Print

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

[mwillems]

So I've been experimenting with the Linkwitz Transform in PEQ, and the four primary parameters seem self explanatory enough (current Fs, current Q, target Fs, target Q), and the signal shaping in Analyzer appears to do what I'd expect based on those parameters. 

One parameter I didn't notice immediately, but I'm not sure I fully understand the operation of is the check box labelled "Adjust volume of all channels to prevent clipping."  Obviously the linkwitz transform boosts the sub signal, so it's going to have higher peaks than the other channels, and might engage clip protection.  So I'm guessing the check box recenters all the outputs downward by some amount to accommodate the higher sub peaks without engaging clip protection. 

If this is correct, how does it determine how far down to recenter the signal?  Is it a fixed amount or is it based on the slope of the specific linkwitz transform?  Does it account for any other gain applied to the sub channel (whether through room correction or otherwise)?  I'm asking because it seems to decrease the volume drastically (18 dB or so), which is higher than I would expect based on the slope of my specific linkwitz transform and I'm trying to figure out why that would be.

Thanks for any insight.

[mwillems]

Bump. 

I'd appreciate any info on how the Linkwitz transform volume reduction is determined as described above (is it fixed or based on the specific transform dialed in)?  I'm getting what appears to be an 18 dB reduction, but based on the linkwitz transform spreadsheet (from linkwitz labs), the Eq curve for the linkwitz transform I've dialed in should only involve 9 or 10 dB of gain (meaning the volume reduction from JRiver is much steeper). 

If JRiver just applies a fixed reduction, I can disable and manually handle the offset based on the actual gain, but if the reduction is based on JRiver's estimation of the actual gain applied, I wouldn't necessarily want to do that (i.e. it may mean JRiver is applying a different gain curve than the spreadsheet, or that my math is wrong about the gain curve). 

Any information on this (especially from the devs) would be appreciated.

[Matt]

I'd appreciate any info on how the Linkwitz transform volume reduction is determined as described above (is it fixed or based on the specific transform dialed in)?

It's just a simple -12dB.

As an aside, if you're using a Linkwitz I'd recommend 'Internal Volume' so you have some extra headroom:
http://wiki.jriver.com/index.php/Volume#Internal_Volume_Headroom

[mwillems]

It's just a simple -12dB.

As an aside, if you're using a Linkwitz I'd recommend 'Internal Volume' so you have some extra headroom:
http://wiki.jriver.com/index.php/Volume#Internal_Volume_Headroom

Interesting.  I guess I need to figure out where the other 6 dB of attenuation is coming from.  I had to change around my PEQ settings when I enabled the Linkwitz (removing other shaping I'd added to the sub, etc.), so that may be the source of the delta.

I do use internal volume, which is actually why I was asking.  I have about 11dB of headroom at my reference level, so I was thinking about unchecking the downward adjustment box.  I just wanted to make sure it wasn't adaptive first; i.e. I wanted to make sure that it wasn't smarter than I was :-)

Thanks so much!

[Jargon]

I am currently using a Hardware LT device for my two sealed subwoofers, the BASSIS from Marchand Electronics http://www.marchandelec.com/wm8.html

Previously I was also using a Velodyne SMS-1 in conjunction to BASSIS that I successfully "eliminated" through the use of PEQ in JRiver with better results due to increased accuracy.

Now, I was thinking that I could possibly also remove this "box" that stands just before the sub amplifier meaning that it is adding the audio signal last after it has left my HTPC and any audio processing done through it.

Is there any possible downside in using the DSP Studio Linkwitz Transform compared to a Hardware solution like the BASSIS?

Is there any benefit either? What about the headroom you discussed by applying the signal? The way I have it connected now, the LT is applied ONLY to the sub signals independently of the rest of the channels. The BOOST I apply is 12db.

Any input would be greatly appreciated.

[mwillems]

I am currently using a Hardware LT device for my two sealed subwoofers, the BASSIS from Marchand Electronics http://www.marchandelec.com/wm8.html

Previously I was also using a Velodyne SMS-1 in conjunction to BASSIS that I successfully "eliminated" through the use of PEQ in JRiver with better results due to increased accuracy.

