Dynamic compression room acoustics, or input clipping?

[SEKLEM]

With some caveats given, I recognize the inconsistency of my gear selection cuts against "audiophile". I'll try to be succinct, so if you feel I've left out critical information please don't hesitate to ask.

System is in a mixed used stereo / 5.1 home theater in a living room. I don't know the exact square footage, but I'm in a typical 1200sq foot ranch, the living area is open concept L shape with low ceilings.

Equipment:

Speakers: Sonus Faber Olympica Nova V

Amplifier: VTV Purifi EVAL1 (1ET400A)

Preamplifier: Onkyo TX-RZ900 (bass management 60Hz crossover)

Source: Apple TV4k (Apple Music)

Subwoofer: Rythmik F18

I'm very satisfied with the system performance overall. Subwoofer is nicely integrated, large open soundstage, the speakers disappear. Detailed, smooth, effortless with most music content and very forgiving with even poorer recordings. I can listen for hours, no fatigue.

BUT

I recently listened to Porcupine Tree's "Gravity Eyelids" with the level set to -10dB on the RZ900 and the instruments begin to smear together. Cymbal detail is washed out at this level of output.

I figure it could be one of three things:

- Amplifier is running out of current to supply the speakers.

- The reflections in the room are amplified at this level of output causing this smearing affect.

- The input level with the gain set to 29dB is overdriving the Purifi buffer board circuit.

 

[Spkrdctr]

You could be overdriving your tweeters. Without real detailed info on your exact set up it is very hard to tell you what is happening with your set up. It could be various things. The best way to fix it is to drop the volume down to where it sounds fantastic! You control the system, it does not control you!

 

[Redacted]

Sensitivity	90 db SPL (2.83V/1 m)
Nominal impedance	4 Ohm
Recommended power amplifier	60W - 400W, undistorted signal

You are probably overloading the speakers. But this will be deafening.

 

[SEKLEM]

You could be overdriving your tweeters. Without real detailed info on your exact set up it is very hard to tell you what is happening with your set up. It could be various things. The best way to fix it is to drop the volume down to where it sounds fantastic! You control the system, it does not control you!

What detailed information is missing that would be instrumental in tacking this down?

The room isn't very large, again, this is a 1200 square foot ranch home. I'm less than 6 feet away from the speakers, and I don't listen at crazy output levels that are uncomfortable, I'm conscious of protecting my hearing. I never go to full reference level (0dB), I don't crank my stereo and leave the room, I listen in the sweet spot.

You are probably overloading the speakers. But this will be deafening.

Pardon my ignorance, are you implying the Purifi is producing more clean power than the Sonus Faber can handle, or that the signal is clipped and overloading it that way?

 

[Keith_W]

Nobody knows the answer. The only answers you will get will be pure speculation. There is only one way to find out:

 

[SEKLEM]

Nobody knows the answer. The only answers you will get will be pure speculation. There is only one way to find out:

View attachment 542226

How is Room EQ Wizard going to show me if amplifier clipping, input clipping or room reflections are the culprit?

 

[OldHvyMec]

and the instruments begin to smear together. Cymbal detail is washed out at this level of output.

Are you saying it gets worse as you play longer? That is pointing to VC getting hot and starting to distort. It's one thing I seldom worry about when using the types of drivers I use. Ribbons and planars.

28mm H28 XTR3 soft dome with DAD (Damped Apex Dome) is pretty exclusive to that manufacturer and has a great reputation for how they subjectively sound.

Having personally listened to that particular speaker quite a few times and over long periods of time, listen to one speaker at a time. I would never associate what you're describing with that speaker unless there was a distortion issue with either a driver developing a problem or something as simple as the driver's screws have started to work loose and need a little maintenance.

Check the drivers and make sure they are secure, and listen to one complete speaker at a time.

THEN go get your ears checked, just in case you can't figure it out.

Regards

 

[SEKLEM]

Are you saying it gets worse as you play longer? That is pointing to VC getting hot and starting to distort. It's one thing I seldom worry about when using the types of drivers I use. Ribbons and planars.

28mm H28 XTR3 soft dome with DAD (Damped Apex Dome) is pretty exclusive to that manufacturer and has a great reputation for how they subjectively sound.

Having personally listened to that particular speaker quite a few times and over long periods of time, listen to one speaker at a time. I would never associate what you're describing with that speaker unless there was a distortion issue with either a driver developing a problem or something as simple as the driver's screws have started to work loose and need a little maintenance.

Check the drivers and make sure they are secure, and listen to one complete speaker at a time.

THEN go get your ears checked, just in case you can't figure it out.

Regards

Thank you for your questions.

It's not a consequence of how long I listen, moreso the level of output. I measured it once before with an inexpensive digital SPL meter and I believe it was peaking around 90-93dB when I noted it started to sound "compressed".

Interestingly enough, the tweeters on these had to be replaced by an authorized dealer in Cincinati (Hanson Audio) by the previous owner before I could take delivery because both tweeters were DEAD. I've not noticed any buzzing or oddities with them. They definitely are detailed while being relaxed with almost my entire music library under normal listening conditions. I'm only looking to fill in the gaps of performance as I'm going along to ensure I'm getting the most out of these speakers, which I consider a sizeable investment.

