Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[mojave]

The Bass Management in Room Correction is not working properly (at least the way I am testing). I have Output Format set to 7.1 and No Mixing. Then I use the Bass Management in Room Correction for the Left and Right speaker. I have it set to 120 Hz and checked "Add removed low frequencies . . ." What is happening is that the bass is being rolled off at the same rate and frequency as the main speakers. This can be seen visually by using Voxengo Span. I use the Underlay button to show both the mains and the LFE track at the same time.

If I use JRSS and set the crossover to 120 Hz and use Room Correction's bass management, then it works okay. However, this way it seems like you are routing the bass twice:  once in Output Format and once in Room Correction. This can also be seen visually by where the Span plugin is located in the chain.

[mojave]

Regarding the bass management, I didn't check but are you making sure to set the distance settings after you apply the bass management? There has been some discussion at AVSForum about the problems with the the bass management in the Emotiva UMC-1. Here is what Roger Dressler had to say:

When I read Lonnie's explanation on their bass  (mis)management, I was almost convinced that they went that route intentionally. Now that we see the time delays are applied before bass management, as evidenced by the comb filter cancellations, we must conclude it is one of two possibilities: a) they designed it this way intentionally, but had no idea what they were doing, or b) it was not designed this way intentionally, by someone who had no idea what they were doing, and are now trying to "feature it" rather than fix it.

Funny, since it takes no more effort, and no more DSP resources to do it right, I wonder what (E)motivated them to do this?

Reading the discussion of Emotiva's handling of bass management might be of some help as you develop this further.

Edit:  I forgot to mention that Roger Dressler designed the bass management for Dolby. Why don't you offer him a free copy of MC and ask for his input on Bass Management. 

[Matt]

Regarding the bass management, I didn't check but are you making sure to set the distance settings after you apply the bass management?

It's wrong in build 65, but will be corrected next build.

I think you're right that bass management should come before anything else.

[Alex B]

Bass management should work with any channel configuration.

I think it wold be better to create the channels that contain direction information (e.g. stereo to 5 or 7 channels - or 5 channels to 7 channels) first and after that handle the bass redirection separately if set by the user.

In addition you should provide a volume mixer with two volume sliders for the bass channel. One for the DSP processed "filtered/redirected/managed" bass signal and another one for the unaltered 0.1 aka LFE signal if it exists in the source. The original LFE signal from a 4.1, 5.1 or 7.1 source should bypass the JRSS process and Bass management and be included only by this mixer. Otherwise the managed bass signal and the original low frequency effect signal can be in a good balance only by chance.

EDIT

Of course the final speaker "volume and distance" adjustments must be "final" - i.e. become after the content for each channel is created in a way or another.

EDIT 2

11. Changed: Cross-over used for LFE channel during JRSS mixing uses a 24db/octave slope. (previously used a lesser slope)

This is probably obvious and you may already have it correct, but at the crossover frequency the curves are supposed to cross at -6 dB. I stole a picture from another site (attached).

[mojave]

In addition you should provide a volume mixer with two volume sliders for the bass channel. One for the DSP processed "filtered/redirected/managed" bass signal and another one for the unaltered 0.1 aka LFE signal if it exists in the source. The original LFE signal from a 4.1, 5.1 or 7.1 source should bypass the JRSS process and Bass management and be included only by this mixer. Otherwise the managed bass signal and the original low frequency effect signal can be in a good balance only by chance.

I am currently using the Voxengo BMS plugin for Bass Management. It provides both an LFE and a Mix Gain. I think this is what you are describing. Here is how the manual explains them:

LFE Gain parameter adjusts the loudness of the selected LFE channel. Please note that this parameter does not affect the extracted low frequency content being added to the LFE channel.

Mix Gain parameter adjusts the loudness of the extracted low frequency content summed from all non-LFE channels before it is mixed to the selected LFE channel.

[Matt]

This can be seen visually by using Voxengo Span. I use the Underlay button to show both the mains and the LFE track at the same time.

How are you viewing more than just the L or R channels?  The 'Channel' button doesn't list all the channels, even though I'm playing 6 or 8 channels.  I'm using 1.9.1 as a VST plugin.

Thanks.

[mojave]

You can view 8 channels. Click on Routing (not the drop down arrow but the actual word) and you can set up the routing and names for each channel. I set each channel as its own group. Once this is done, you can then view each channel independently or underlay one channel with another.

[Alex B]

I think this is what you are describing...

Hmm... Perhaps, I would need to see the plugin before I can say for sure. However, I created a drawing that illustrates what I mean. It's attached.