Now, I was thinking that I could possibly also remove this "box" that stands just before the sub amplifier meaning that it is adding the audio signal last after it has left my HTPC and any audio processing done through it.

Is there any possible downside in using the DSP Studio Linkwitz Transform compared to a Hardware solution like the BASSIS?

Is there any benefit either? What about the headroom you discussed by applying the signal? The way I have it connected now, the LT is applied ONLY to the sub signals independently of the rest of the channels. The BOOST I apply is 12db.

Any input would be greatly appreciated.

I don't see any serious downsides; two immediate upsides that jump out at me are 1) more precise transforms (you can plug in the exact numbers instead of trying to line up a pot with hash marks) and 2) one less analog stage in the signal chain reducing (very incrementally) your total added distortion, allowing fewer opportunities for ground hum, etc.  

The only potential downside would be slightly reduced software volume output to account for the 12 dB (in your case) of gain, which, if you use the internal volume feature, may not even be an issue in your circumstances.  For example, I use internal volume as my volume control and leave my amps all the way open (after calibrating my speaker levels appropriately).  That ensures I have lots of headroom available for DSP and don't run the risk of engaging clip protection.  

[Jargon]

Ah you spelled the right words! Less ground hum!

The main reason I want to remove it is this! Hum! At the moment I am passing audio through the HDMI to an Onkyo 5507 and from there through balanced outputs to balanced inputs on amplifiers. The BASSIS is the only non balanced device and the only one producing a hum as long as I turn it on. It is connected to a Bridged big amplifier (QSC RMX5050) that is driving the subs and I had to turn its gain all the way down to make the hum tolerable! The benefits you mentioned are certainly going to be appreciated.

My main concern is volume headroom but as you said I will switch to internal volume to avoid any issues. But what happens when I listen to reference level? Yet, for now and untill I switch to another device I will still have to use the Preamps volume (Onky 5507) since it will still be the last device on the chain feeding the amps and doing the D/A conversion.

And one more worry. Even if was to remove the preamp with the hardware volume, what about any other sound outside J River that could be let out without any volume control then damaging the speakers? I know I could disable all system sounds but still...I am worried of having nothing between the PC and some thousands of watts on the loose on the amplifiers! If any signal goes out unattenuated it could cause damage!

[mwillems]

Ah you spelled the right words! Less ground hum!

The main reason I want to remove it is this! Hum! At the moment I am passing audio through the HDMI to an Onkyo 5507 and from there through balanced outputs to balanced inputs on amplifiers. The BASSIS is the only non balanced device and the only one producing a hum as long as I turn it on. It is connected to a Bridged big amplifier (QSC RMX5050) that is driving the subs and I had to turn its gain all the way down to make the hum tolerable! The benefits you mentioned are certainly going to be appreciated.

My main concern is volume headroom but as you said I will switch to internal volume to avoid any issues. But what happens when I listen to reference level? Yet, for now and untill I switch to another device I will still have to use the Preamps volume (Onky 5507) since it will still be the last device on the chain feeding the amps and doing the D/A conversion.

And one more worry. Even if was to remove the preamp with the hardware volume, what about any other sound outside J River that could be let out without any volume control then damaging the speakers? I know I could disable all system sounds but still...I am worried of having nothing between the PC and some thousands of watts on the loose on the amplifiers! If any signal goes out unattenuated it could cause damage!

In re: reference level, JRiver allows you to set your reference level in JRiver based on calibration, and (with your amps open) it should be significantly less than 100% software volume.  Every dB below 100% is additional signal overhead available for DSP, even when you have it set to your reference volume.  Check out the wiki entry on volume and reference calibration for more info: http://wiki.jriver.com/index.php/Volume

As to the risk of stray sounds, here's my solution:  I set my windows default sound device to a sound device that isn't hooked up to any speakers (i.e. a motherboard soundcard, or (in my case) an HDMI out to my television (which is muted)).  The only program that is set to output sound to my actual sound device (i.e. the one hooked to my amp and speakers) is JRiver.  If I want to hear system sounds or a video game, I turn on JRiver's "loopback" function, which redirects the audio from the "default" sound card  and feeds it through JRiver to the "real" sound card.

