HT Processor Bass Management impact on SINAD

[RichB]

As we have seen with ASR measurements, AVR/AVP's have a difficult time approaching the SINAD of relatively inexpensive desktop DACs and sometimes not even matching a dongle-DAC.

These devices are clearly more complex and shielding and other design issues are present on some but, there may be a theoretical limit to SINAD due the nature of the device. That is what is being discussed here.

The theoretical processors under discussion has the following assumptions:

- All non-subwoofer channels can accept a 0 DBFS signal from the source
- The non-subwoofer channels (for example 13) can be bass-managed
- The subwoofer implementation does not have any capability for additional analog gain (for higher that 0 DBFS) so the added bass from other channels cannot exceed the digital maximum.

In this example, does bass management require attenuation of main channels in order to provide headroom for the subwoofer channel?
If so, what is the required attenuation?
If for example, it is 10 to, say, 18 dB then wouldn't that limit the achievable SINAD?

- Rich

 

[dlinsley]

On a similar vein, I asked Amir if during his Dirac testing he planned on seeing how SINAD was affected. For example, if you set the right-side curtain to say 3-500Hz to limit correction to Schroeder, how is the 1kHz SINAD affected?

 

[amirm]

First order of business is to look at whether SINAD is dominated by distortion or noise. If it is distortion, then the headroom issue is not a problem until it get so bad as to exceed distortion.

A few times I have measured SINAD but then realized bass management was on and had to go back and repeat all the tests. In general, removing bass management did not impact SINAD since it is at 1 kHz so well above the crossover region.

 

[RichB]

Here is information about the LFE channel and Subwoofer:

https://www.dolby.com/uploadedFiles/Assets/US/Doc/Professional/38_LFE.pdf

LFE does not equal subwoofer Dolby Digital programs may include a bass-only LFE channel, but this channel does not correspond directly to a subwoofer output. It is possible for a program to contain an LFE channel, but a decoder may provide no subwoofer output because all of the bass information in the program, including the LFE channel, can be reproduced by the main speakers. The opposite is also true: it is possible for a program to not contain an LFE channel, yet a decoder may provide a subwoofer output because some or all of the main speakers are unable to reproduce the bass information in the program.

The difference between the LFE channel and the subwoofer output is that the LFE channel is used to carry additional bass information in the Dolby Digital program while the subwoofer output represents how some or all of the bass information will be reproduced.

The LFE channel carries additional bass information to supplement the bass information in the main channels. The signal in the LFE channel is calibrated during soundtrack production to be able to contribute 10 dB higher SPL than the same bass signal from any one of the screen (front) channels. Even if all three screen channels are active, enough bass could be delivered by the LFE channel alone to bring the theatre’s subwoofer into acoustic balance with the screen channels. This allows filmmakers to unburden the main channels by diverting the strongest bass to the separate LFE channel, as needed. Under the most demanding program conditions, where the bass is fully loading the left, center, and right channels, the LFE channel could increase the bass intensity by up to 6 dB.

The subwoofer output, on the other hand, is bass information from up to all six channels that has been selected to be reproduced by a subwoofer. The specific combination of bass information in the subwoofer output is determined by the bass management settings chosen for that particular playback system’s speakers. For example, in addition to the bass information from the LFE channel, the subwoofer output may include the bass information from the center and surround channels when those speakers are unable to adequately reproduce the bass frequencies. As can be seen from the above explanation, the terms LFE and subwoofer are not interchangeable, and the distinction between the two terms is very important. Care should be taken to avoid confusion by using these terms appropriately.

If there are 13 bass managed channels, additional headroom on the subwoofer channel is required.
If channel balance is to be maintained, all channels must be attenuated to have digital headroom in the subwoofer channel.

If this is the case, manufacturers could have complex rules based on channel settings to adjust attenuation for the subwoofer channel.
For example, there could be 10 to 18 dB of attenuation of all channels implemented consuming some of the SINAD.

There can be SINAD issues with AVP/AVR implementation unrelated to bass management but, it may be that an AVP/AVR may not be able to achieve the SINAD of a desktop DAC. That is the point I am exploring.

