• 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!

Jriver (DSP)Surround Set Up, JRSS And Another Upmixing Discussion

Trdat

Addicted to Fun and Learning
Forum Donor
Joined
Sep 6, 2019
Messages
967
Likes
396
Location
Yerevan "Sydney Born"
I have 5.2 set up with my rear surrounds up the front as side surrounds, actually in Toole's book you can read that this option is one of the more preferable layouts for surround sound and personally I hate sound emanating from the back.

I know discussions around up-mixing will still rage and I have posted a couple fascinating threads myself, but I thought ill ask with Jriver, is there any techniques that can be used in the DSP tab to emulate an up-mixer on top of the JRSS that is recommended for up-mixing. Can delay be used for the surrounds? If so by how much...?

Overall, the surround format definitely adds a greater immersion but it only works for live music and recorded music that has a touch of reverb as it enhances the already live feeling. For studio music doesn't really work, unless there exists a touch of reverb. Of course don't quote me on the terminology of reverb as I am guessing there is some recording or production process that works well with the 5.2 set up, I am just naming it reverb in the recording.

I have a unique set up with my mains a horn/CD down to 900hz and a 15 inch woofer and wide directivity speakers for the centre and side surrounds. The side surrounds and centre speaker match and I feel it bridges the gap in the lack of reflections I have from the point source speakers.

I thought this thread can serve for anyone to provide input on how they manage surrounds within digital crossovers and can't use a AVR for up-mixing the one downside of digital crossovers and DSP for setting up surrounds. Also, if anyone has any techniques used to up-mix in Jriver or has any experience with JRSS...?
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
I have 5.2 set up with my rear surrounds up the front as side surrounds, actually in Toole's book you can read that this option is one of the more preferable layouts for surround sound and personally I hate sound emanating from the back.

I know discussions around up-mixing will still rage and I have posted a couple fascinating threads myself, but I thought ill ask with Jriver, is there any techniques that can be used in the DSP tab to emulate an up-mixer on top of the JRSS that is recommended for up-mixing. Can delay be used for the surrounds? If so by how much...?

Overall, the surround format definitely adds a greater immersion but it only works for live music and recorded music that has a touch of reverb as it enhances the already live feeling. For studio music doesn't really work, unless there exists a touch of reverb. Of course don't quote me on the terminology of reverb as I am guessing there is some recording or production process that works well with the 5.2 set up, I am just naming it reverb in the recording.

I have a unique set up with my mains a horn/CD down to 900hz and a 15 inch woofer and wide directivity speakers for the centre and side surrounds. The side surrounds and centre speaker match and I feel it bridges the gap in the lack of reflections I have from the point source speakers.

I thought this thread can serve for anyone to provide input on how they manage surrounds within digital crossovers and can't use a AVR for up-mixing the one downside of digital crossovers and DSP for setting up surrounds. Also, if anyone has any techniques used to up-mix in Jriver or has any experience with JRSS...?

Last time I checked JRiver forum, it was mentioned that JRSS works similarly to Dolby pro-logic II and DTS Neural-X -- one of the developers or more knowledgeable members said it, but cannot remember exactly who. The technical details and differences between JRSS and other upmixers are opaque to me...

But I agree with your assessment in that for dry studio type recordings it may not "work" as well as expected.

Additionally, there is an issue with the bass response when you combine JRSS with "pseudo surround" detection mode -- i.e. bass EQ needs to be re-corrected due to some odd summing irregularity.

Since "reverb" enhancement is what we're primarily aiming for here, I started disabling bass-management (manual method) of the surrounds and apply a linear phase HPF/low cut at around 140Hz (in addition to some some manual "post" phase linearization) to avoid any issue altogether. So I then get much less interference in the bass this way when all channels are swept simultaneously with a coherent signal (for testing).

JRiver DSP studio PEQ section
1652193226023.png

*I have no center hence its absence


The red magnitude curve below is simply the result when a mono L+R log sweep is combined with the matrixed surround signals (20 ms delayed).

FDW 15 cycles
1652193959629.png


One could either adjust the volume level of the surrounds or equalize them further (for this DSP preset mode) if you still find them obtrusive. I do not add any more delays than what is already provided.


The more similar the phase profile of the bass section between multiple channels, the less "trouble" or irregularity in the summed response.

However, phase doesn't particularly matter much the higher up in frequency since the surrounds are delayed by 20ms anyway -- actually, more or less the further away one sits from the center position.