I will try listening to one speaker at a time though, that's an interesting idea.

I haven't had my hearing tested in decades, but I can still hear up to 16kHz.

Some additional information.

The Onkyo RZ900 can produce 2.83v cleanly according to Sound & Vision's bench testing. It's a curious result, leading one to wonder if it may supply more clean voltage. The input buffer on the VTV amplifier is 27dB as configured. One of my assumed possible culprits would be the Onkyo's preamplifier is overdriving the input buffer creating this smearing effect. While the speakers (read tweeters) don't glare the instrumentation becomes less delineated with "Gravity Eyelids" at higher output levels. When output from the Onkyo is pulled back the details are more vivid.

Hope this sheds some light on what I'm dealing with.

 

[Keith_W]

How is Room EQ Wizard going to show me if amplifier clipping, input clipping or room reflections are the culprit?

By measuring at different volumes and comparing them. You won't be able to tell if it's the amplifier or the speaker by looking at the distortion products alone, unless you swap the amplifier and re-measure. And examining room reflections are trivial with REW. You just need to know where to look. See the eBook in my sig.

 

[Chris A]

...BUT

I recently listened to Porcupine Tree's "Gravity Eyelids" with the level set to -10dB on the RZ900 and the instruments begin to smear together. Cymbal detail is washed out at this level of output.

I figure it could be one of three things:

- Amplifier is running out of current to supply the speakers.

- The reflections in the room are amplified at this level of output causing this smearing affect.

- The input level with the gain set to 29dB is overdriving the Purifi buffer board circuit.

You are right about the clipping (if referring to the particular track that you identified). The problem is...it's not your gear that's clipping. Here's a downloaded version of that track in the waveform view from YouTube, viewed from within Audacity:

I ran Clip Fix and found 4.25 dB of clipping was used on this particular version of the track. All that red in the waveform view is clipping. The ReplayGain figure is -8.85 dB, which is also at least 4.9 dB too high (i.e., any typical track maximum should never be higher than -3.9 dB on the ReplayGain scale, I've found empirically).

After I ran Clip Fix and then Amplify (at -4.25 dB), here is the resultant track waveform:

Here is the spectrogram view of that same de-clipped track, above:

Here is the resultant cumulative spectrogram view of the track, after all the artificial odd-order clipping has been removed from the track. Notice the broad hump centered on 3 kHz and the attenuation below 50 Hz:

I then used the following demastering EQ curve to bring the track to a more listenable result (i.e., partially undoing the mastering EQ that was added to the track before release):

Here is the resultant cumulative spectrogram (the "plot spectrum" view) to verify that the demastering curve produces a reasonable result:

Just listening to the newly declipped and de-EQed (i.e., the new demastered) track, it's pretty obvious that both the piano and drum stems in the mix have been very heavily compressed--to the point that it sounds a bit ragged and artificial--nothing like the original instruments sound like.

Bottom line: you would do well to have a look at your "reference music" tracks using something like Audacity to see what is actually happening.

Personally, I've found that Porcupine Tree albums are quite heavily clipped and sometimes very heavily compressed. After declipping and re-EQing, I found that the multichannel version of Fear of a Blank Planet and the stereo (CD) version of Stupid Dream are both very listenable, but nothing like any of the more hi-fi recordings that I own. I would not use any Porcupine Tree track to judge fidelity of a setup.

There is a LOT of very badly damaged music out there (virtually all popular music genres produced since 1991).

JMTC/YMMV.

Chris

 

[SEKLEM]

You are right about the clipping (if referring to the particular track that you identified). The problem is...it's not your gear that's clipping. Here's a downloaded version of that track in the waveform view from YouTube, viewed from within Audacity:

View attachment 542306

I ran Clip Fix and found 4.25 dB of clipping was used on this particular version of the track. All that red in the waveform view is clipping. The ReplayGain figure is -8.85 dB, which is also at least 4.9 dB too high (i.e., any typical track maximum should never be higher than -3.9 dB on the ReplayGain scale, I've found empirically).

After I ran Clip Fix and then Amplify (at -4.25 dB), here is the resultant track waveform:

View attachment 542307
Here is the spectrogram view of that same de-clipped track, above:

View attachment 542309

Here is the resultant cumulative spectrogram view of the track, after all the artificial odd-order clipping has been removed from the track. Notice the broad hump centered on 3 kHz and the attenuation below 50 Hz:

View attachment 542311

I then used the following demastering EQ curve to bring the track to a more listenable result (i.e., partially undoing the mastering EQ that was added to the track before release):

View attachment 542310

Here is the resultant cumulative spectrogram (the "plot spectrum" view) to verify that the demastering curve produces a reasonable result:

View attachment 542318

Just listening to the newly declipped and de-EQed (i.e., the new demastered) track, it's pretty obvious that both the piano and drum stems in the mix have been very heavily compressed--to the point that it sounds a bit ragged and artificial--nothing like the original instruments sound like.