I have not had any "stray" sounds (e.g. unexpected sounds) in several months with this set up.  The only real "danger" moment was when I physically whacked the analog outputs of my sound card (I stumbled while trying to turn off an amp).  One of them came slightly loose and the "loose connection" hum was loud and alarming.  But that is not a software fault, that is a user fault :-)

For more info on setting up loopback, check out mojave's excellent posts in this thread: http://yabb.jriver.com/interact/index.php?topic=70242.50

[Jargon]

Ok now I get so it works like the receivers volume who implement Room Correction and in the case of my Onkyo it allows +18db although if I calibrate with Audyssey the ref level is at 0db volume.

So I would have to set the Onkyo at +18db (ouch!) to completely bypass its hardware volume control then do as you say with setting the default audio device and be done with it.

I am already using the loopback feature for REW measurements and to test EQ settings in REW as well so I am familiar with that and it is also great since you can use all the DSPs of JRiver with any audio application as well. I am also planning to get Audiolense DRC as well.

I will look into that tomorrow.

[mwillems]

Ok now I get so it works like the receivers volume who implement Room Correction and in the case of my Onkyo it allows +18db although if I calibrate with Audyssey the ref level is at 0db volume.

So I would have to set the Onkyo at +18db (ouch!) to completely bypass its hardware volume control then do as you say with setting the default audio device and be done with it.

I am already using the loopback feature for REW measurements and to test EQ settings in REW as well so I am familiar with that and it is also great since you can use all the DSPs of JRiver with any audio application as well. I am also planning to get Audiolense DRC as well.

I will look into that tomorrow.

Be careful and don't do anything that might jeopardize your system.  

I should mention that all of my speakers have RMS power handling ratings significantly higher than my amps' rated outputs so my "worst case" scenario (barring something truly unusual) is just earsplitting volume.  If your amp can put out much more power than your speakers can safely handle, you might want to find some middle ground setting instead of rolling it up to the maximum (as a safety measure).

[dean70]

Ok now I get so it works like the receivers volume who implement Room Correction and in the case of my Onkyo it allows +18db although if I calibrate with Audyssey the ref level is at 0db volume.

So I would have to set the Onkyo at +18db (ouch!) to completely bypass its hardware volume control then do as you say with setting the default audio device and be done with it.

I am already using the loopback feature for REW measurements and to test EQ settings in REW as well so I am familiar with that and it is also great since you can use all the DSPs of JRiver with any audio application as well. I am also planning to get Audiolense DRC as well.

I will look into that tomorrow.

Actually the Onkyo has zero cross detection in the volume circuit, so it is actually boosting the signal by +18db via an internal op-amp in this case.

[Jargon]

Be careful and don't do anything that might jeopardize your system.  

I should mention that all of my speakers have RMS power handling ratings significantly higher than my amps' rated outputs so my "worst case" scenario (barring something truly unusual) is just earsplitting volume.  If your amp can put out much more power than your speakers can safely handle, you might want to find some middle ground setting instead of rolling it up to the maximum (as a safety measure).

My LCRs for example can handle up to 2,000 Watts peak each (3 X Danley SH50s) and my subs are fed by a bridged QSC RMX 5050 set to output 5,000 Watts. A Bryston 14BSST which I also use for L&R can put out 900-1000 watts at 4Ohms.

The Danley's have a sensitivity of 101db (maybe more in room) and I can't imagine what a high pitched noise could do to my ears if they are fed with 1,000 watts instantly. It's not about the speakers I'm only worried but my ears as well.

Maybe some middle ground as you say is the best. I will probably settle for 0db setting on the Onkyo. Right now I have it locked at -4db max and the reason is not speaker protection but ear safety.