Envelope Time-Curve
1652194048821.png



Wavelet: 1/6 res, 40 dB scale, normalized
1652194061219.png



I do not use JRiver's "Room Correction" module (I find it limited and somewhat unwieldy), and much prefer to manually combine the PEQ section(s) with their convolution file loader -- a little additional FIR phase manipulation is used in order to improve the summed response between all channels and minimizing cancellation/dips.
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
Last time I checked JRiver forum, it was mentioned that JRSS works similarly to Dolby pro-logic II and DTS Neural-X -- one of the developers or more knowledgeable members said it, but cannot remember exactly who. The technical details and differences between JRSS and other upmixers are opaque to me...


Actually, some of the mixing algorithms/methods may be very similar or same to what has been discussed here:

 
OP
Trdat

Trdat

Addicted to Fun and Learning
Forum Donor
Joined
Sep 6, 2019
Messages
967
Likes
396
Location
Yerevan "Sydney Born"
Last time I checked JRiver forum, it was mentioned that JRSS works similarly to Dolby pro-logic II and DTS Neural-X -- one of the developers or more knowledgeable members said it, but cannot remember exactly who. The technical details and differences between JRSS and other upmixers are opaque to me...

Dolby Pro logic 2 has some squealy high pitched extraction of one part of the bandwidth which I absolutely hate, JRSS seems like its playing a larger bandwidth or minimally less than the mains with a delay and in no way is it obtrusive I actually enjoy it.
But I agree with your assessment in that for dry studio type recordings it may not "work" as well as expected.
Good to see that I am on the right track.
Additionally, there is an issue with the bass response when you combine JRSS with "pseudo surround" detection mode -- i.e. bass EQ needs to be re-corrected due to some odd summing irregularity.

Since "reverb" enhancement is what we're primarily aiming for here, I started disabling bass-management (manual method) of the surrounds and apply a linear phase HPF/low cut at around 140Hz (in addition to some some manual "post" phase linearization) to avoid any issue altogether. So I then get much less interference in the bass this way when all channels are swept simultaneously with a coherent signal (for testing).

Well, I use Audiolense anyway for my DSP and my mains are digital crossovers, do you think I should still be aware of the anomaly?
JRiver DSP studio PEQ section

*I have no center hence its absence


The red magnitude curve below is simply the result when a mono L+R log sweep is combined with the matrixed surround signals (20 ms delayed).
I always appreciate your graphs and measurement Ernest, I will analyze them. Learn a lot from them.
One could either adjust the volume level of the surrounds or equalize them further (for this DSP preset mode) if you still find them obtrusive. I do not add any more delays than what is already provided.
I tried 20ms delay to the side surrounds, if the recording has quite a bit of reverb it does sound sensational its like you are recreating the live event in your listening room. But 20sec delay or any extra delay for that matter with say techno doesn't work.
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
Well, I use Audiolense anyway for my DSP and my mains are digital crossovers, do you think I should still be aware of the anomaly?

This is easy to check... do a L+R sweep with the pseudo-surround checkbox turned off and then on. You would need to check that the surrounds are really playing as it takes time for JRiver to activate the surrounds. Maybe play some pink noise and check the analyzer tab just to make sure they have started playing before doing the sweep -- if not press stop to reset the loaded DSP settings. If the mains and surrounds are not all centered exactly, the upper frequencies will look weird which isn't really important here. You mainly want to see the difference in the bass response with pseudo-surround mode engaged or disabled.

I tried 20ms delay to the side surrounds, if the recording has quite a bit of reverb it does sound sensational its like you are recreating the live event in your listening room. But 20sec delay or any extra delay for that matter with say techno doesn't work.

If one were to upmix to 7.1c instead, JRSS's pseudo-surround mode will apply an additional 10 ms delay to the rear surround channels so that's 30 ms in total (perhaps to simulate rear wall reflections) -- the side surrounds remain at 20 ms. Disable the pseudo-surround mode, and those additional delays are effectively removed from the surrounds. The small amount of delay added is enough to keep the surround effect well outside the (direct sound) integration time of the ears.

1652282603360.png


Bottom IR envelope graph is supposed to be an 'ideally' treated room.
 
Last edited:

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
Additionally, there is an issue with the bass response when you combine JRSS with "pseudo surround" detection mode -- i.e. bass EQ needs to be re-corrected due to some odd summing irregularity.

Since "reverb" enhancement is what we're primarily aiming for here, I started disabling bass-management (manual method) of the surrounds and apply a linear phase HPF/low cut at around 140Hz (in addition to some some manual "post" phase linearization) to avoid any issue altogether. So I then get much less interference in the bass this way when all channels are swept simultaneously with a coherent signal (for testing).