[mojave]

That is exactly what it does . . . although without the JRSS Magic. 

[Matt]

You can view 8 channels. Click on Routing (not the drop down arrow but the actual word) and you can set up the routing and names for each channel. I set each channel as its own group. Once this is done, you can then view each channel independently or underlay one channel with another.

I was using the old version.  They changed the name of the DLL, so I had the new one on disk but the old one installed.

Thanks.

[Matt]

First, thanks for the help everyone.  It may take a few go-rounds to get this right, but we will with your help.

I think there are a few things we're solving:

1) Filtering low frequencies from speakers

I think we've got this right.  Build 66 uses a higher quality high pass filter.  Currently we're always using a 12 dB/octave slope.  If there's a good argument for adding other frequencies or dB/octave slopes, let us know.


2) Redirecting filtered bass

We still have questions for how to handle this:

Now we have a per speaker, "add removed frequencies to the subwoofer."

However, would it be more straightforward just to choose how to fill in the subwoofer?  For example:

=== If source already has LFE channel ===

• Add bass from front speakers to LFE channel at: [ cutoff frequency] [ level ]
• Apply cross-over: [ frequency ]
• Subclarity

=== If source has no LFE channel ===

• Create LFE channel from input: [ level ]
• Apply cross-over: [ frequency ]
• Subclarity

I'm not sure if there are any other "big picture" bass management things we should solve.  Please let us know.

[Matt]

And one more Voxengo SPAN question.

It's really helpful working on this, but the newer version shows our pink noise with more energy in the high frequencies.  The old version (v1.9) shows a perfectly flat curve, which is what I would expect.

Any ideas?

[mojave]

And one more Voxengo SPAN question.

It's really helpful working on this, but the newer version shows our pink noise with more energy in the high frequencies.  The old version (v1.9) shows a perfectly flat curve, which is what I would expect.

Any ideas?

I am actually using GlissEQ, because it includes parametric EQ. SPAN is a scaled back free version. I think you click the Edit button in the upper right and then change the slope.

[mojave]

I had a little time to listen to the changes last night with build 15.0.66. It really sounds good now and it looks like you have the high and low pass working properly. I stayed up later than I should listening to music using the new bass management.

I've thought about the implementation and have a few suggestions.

Output Format

• I would get rid of the crossover setting here. JRSS should just send full range signals to however many speakers it is upmixing to.
• Instead of "For Stereo sources only mix to 2.1" you should have the selection "Do not upmix stereo sources." Don't get rid of this option, just tweak it! I like it.

Room Correction

• I would use the Bass Management heading with a few options under it - Crossover, Bass Routing, Low Pass Filter.
• You aren't removing frequencies below the crossover, you are rolling them off at 12 dB/octave. People don't really need to know that. Therefore, I would just put "Crossover" and the frequencies like you currently have in Output Format. Regarding the high pass filter, THX mandates a 24db/Octave high pass, but all other implementations I have seen use a 12dB/octave filter like you are currently using. You can probably just leave it at 12dB/octave.
• You need more frequencies available.
• I don't think you need the "add removed low frequencies . . ." option. Why would someone remove frequencies, but not want them routed to the subwoofer? I think there is a better way as I will explain in the next point.
• Add a drop down box for how to handle the bass below the crossover. It could be called "Bass Routing." The options would be "Move bass to LFE channel" and "Copy bass to LFE channel and keep speaker fullrange." I guess you could add the option "Make bass disappear."  
• A Low Pass Filter option could have a drop down box with 12dB/octave, 18dB/octave, 24dB/octave, 48dB/octave. If you use Copy in the above option, you might be using the natural roll-off of your speakers and will need a little more flexibility in the Low Pass Filter. You do need to make sure the summing is correct with each one when using Move so the frequency response remains flat. Another reason for these options is so you can set the crossover higher, but keep the subwoofer within its limits by using a steeper low pass filter.
• Under Tools, you could add the option to "Monitor redirected bass." I find something like this helpful because I can listen to the redirected bass in the speaker. This just plays, in the selected speaker, the bass below the crossover. To listen to the same bass in the subwoofer, one can then have just that speaker set with a crossover and can listen to the subwoofer with Mono selected. This allows for a listening comparison of the bass below the crossover and helps with crossover setup. This feature isn't necessary, but it can be helpful.

[Matt]

Thanks for all the feedback.