- Rich

 

[Dimifoot]

I believe that this has been explained by an Emotiva guy in the RMC measurements thread.

 

[RichB]

I believe that this has been explained by an Emotiva guy in the RMC measurements thread.

@amirm‘s measurement have exposed real design issues which have already had fixes from manufacturers.
I thought it may be worth a deeper dive into the need for the good design and headroom in these products.

- Rich

 

[Dimifoot]

@amirm‘s measurement have exposed real design issues which have already had fixes from manufacturers.

True.

Also true that the Low Frequency Effects -LFE channel (separate channel, not to be confused with bass management) runs 10db hotter than the rest, or to state it diffently, all channels are configured at -10dbs from LFE in AV products

 

[markus]

True.

Also true that the Low Frequency Effects -LFE channel (separate channel, not to be confused with bass management) runs 10db hotter than the rest, or to state it diffently, all channels are configured at -10dbs from LFE in AV products

That would be 10dB wasted. Headroom is hopefully managed more intelligently. Dirac Live already "wastes" 10dB.

 

[Dimifoot]

That would be 10dB wasted. Headroom is hopefully managed more intelligently. Dirac Live already "wastes" 10dB.

Thats how movies are encoded.
I have no idea how it's implemented in the AV architecture/chips, but if you don't built it like this, then the LFE channel will clip.

 

[markus]

Thats how movies are encoded.
I have no idea how it's implemented in the AV architecture/chips, but if you don't built it like this, then the LFE channel will clip.

Only the redirected LF needs to be reduced in level (you have to make up for the loss at the end of course, preferably in the analog domain). The more channels, the more it needs to be reduced. In the end this is not really tragic as the signal ends up in the sub where our hearing isn't that sensitive anyway.

 

[Dimifoot]

The LFE channel is a separate channel, with discrete information.
Has nothing to do with bass management.

"The low-frequency effects (LFE) channel is the name of an audio track specifically intended for deep, low-pitched sounds ranging from 3-120 Hz. This track is normally sent to a speaker that is specially designed for low-pitched sounds called the subwoofer. While LFE channels originated in Dolby Stereo 70 mm film prints, they became commonplace in the 1990s and 2000s in home theater systems used to reproduce film soundtracks for DVDs and Blu-ray discs.......
The LFE channel originated in Dolby Stereo 70 mm Six Track film prints, as a way of providing louder bass and sub-bass effects, without detracting from the quality of the standard audio channels. The LFE channel is conventionally played back 10 dB louder than the main channels...."

From Wiki.. you could also search Dolby for more info.

 

[markus]

The LFE channel is a separate channel, with discrete information.
Has nothing to do with bass management.

"The low-frequency effects (LFE) channel is the name of an audio track specifically intended for deep, low-pitched sounds ranging from 3-120 Hz. This track is normally sent to a speaker that is specially designed for low-pitched sounds called the subwoofer. While LFE channels originated in Dolby Stereo 70 mm film prints, they became commonplace in the 1990s and 2000s in home theater systems used to reproduce film soundtracks for DVDs and Blu-ray discs.......
The LFE channel originated in Dolby Stereo 70 mm Six Track film prints, as a way of providing louder bass and sub-bass effects, without detracting from the quality of the standard audio channels. The LFE channel is conventionally played back 10 dB louder than the main channels...."

From Wiki.. you could also search Dolby for more info.

I don't need an explanation what the LFE channel is. Not sure why you think I would. We were talking about how to manage gain in a digital system.

Attached is a diagram showing how bass management works (also shows where DL applies filters). Obviously you can not apply +10dB to a channel that might already carry signals up to 0dBFS. In the end you have to make up for that 10dB somewhere though. Best way to do this is in the analog domain.
The redirected LF signal from speakers set to small needs to be lowered by 10dB when mixed into the subwoofer channel so the level relationship between LFE and main channels is maintained. Depending on how many speakers are set to small the overall gain of the whole subwoofer channel needs to be lowered further so the summed signal doesn't clip. Again makeup gain has to be applied in the end.