[Jargon]

Actually the Onkyo has zero cross detection in the volume circuit, so it is actually boosting the signal by +18db via an internal op-amp in this case.

Isn't this feature implemented to avoid clicking when changing volume fast? What do you mean by boosting via an internal opamp with regards to headroom?

And since you brought it up I've always wondered what the +18db setting really represents with regards to the input signal and 0dbfs. Is the +18db setting for example the same as 100% in JRiver internal volume?

Damn I need to do some reading to clear all this up in my head.

[Jargon]

Ok I've done some reading and the way I understand it is that the Onkyo at 0db (relative scale) volume is at unity gain meaning no negative or positive attenuation.

The XLR outs give a 2 Volts / 490 Ohms and that comes from Onkyo. Does this mean that by setting the volume at zero I let the signal at 2V and any +/- adjustment increase or decrease this value?

I am just wondering what is the meaning of +18db volume setting on Onkyos scale (the relative)? What happens to the signal when you go above 0db supposing your speakers/amps can take it? I mean is the + attenuation harmful?

Is there some material I can read on the subject?

[mwillems]

Ok I've done some reading and the way I understand it is that the Onkyo at 0db (relative scale) volume is at unity gain meaning no negative or positive attenuation.

The XLR outs give a 2 Volts / 490 Ohms and that comes from Onkyo. Does this mean that by setting the volume at zero I let the signal at 2V and any +/- adjustment increase or decrease this value?

I am just wondering what is the meaning of +18db volume setting on Onkyos scale (the relative)? What happens to the signal when you go above 0db supposing your speakers/amps can take it? I mean is the + attenuation harmful?

Is there some material I can read on the subject?

2 volts is the usual "consumer" line output level.  Pro audio devices sometimes have a higher line output.  Does changing the gain setting affect the volume coming out of the XLRs?  Often line outputs are isolated from volume adjustments (i.e. they bypass the internal high-gain amp architecture).  If the volume control does affect the line out volume, then it sounds like the Onkyo allows you to boost the line output using the op-amps built into it.  Generally multiple high gain amplification steps aren't desirable, but some power amps require a high level input to drive them to full potential.  I would have a look at your power amps specs and see if they have any notation about what kind of input signal is required.  Most amps run just fine on a regular line level signal.

If your amps can be fully driven by a regular line out, I'm not sure there's any obvious advantage to boosting it at the Onkyo.  The potential disadvantages (especially with 18 dB of gain) are two high gain amplifier stages (with double the distortion), and the fact the amplifier stage in the Onkyo might not be as powerful and/or low noise as your power amplifiers (so you may be tripling or quintupling your distortion rather than doubling it).  

It sounds to me as though the 0dB setting (unity gain) is probably a good starting point, but I'd suggest paddling around and see what sounds good to you.

[mojave]

Maybe some middle ground as you say is the best. I will probably settle for 0db setting on the Onkyo. Right now I have it locked at -4db max and the reason is not speaker protection but ear safety.

In JRiver you can go to Tools > Options > Audio > Volume and set the maximum volume there. This is zone independent so you could setup a zone for kids that has a lower maximum volume available. There is also the ability to set the Reference Level. Using Tools > Advanced Tools > Audio Calibration you can play bandwidth limited pink noise and set the level of your speakers to 83 dB. This might be at 75% volume in JRiver so you put 75 in the Internal Volume Reference Level field. Once you do that, JRiver will show you how many dB you are above/below Reference Level.

I had the Danley SH50 spec page open in another browser tab when I read your posts. There is a remote chance I may hear them in a couple of months. JRiver member TheLion also has Danley SH50's.

Let us know how it goes using the L-T in JRiver vs the BASSIS.

[dean70]

Isn't this feature implemented to avoid clicking when changing volume fast? What do you mean by boosting via an internal opamp with regards to headroom?

And since you brought it up I've always wondered what the +18db setting really represents with regards to the input signal and 0dbfs. Is the +18db setting for example the same as 100% in JRiver internal volume?

darn I need to do some reading to clear all this up in my head.