Bottom line: you would do well to have a look at your "reference music" tracks using something like Audacity to see what is actually happening.

Personally, I've found that Porcupine Tree albums are quite heavily clipped and sometimes very heavily compressed. After declipping, I found that the multichannel version of Fear of a Blank Planet and the stereo (CD) version of Stupid Dream are both very listenable, but nothing like any of the more hi-fi recordings that I own. I would not use any Porcupine Tree track to judge fidelity of a setup.

There is a LOT of very badly damaged music out there (virtually all popular music genres produced since 1991).

JMTC/YMMV.

Chris

Then perhaps there's a psychoacoustic reason why I felt the details were more defined at lower levels. I'll have to relisten to the track again this evening.

 

[formdissolve]

That type of loudness war clipping and brickwall has been the bane of music for 25+ years..

 

[babadono]

Get the de clipped, de mastered version from Chris...or better yet learn how to do what he does.

 

[SEKLEM]

I listened again and it seems like my initial impression that it sounded more coherent at lower levels may not be true. I’m not in love with the album or anything, but I thought it maybe I found a limit to the system rather than a badly clipped song.

I played Tool’s Chocolate Chip Trip, and while not as complex I was able to run it at similar levels without any noticeable compression or loss of resolution. Seems I don’t have much to be worried about with equipment, just need to learn to recognize bad mastering for what it is and stop worrying about some imagined bottleneck.

 

[Chris A]

Seems I don’t have much to be worried about with equipment, just need to learn to recognize bad mastering for what it is and stop worrying about some imagined bottleneck.

The Tool album Fear Inoculum is very different than the Porcupine Tree In Absentia. Fear Inoculum is very straightforward to run Clip Fix and then Amplify to fix the much smaller amount of clipping found in the Tool album tracks. I would go as far as to say that Fear Inoculum is not a bad album from a hi-fi viewpoint (especially after fixing the clipping). There is a much smaller amount of clipping present on that album than most other hard rock or metal albums, and the sound is very full--making the Tool tracks not a bad album to choose to listen to different hi-fi setups to hear their capabilities.

Like I mentioned, you can download and install a free copy of Audacity, and then use that to see what's happening on your problematic music tracks.

More often than not, it's immediately apparent from just looking at any "freshly mowed grass" in the waveform view that a track has been limited too much--in order to make the tracks sound louder. If that's the case, then usually there are other issues with the track that may not be immediately visible, but are definitely audible.

Chris

 

[OldHvyMec]

Interesting and informative thread.

Regards

 

[Philbo King]

With some caveats given, I recognize the inconsistency of my gear selection cuts against "audiophile". I'll try to be succinct, so if you feel I've left out critical information please don't hesitate to ask.

System is in a mixed used stereo / 5.1 home theater in a living room. I don't know the exact square footage, but I'm in a typical 1200sq foot ranch, the living area is open concept L shape with low ceilings.

Equipment:

Speakers: Sonus Faber Olympica Nova V

Amplifier: VTV Purifi EVAL1 (1ET400A)

Preamplifier: Onkyo TX-RZ900 (bass management 60Hz crossover)

Source: Apple TV4k (Apple Music)

Subwoofer: Rythmik F18

I'm very satisfied with the system performance overall. Subwoofer is nicely integrated, large open soundstage, the speakers disappear. Detailed, smooth, effortless with most music content and very forgiving with even poorer recordings. I can listen for hours, no fatigue.

BUT

I recently listened to Porcupine Tree's "Gravity Eyelids" with the level set to -10dB on the RZ900 and the instruments begin to smear together. Cymbal detail is washed out at this level of output.

I figure it could be one of three things:

- Amplifier is running out of current to supply the speakers.

- The reflections in the room are amplified at this level of output causing this smearing affect.

- The input level with the gain set to 29dB is overdriving the Purifi buffer board circuit.

Using a single album to judge fidelity is a bit of a pitfall. It could be due to how that album was mixed or mastered.

 

[SEKLEM]

Using a single album to judge fidelity is a bit of a pitfall. It could be due to how that album was mixed or mastered.

I was only using it as a test for high output. Most music I listen to is not that heavy and also doesn't beg me to turn up the volume. I was under the impression that this album was well mastered with excellent dynamic range and low distortion, that was a mistake apparently.

 

[Spkrdctr]

I was only using it as a test for high output. Most music I listen to is not that heavy and also doesn't beg me to turn up the volume. I was under the impression that this album was well mastered with excellent dynamic range and low distortion, that was a mistake apparently.

We all make mistakes. No worries.

 

[MA043HA]

I wonder what Porcupine Tree lead man Steven Wilson would say about his mixing and mastering choices (or approval of others choices) since he is renowned for remixing and remastering many classic rock albums in both stereo and multi-channel. I wonder if the YouTube version Chris downloaded is significantly different than the stereo mix/master from a CD or Bluray. I may have to check it when I get home tonight or tomorrow.