We have a little different proposal:

In DSP Studio > Output Format, rename "Subwoofer crossover" to "Bass management".  It would be a new section that contains:

Crossover frequency
Subclarity (yes / no)
What to do if source already has LFE: no processing / crossover & subclarity only / add bass from other speakers below crossover & subclarity
Levels: original, JRSS simulated, redirected  (not exactly sure how to do this, but there needs to be a way to dial in relative volumes)

The option in 'Room Correction' to "Add removed low frequencies to the subwoofer" would be removed.  I don't think this is a per-speaker setting, and it doesn't need a frequency since the subwoofer crossover handles this.

Thoughts?

[Alex B]

I don't think this is a per-speaker setting, and it doesn't need a frequency since the subwoofer crossover handles this

Someone might want to let the big front speakers to reproduce the full range and filter & redirect bass only from the center and surrounds (in most AV receivers you configure the individual speakers to be "large" or "small").

EDIT

BTW, I am just heading to our summer cottage for the weekend. Here are some pictures of it: http://www.pix01.com/gallery/4AD73041-94E6-4810-80BE-EDB7F3274485/My_20Summer_20Cottage/. Matt, I think the lake and forest look quite similar to what you have in Minnesota.

I can participate this discussion next time on Tuesday. I wish you all a nice weekend.

[Matt]

Someone might want to let the big front speakers to reproduce the full range and filter & redirect bass only from the center and surrounds (in most AV receivers you configure the individual speakers to be "large" or "small").

The frequencies to filter would still be available per speaker.

The 'add bass from other speakers' could be expanded to include other arrangements (add from all, add from rears only, etc.)

This proposal works better technically because the subwoofer crossover (and subclarity) happen in the 'Output Format' DSP which is always first.  It would be best to accumulate redirected bass before the subwoofer crossover is applied.

(p.s. sorry if I mix the words "subwoofer" and "LFE" loosely -- technically I should say "LFE" but I think "subwoofer" is better understood by normal people)

[mojave]

I don't see how statements "The frequencies to filter would still be available per speaker" and "It would be best to accumulate redirected bass before the subwoofer crossover is applied" work together. It looks like you are making the option of having a low pass at one frequency and a high pass at another frequency. The frequency at which these occur should be the same. You shouldn't combine all channels into the LFE channel and then roll it off at one frequency while at the same time allowing the speakers to roll off at a different frequency. You can end up with gaps in frequency response and potential phase issues.

What should happen is that you have up to seven channels of redirected bass. You then combine this bass and add it to the LFE channel. If this is too difficult technically, then you should only allow one crossover frequency that is applied to all speakers.

[Matt]

What should happen is that you have up to seven channels of redirected bass. You then combine this bass and add it to the LFE channel. If this is too difficult technically, then you should only allow one crossover frequency that is applied to all speakers.

That's no problem.

But isn't it typical to want a low-pass filter on your subwoofer?  I'd prefer a high-quality software lowpass to the hardware lowpass built into the subwoofer.  I'm fuzzy with how that should interact with redirected bass.

[JimH]

BTW, I am just heading to our summer cottage for the weekend. Here are some pictures of it: http://www.pix01.com/gallery/4AD73041-94E6-4810-80BE-EDB7F3274485/My_20Summer_20Cottage/. Matt, I think the lake and forest look quite similar to what you have in Minnesota.

I thought the same thing when I first saw your pictures.  Beautiful.

I can participate this discussion next time on Tuesday. I wish you all a nice weekend.

And you also.

[mojave]

That's no problem.

But isn't it typical to want a low-pass filter on your subwoofer?  I'd prefer a high-quality software lowpass to the hardware lowpass built into the subwoofer.  I'm fuzzy with how that should interact with redirected bass.

The LFE channel contains content up to 120 Hz per its specification (and some at AVSForum have measured content higher than that). You don't want to truncate that by putting a low pass filter on the subwoofer. The lowpass built into a subwoofer isn't used by many anymore. That is usually used when matching the natural roll off of a speaker to the subwoofer when using a 2 channel system without bass management.

The way an active crossover usually works for a channel is that the signal is duplicated. A crossover frequency is then specified. The high pass filter uses this crossover frequency and applies a 12 dB/octave rolloff. This information is sent to the speaker. The other duplicated channel has a low pass filter applied (at the same crossover frequency as the high pass) with a 12-48 dB/octave roll off. This information is sent to LFE channel (or combined with the other redirected bass before being sent).

By the way, Matt, I just sent you an e-mail.

[mojave]

I have set Output Format to 7.1 and No Mixing and then set the crossovers in Room Correction. The bass being sent already has a low pass filter applied as can be seen using the Span plugin. I don't understand why you are adding an additional low pass filter in Output Format.  ?