I am only going by the block diagram in the service manual. It has 1 section for -99 to 0 and a second stage for +1 to +18db. You can find these on hifi-manuals dot com.

[Jargon]

2 volts is the usual "consumer" line output level.  Pro audio devices sometimes have a higher line output.  Does changing the gain setting affect the volume coming out of the XLRs?  Often line outputs are isolated from volume adjustments (i.e. they bypass the internal high-gain amp architecture).  If the volume control does affect the line out volume, then it sounds like the Onkyo allows you to boost the line output using the op-amps built into it.  Generally multiple high gain amplification steps aren't desirable, but some power amps require a high level input to drive them to full potential.  I would have a look at your power amps specs and see if they have any notation about what kind of input signal is required.  Most amps run just fine on a regular line level signal.

If your amps can be fully driven by a regular line out, I'm not sure there's any obvious advantage to boosting it at the Onkyo.  The potential disadvantages (especially with 18 dB of gain) are two high gain amplifier stages (with double the distortion), and the fact the amplifier stage in the Onkyo might not be as powerful and/or low noise as your power amplifiers (so you may be tripling or quintupling your distortion rather than doubling it).  

It sounds to me as though the 0dB setting (unity gain) is probably a good starting point, but I'd suggest paddling around and see what sounds good to you.

My amps are a Bryston 14BSST (Danley SH50s) for the L&R two Crown XLS 802D for Center and Surrounds and a QSC RMX 5050. Among my plans is to get a 3rd Bryston 7BSST for the center (also SH50) or go sell everything and go to Hypex Ncore monoblocks after I test sometime within the next 3 months from a friend.

For now the Bryston accepts either unbalanced input (no gain) Balanced REG (no gain) or Balanced +6db. As I found out that the +6db is through attenuation I left the switch on Balanced REG that had an immediate 6db effect at the speaker levels on the calibration (i.e. from -11db to 5.5db).

The XLRs on the Onkyo output 2 volt /490 ohms but the RCAs output 1 volt /470 ohms according to Onkyo (a friend mailed them and replied with this info). 

Since all those are new areas for me that I am currently reading into, how does the preamp increase / decrease the volume if the line out voltage remains constant? (Is it through current alternations or what?).

The 5507 has no amplifier stages since it is a preamp only so any distortion can only occur at the preamplification stage. I am mentioning this since you talked about noise levels in its amp stages unless you mean the same (preamplification).

So in essence what is the meaning of a +18db volume level and what is the reason of its existence?

[Jargon]

In JRiver you can go to Tools > Options > Audio > Volume and set the maximum volume there. This is zone independent so you could setup a zone for kids that has a lower maximum volume available. There is also the ability to set the Reference Level. Using Tools > Advanced Tools > Audio Calibration you can play bandwidth limited pink noise and set the level of your speakers to 83 dB. This might be at 75% volume in JRiver so you put 75 in the Internal Volume Reference Level field. Once you do that, JRiver will show you how many dB you are above/below Reference Level.

I had the Danley SH50 spec page open in another browser tab when I read your posts. There is a remote chance I may hear them in a couple of months. JRiver member TheLion also has Danley SH50's.

Let us know how it goes using the L-T in JRiver vs the BASSIS.

For the moment I have set the levels and distances for speakers in Onkyo but the goal is to remove it from the Audio chain. This will happen when I get a Metric Halo ULN-8 or LIO8 with 4 preamps and a Lynx AES16 SRC to directly output from the PC to it.  The problem is that these cost about 5,000 Euros and right now I am "dry" lol!