Just an update to the couch measurements from before in post #2 in this thread:

I understand a little clearer now why an "anomaly" occurs such that the ff. white/gray traces below could happen:

1661010836147.png

*Light gray and white traces

If the BM crossover EQ to the surrounds is not disabled during stereo LR "pseudo-surround" upmixing, non-linearity in both summed electrical (DSP section) and acoustic responses occurs due to timing and level differences -- though, it is much more emphasized in the electrical side of things.

We know that the bass managed (re-directed) sub response of the surrounds are delayed by 20ms in the above case by the JRSS upmixer causing some amount of phase & magnitude difference induced cancellation. The effect is particularly pronounced in the subwoofer bass region.

1661010403212.png


Right above 100 Hz or so we barely see any effect and so SPL remains constant. Do note, however, that we also wouldn't want to remove the 20ms delay in the surrounds as it would cause a wholly different effect -- for example, there will be an acoustic SPL increase of about 2 dB in the summed response.

So, ideally, the delayed surrounds (green trace) should be playing on their own effectively in full-range or high-passed mode without any active bass re-distribution or bass management routing to the subwoofer.

I started disabling bass-management (manual method) of the surrounds and apply a linear phase HPF/low cut at around 140Hz (in addition to some some manual "post" phase linearization) to avoid any issue altogether.

This step is actually unnecessary and can be skipped unless the surrounds' extension is very deep, and where the equalized low bass magnitude level is shelved way up high possibly causing significant unwanted acoustic interaction.

1661013767467.png



1661014808738.png



The upmixed surround signal is automatically delayed from t=0, however, the front LCR will still start at exactly t=0.

1661013784371.png



1661013789393.png



Some more measurements to come...
 
Last edited:

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
Continuing from a previous post in another thread: https://www.audiosciencereview.com/...-to-calibrate-any-speakers.36283/post-1282099


The ff. are simple swept measurements at my desk MLP:

1661026764875.png


The steady state, full-range frequency responses were adjusted objectively (via measurements) and subjectively -- through temporary channel re-mapped DSP presets so that isolated solo listening (with and without bass management) could be cycled through all the channels repeatedly back-and-forth.


The center channel (white trace) looks unnaturally elevated from about 7 kHz to 20 kHz. Please, ignore that rise in the steady state response as what you see does not directly correlate to what is actually perceived. The speaker in question is a single driver Fostex 6301 that is situated off-axis above my monitor screen, and its HF decay and damping charactistic (besides narrow directivity design) makes it so that what's heard in the HF region is largely muted/padded. The speaker sounds very neutral side-by-side my Neumann KH120s. There is definitely a difference in timber and tonality, wherein the 6301 sounds "dark" or heavily mid-range-focused -- but this is true in comparison to just about every other studio monitor speaker with a dedicated HF driver -- the ones I've heard, at least.

Some of the damping and decay profile (only non-anechoic and in-room MLP measurements) can be observed and compared in the spectral decay graphs:

1661020885407.png 1661020891308.png 1661020896162.png 1661020900781.png
*hopefully some insight why this is so can be gained from looking at these graphs
**also note the rise time setting

Now, while the rear surrounds look attenuated in the first "baseline" graph (dotted curves), those speakers are actually positioned at a further distance away, and so give out more late reflected energy back into the MLP from all the room boundaries -- i.e. they sound much louder in level than what you would expect if only looking at the simple steady state curves. Yep, the sound levels in-between channels fine and sounds relatively balanced in level.


1661021440156.png



1661026780896.png



Much of any boost in the measured summed (stereo upmixed) steady state response comes from the contribution of the synthesized center channel mix. This is because its phase response is coherent with the LR mains. All the surround speakers should be sufficiently de-correlated (from the original LR stereo mix) and mainly only affect the overall perceived loudness and envelopment level.

EDIT: previously made an error by omitting bass management to the center mix channel
1661055194991.png


Since I didn’t feel the bass was lacking before when the center was not bass managed, this should not be so much a requirement but rather a recommendation. Ultimately, though, the only way to find out which sounds better for most material is through conducting a lot of more listening.

1661055275397.png


*The LF level would be lower if no bass management were applied to the center channel. On the other hand, if non-coherent low bass content from any of the pseudo-surround channels were inadvertently added to the subwoofer, there's bound to be a clash in their summed response.



If the default stereo pseudo-surround upmix comes out sounding excessive to you (depending on recording material), it's relatively easy to reduce or completely mute out the offending channels.

1661021454706.png

*easy!

Eh, I know... I'm coming off really redundant by now. :p
 

Attachments

  • 1661021450559.png
    1661021450559.png
    89.9 KB · Views: 60
Last edited:
Top Bottom