[Matt]

I have set Output Format to 7.1 and No Mixing and then set the crossovers in Room Correction. The bass being sent already has a low pass filter applied as can be seen using the Span plugin. I don't understand why you are adding an additional low pass filter in Output Format.  ?

In that case, there would be no need for a low pass.  But remember the LFE channel can be filled any of the following ways:

1) With data from a 5.1 or 7.1 source (DVD, etc.)
2) By JRSS during 2.0 upmixing
3) With 'Room Correction' using bass redirection

To make it a little more complicated, #3 could be added to #1 or #2.

[mojave]

If you eliminate #2 above, then you can just use the bass management in Room Correction. #1 only contains material up to 120Hz and you don't want to low pass that info. You want to combine it with the redirected bass using #3.

For upmixing from 2.0 to 3.0, 3.1, 4.0, 4.1, 5.1 or 7.1, just have JRSS upmix to the number of channels and create a "silent" .1 channel. Then use the information in Room Correction > Bass Management to route the sound to the LFE channel. This allows people to have either 3.0 or 3.1, 4.0 or 4.1, 5.0 or 5.1, 7.0 or 7.1 all depending on if they are starting with a 2 channel source and what settings they are using in Bass Management.

[Matt]

If you eliminate #2 above, then you can just use the bass management in Room Correction. #1 only contains material up to 120Hz and you don't want to low pass that info. You want to combine it with the redirected bass using #3.

That could be optional, but I think it's an overly complicated default.  Besides that, I like a lower crossover and subclarity for music upmixing.  From some limited testing, doing both JRSS and bass redirection yields good results, so this may not be a big issue.

The coming build 67 will have some improvements (including your routing idea):
NEW: Bass management in DSP Studio > Output Format allows specifying a bass routing mode of remove, move, or copy.
Changed: Bass management uses a 24 dB/octave lowpass for creating redirected bass (previously used a 12 db/octave slope).
Changed: Tuned JRSS mixing level used when creating a LFE channel.

[mojave]

That could be optional, but I think it's an overly complicated default.  Besides that, I like a lower crossover and subclarity for music upmixing.  From some limited testing, doing both JRSS and bass redirection yields good results, so this may not be a big issue.

I think it is overly complicated and confusing to have crossover settings at two different locations when one can handle it fine. I understand wanting different settings for music, but couldn't you just create a different zone and use one for music and one for movies. 

What if JRSS is set with a 40 Hz crossover in Output Format and the L/R speakers are set with a 60 Hz crossover in Room Correction? You then have a 20 Hz gap in frequency response. If you keep it all in Room Correction, couldn't you just add the Subclarity option to the subwoofer channel where the greyed out Bass Management stuff is located.

Speaking of wanting a different setting for music, if you leave it in Output Format, now seems like a good time to have a JRSS crossover setting of zip, zero, zilch so that no LFE channel is created. Just create a silent .1 channel. Then I can always have Output Format set to 7.1 and JRSS so that music is 2.1 using the crossover settings in Room Correction and 5.1 content is upmixed to 7.1. When there is already a .1 channel, JRSS doesn't create one, right?

NEW: Bass management in DSP Studio > Output Format allows specifying a bass routing mode of remove, move, or copy

I suggested "disappear"! 

[Matt]

Speaking of wanting a different setting for music, if you leave it in Output Format, now seems like a good time to have a JRSS crossover setting of zip, zero, zilch so that no LFE channel is created.

This will also be in build 67.

[Matt]

BTW, I am just heading to our summer cottage for the weekend. Here are some pictures of it: http://www.pix01.com/gallery/4AD73041-94E6-4810-80BE-EDB7F3274485/My_20Summer_20Cottage/. Matt, I think the lake and forest look quite similar to what you have in Minnesota.

That's beautiful.  It looks very much like the northern Minnesota and the Boundary Waters.  It makes me want to paddle a canoe.

[retro]

Changed: Bass management uses a 24 dB/octave lowpass for creating redirected bass (previously used a 12 db/octave slope).

Any chance of making this user-selectable like mojave suggests, 12-48 db/octave...?

Im using a big Thiel-subwoofer and I can "hear" it too much sometimes....I think 48 db/octave would be perfect...

[flac.rules]

Might i be an idea to clean this up a little? AFAIK, what conten is in LFE, and so on isn't really relevant on a modern receiver iwith digital in, the combined input of all channels are added together (including LFE-channel), for all content below the crossover, an routed to the subwoofer. Content above crossover is sent to the respective speakers. Exactly what is sent to the LFE-channel, and crossoversettings is thus mostly relevant for analog out as far as I can see. Maybe this should be clearer in the settings?