If this isn't managed dynamically as described above you're making SNR worse, throwing away bits.

 

[RichB]

I don't need an explanation what the LFE channel is. Not sure why you think I would. We were talking about how to manage gain in a digital system.

Attached is a diagram showing how bass management works (also shows where DL applies filters). Obviously you can not apply +10dB to a channel that might already carry signals up to 0dBFS. In the end you have to make up for that 10dB somewhere though. Best way to do this is in the analog domain.
The redirected LF signal from speakers set to small needs to be lowered by 10dB when mixed into the subwoofer channel so the level relationship between LFE and main channels is maintained. Depending on how many speakers are set to small the overall gain of the whole subwoofer channel needs to be lowered further so the summed signal doesn't clip. Again makeup gain has to be applied in the end.

If this isn't managed dynamically as described above you're making SNR worse, throwing away bits.

Once the number of bass-managed speakers are in place there is nothing that can be done dynamically that preserves the signal since the source must be supported.

AVR/AVP bass management is digital. Dirac Bass Control is digital.
It is possible to provide the additional gain in the analog chain but this is all a black box in these processors.
I have not seen any schematics showing an additional gain stage.

I suspect that each manufacturer is making their own decisions.
Some may consider clipping not typical and either let it clip or soft-clip (reducing the steered bass). Some may simply attenuate all channels to preserve balance at the cost of some S/N or some mix of both.

There seems to be little or no exposure of bass handling implementation. I am unaware of any professional reviewers targeting a deep dive into bass management, although, enthusiasts may have.

- Rich

 

[markus]

Once the number of bass-managed speakers are in place there is nothing that can be done dynamically that preserves the signal since the source must be supported.

Doesn't compute. Best case, gain staging is dynamically managed based on user bass management settings. It could even be dynamically optimized based on channel count of input signal but I doubt any manufacturer went that far.

 

[RichB]

Doesn't compute. Best case, gain staging is dynamically managed based on user bass management settings. It could even be dynamically optimized based on channel count of input signal but I doubt any manufacturer went that far.

Depends on what you mean by dynamically. Once the channel count and bass management then the configuration is static.
It does not change with program material.
Agreed, they are not likely changing the configuration with the source channel count.

The point of the thread, is that attenuation is occurring that is not needed for a 2-channel desktop DAC, that don't have room correction or crossovers.

ASR has found real issues measuring AVR/AVP's so there is room for improvement.
Due the the nature of the beast, an HT processor will never best the SINAD of a well designed 2-channel DAC.
An HT processor needs excellent SINAD to provide headroom for processing because that processing is in the digital domain.
Digital processing is a constraint.

- Rich

 

[markus]

Depends on what you mean by dynamically. Once the channel count and bass management then the configuration is static.
It does not change with program material.
Agreed, they are not likely changing the configuration with the source channel count.

The point of the thread, is that attenuation is occurring that is not needed for a 2-channel desktop DAC, that don't have room correction or crossovers.

ASR has found real issues measuring AVR/AVP's so there is room for improvement.
Due the the nature of the beast, an HT processor will never best the SINAD of a well designed 2-channel DAC.
An HT processor needs excellent SINAD to provide headroom for processing because that processing is in the digital domain.
Digital processing is a constraint.

- Rich

Just to clarify, no reason to change anything with program material but gain management based on input channel count could make sense. Playing 2 channel stereo on a 16 speaker configuration could recover a lot of headroom.

It's true that systems without bass management have an advantage as no gain management is necessary. But the issue with AVRs is probably not the processing (really easy to do high precision calculations these days) but how good or bad gain staging is managed. Dirac for example never "releases" any headroom it reserves, regardless whether it uses it or not. There are probably a lot more places inside an AVR/P where such things happen.
Now one could argue that most manufacturers are just integrators of products others have designed (Texas Instruments, ADI, AKM, etc) and they have no control over their inner workings but even Trinnov who do claim they have a radically different solution and do everything in software they write/control don't yield superior performance: https://www.avsforum.com/forum/90-r...lith-htp-1-owners-thread-54.html#post59242736