export impulse response as WAV, not Export all impulse responses as WAV.
This way you‘ll be presented with all the needed options as explained by the video.
Not sure about it when using mono files, but if exporting stereo files as explained above, no config file is needed.
@Jon_Belak, give the above a try to see if your volume problem is solved and get back to us.
I have a suspicion, though, that you‘re loosing too much gain by the way you‘re performing the correction.
But let‘s not get ahead of ourselves and wait for your response …
1 Like
But then, it’s not really Roon’s DSP volume. If I remember correctly, the setting is ‘device volume’, and the device implements it somehow.
The OP‘s screenshot shows DSP volume, and it‘s Roon‘s code executed in the end point (RPi in this case) as detailed by Danny himself here.
To confirm, we would need @Jon_Belak‘s screenshot of Roon’s device setup dialog for his PecanPi.
2 Likes
Makes sense, thanks for the explanation. I was under the impression that Roon stops after RAAT.
1 Like
Hello, sorry for the delay. @Marin_Weigel I exported the stereo files at 44.1 and 48k using the export impulse response as wav option and fed Roon that zipped file. Also each individual file. All have the reduced volume issue, and I agree that the most likely culprit is that the correction is losing too much gain. I posted in the comments section of OCA’s video about this linked to in my original post. So far I have tried his suggestion about IR Windows, which didn’t work, and am awaiting further feedback (I’ll update once I do).
No problem.
Can you share your complete REW mdat file via some upload service to have look at it?
@Marin_Weigel here’s a link to the archive .mdat with all the measurements (the ones with MP on the end are the final ones)
https://file.io/PwF9hz58nWOC
Please try this link:
https://github.com/BelakScientific/REWConvolution/blob/fb45ae7694349d839f9f06e3af835992117b8b80/Archive2.mdat
There are so many things way off, that I don’t know how help you out, other than pinging @OCA for help.
To highlight your loudness problem, I just generated the curve L Room Corrected in green by multiplying LeftRoom with L Inverted-MP.
… the result is ~60dB lower in volume with actually no meaningful correction worth talking about at all.
See what I’m talking about by comparing it with my graphs (note: exact same scale!).
1 Like
Super. This could be a problem with the analysis or the Omnimic I’m using. I switched to REW from the Omnimic software because the results I was getting didn’t make sense. I have pinged the OCA.
I hope these are faulty measurements (ie faulty mic, speakers crossed over with a sub which was off during the measurement or just micro size desktop speakers) because among other things, the bass rolls off at 200Hz which is before where most EQ is normally applied, hence no chance of applying descent filters to these responses.
Assuming there was no error and it is what it is and these are full range speakers, the attached convolution file is the best I could do.
It’s 48kHz, just browse to it in Muse convolution. Roon will up/downsample as needed. You don’t need to zip it. I guess It will require a minimum of 15dB clipping headroom. You should also deduct -0.34dB from the left speaker volume in Muse speaker set up. I would not expect them to sound great but it should at least sound balanced.
2 Likes
EDIT:
@OCA posted while I composed my message …
Please don’t get my comments the wrong way, I don’t mean to be condescending!
By looking at the speakers’ impulse responses, I’ve a hunch your problems start with basic room acoustics and suboptimal mic positioning, unless they’re broken somehow.
Then, it seems you omitted preliminary REW settings for the procedure to produce optimal results, like using 48kHz sampling rate and an individual calibration file that comes with the mic.
Finally, you may have misunderstood or overlooked some of the steps.
Thank you both, your willingness to help is much appreciated. Yes, I don’t know my ass from my elbow when it comes to this. A few things: the vertical mic placement recommended by OCA did not seem right to do since the calibration files provided for Omnimic are only for horizontal mic placement. I did use that file, though. Also, I think I did select the 48k rate. Mic was placed horizontally at ear level exactly between the speakers. I really don’t have any way to change my room acoustics significantly, unfortunately. It is not a dedicated listening room. The speakers are oddball full range open baffles (some might consider them broken by definition) and although they’re 15" they’re placed higher in the baffle so not tons of bass. Bass swarm sub array that I normally used was turned off…This is set up to run through a MiniDSP and my plan was to do this afterwards. I wanted to get the mains as good as possible and do subs after. OCA, I appreciate you trying to come up with a workable result but based on Marin’s comment I need to revisit it all first and see if I get the same result. If any of the info above gives you additional thoughts please let me know.
Ok open baffles explain the lack of bass response. And Marin’s suggestion is right on. Also measure 0-24000Hz next time.
Thanks, I will. If I’m still not able to get a good result I’ll probably buy or borrow a new microphone.