[Matt]

Might i be an idea to clean this up a little? AFAIK, what conten is in LFE, and so on isn't really relevant on a modern receiver iwith digital in, the combined input of all channels are added together (including LFE-channel), for all content below the crossover, an routed to the subwoofer. Content above crossover is sent to the respective speakers. Exactly what is sent to the LFE-channel, and crossoversettings is thus mostly relevant for analog out as far as I can see. Maybe this should be clearer in the settings?

We're adding the audio processing found in high-end receivers to Media Center.  Here's a thread from when we started down this path:
http://yabb.jriver.com/interact/index.php?topic=55607

I think using a computer to do this work makes a lot of sense.  Computers offer a nice user interface and have a lot of processing muscle.  There's really no way a receiver can compete.

We realize this is not for everyone.  But it should be interesting to anyone looking for the highest possible audio quality when using a computer.

[Matt]

Any chance of making this user-selectable like mojave suggests, 12-48 db/octave...?

Im using a big Thiel-subwoofer and I can "hear" it too much sometimes....I think 48 db/octave would be perfect...

Are you hearing frequencies above the cross-over, or do you mean it's just too loud?

Switching to 48 db/octave would be no technical problem, but higher order filters require more processing power so we want to make sure we're getting audible gains for the extra processing work required.

Any other opinions?  Thanks.

[flac.rules]

We're adding the audio processing found in high-end receivers to Media Center.  Here's a thread from when we started down this path:
http://yabb.jriver.com/interact/index.php?topic=55607

I think using a computer to do this work makes a lot of sense.  Computers offer a nice user interface and have a lot of processing muscle.  There's really no way a receiver can compete.

We realize this is not for everyone.  But it should be interesting to anyone looking for the highest possible audio quality when using a computer.

Hmm, I think i am not being totally clear. I am not opposed to the functions being available.  In fact, I think its a good direction to let a processor (computer or otherwise) to handle this.

However, with digital out, the receiver will always(?) do the crossover-job, and route content to the sub. I am not even sure this function can be turned off in receivers, and these particular features are not that computanionally advanced.

My point was more of an interface-one, AFAIK, setting crossover, and what gets sent to LFE-channel has no practical relevance when you have a receiver and output the sound digital, so maybe it would be possible to somehow in the interface show that these options are only relevant for people running analogue out (or straight to a dac). I think it can be confusing for people to set the "right" options. Somehow conveying the information that these settings are not important for a "normal" hdmi to receiver setup would maybe help.

[retro]

Are you hearing frequencies above the cross-over, or do you mean it's just too loud?

Switching to 48 db/octave would be no technical problem, but higher order filters require more processing power so we want to make sure we're getting audible gains for the extra processing work required.

Any other opinions?  Thanks.

Yes, I'm hearing "voices" from the sub. I'm sure 36 or 48db/oct would make a big difference. Please make it selectable in 12-24-36-48db steps....

More crossover points would also be really useful in order to seamlessly blend the sub with the mains.

Thanks for a great piece of software...

[Matt]

Coming in build 74 and later:

Changed: By default JRSS and Bass Management will use 48 dB/octave cross-overs for the LFE / subwoofer.
NEW: Bass Management in Output Format DSP allows selection of cross-over slopes per channel (independently configurable slopes for speaker and redirected bass).

[Bernt]

Coming in build 74 and later:

Changed: By default JRSS and Bass Management will use 48 dB/octave cross-overs for the LFE / subwoofer.
NEW: Bass Management in Output Format DSP allows selection of cross-over slopes per channel (independently configurable slopes for speaker and redirected bass).

Wow, looks like itīs time to upgrade from 14. This will for sure be a significant improvement. 

[GreggP]

It's great to see the new bass management features. I've just downloaded and upgraded to v15 because of this.

Now I need a little help to figure out how to use this. I have a 5.1 system.

When playing stereo source, I still like it to play back stereo, BUT I'd like to use the bass management feature to play 2.1.

In 'Output Format', should I leave Channels to 'Source number of channels' or to 5.1 and 'No mixing (unused channels silent)'?

and then set the Bass Management features within the 'Room Correction' ??

I assume this will apply to some of my quad (4.0) recordings, so I will get 4.1. I also have some 5.0 recordings which I assume will play as 5.1 (which is what I'd like with bass management, since my speakers only go down to about 60 Hz).

If I can get this to work, I may stop using foobar.   

[mojave]

You have to set the source number of channels to 5.1 and "No Mixing (unused channels silent)."

There have also been some improvements since version 15.0.80 so when make sure you download the newest version as soon as it is posted in this forum. There is more discussion in the thread Bass Management Testing.