The 3 X SH50 Danley's I have I got from The Lion since he moved and had to let them go in the new space. So far I am really excited with them and my previous speaker system was B&W 800Ds (L&R), HTM1D Center (still have this), 4X 803Ds for surrounds. I have not regret it for one second for making the switch but I have to say that the Danley's need A LOT of work to be placed properly and sound at their best. To begin with, I have to custom build stands for them that allow height adjustments, tilt and also have wheels (concealed steel balls like the B&W 800) for toe in angling. Unless you can play with these variables easily you won't get the best of them. Yet, considering the are almost 65Kg each these stands need to be heavy duty and as I want them to look nice they are going to cost (maybe more than 1,000 Euros). I am mentioning this since many people are intrigued by them but they are not an out of the box solution and in my opinion room treatments are also needed but they always do.

The next step will be DRC through convolution and most probably with Audiolense. I will keep you posted if interested.


Finally, the BASSIS. I did test the LT through J River today and tried to compare the two of them. As I have the exact measured Fz /Qz values of my twin subs as well as the Qp I use in the BASSIS (0.707) the only thing to calculate was the Fp.

On the BASSIS there is no dial for the target Fp but a Boost dial instead that I have set at 12db. According to the spreadsheet on the Linkwitz website it corresponds to an Fp of 24Hz when starting with Fz 48Hz and Qz 1 which are my fixed values.

It does seem to work properly but somehow something was wrong with my measurements and REW that by the time I resolved it I had to stop sine it was kid's bed time and sub testing although I am in a dedicated room is till not exactly quiet. So I will retest tomorrow.

My subs with corner loading (concrete walls /floor / ceiling) go down very low almost 10Hz (4X 15" AV15H drivers fed 5,000 Watts) and somehow it seemed to me (before REW crashed) that I was losing extension with the software LT vs the BASSIS but I want to do a thorough check to conclude. There was a problem with WASAPI loopback during the measurements as if it got stuck and some DSP filters were not being applied. Nevertheless, I will test thoroughly again.

FYI, I am using an Earthworks M30 with a Focusrite 2i2 (flat to 10Hz) for measurements and have its calibration file loaded in REW.   

[Jargon]

I am only going by the block diagram in the service manual. It has 1 section for -99 to 0 and a second stage for +1 to +18db. You can find these on hifi-manuals dot com.

It needs "yet" another account to download anything from there but I can't see the 5507 pro there.

In my case and using the relative scale (not the absolute - you can switch) the range is from -00/-81.5db to +18db. I somehow managed to find that at 0db you are at unity gain i.e. no attenuation on the signal positive or negative.

So above 0db the signal gets +gain. So my question was what happens when you add gain above unity gain supposing that your speakers and amps can take it.

The reason I am asking is because the Danleys even at reference levels at 4 meters do not get near their limits at all and there is still headroom for more. Yet, although the speakers are rated for 1000W continuous / 2000W peak and my Bryston can put out a bit more than 1000 Watt @ 4ohms (all Brystons go a bit above rated specs which 900Watts in this case) it is NEVER hot to the touch while with my B&W 800Ds it got really hot after high SPLs. On the other hand the SH50s have 101db sensitivity.

So if I wanted to drive them at their limits (with ear protection) and or sit at 8 meters I would probably have to go above 0db or use + values on the speaker levels menu which are at average -6db each now.


So for one more time, what is the use of the +18db setting on the volume control then? Has anyone ever used it?

[mojave]

The miniDSP website has a biquad calculation spreadsheet that includes a Linkwitz Transform worksheet. You can enter your before and after numbers in the spreadsheet to get a visual idea of what is going on. That has helped me in the past.



I use a Steinberg UR824 and Audiolense. I am very happy with the Steinberg and since it provides phantom power, I can just connect the mic to make measurements. I use an iSEMcon EMM-7101-CHTB mic with custom calibration to 5 Hz.

JRiver member TheLion does not have Danley SH50's.  

[Jargon]

I downloaded this as well but the one I use is also like that but has more graphs, gives a DC gain value, driver and box calculations, phase etc all with charts as well. Damn I can't remember where I got it but have a look here at post #2

http://www.diyaudio.com/forums/multi-way/37341-linkwitz-transform.html You can download it from there.

[mojave]

Jargon, did you do more testing and compare to the BASSIS yet? I know someone that might want to buy a Marchand BASSIS.