ProSoundWeb Community

Sound Reinforcement - Forums for Live Sound Professionals - Your Displayed Name Must Be Your Real Full Name To Post In The Live Sound Forums => LAB Lounge => Topic started by: Peter Hvedstrup on April 28, 2014, 03:04:29 AM

Title: DSP's sounding completely differing
Post by: Peter Hvedstrup on April 28, 2014, 03:04:29 AM
Hi all

I have read Bennetts article on DSP behavior and it seems that they sound and act very different.

My rig is a old Turbosound Floodlight with TSW-718 subs, so the settings are simple 24DB LR filters, delay and gain structure (no EQ). 

This weekend my trusty (very) old BSS FDS-388 started making a lot of noise in the left channel. Just before the show i swapped to a Ashly Protea 4.24c with the exact same settings. The result was a completely different sound. The tops sounded more crisp in the higs but a lot "colder" in the sound. The gain structure was way off, and i had to dampen to top end a lot and raise the sub a little to balance out things. The rig as a whole sounded a lot thinner with less sub and more aggressive highs. It turned out ok but the FDS-388 sounded way better on the Floods and TSW's. On other speakers i have always loved my Ashly.

Is it really impossible to switch DSP even with simple settings?

 
Title: Re: DSP's sounding completely differing
Post by: David Sturzenbecher on April 28, 2014, 07:53:45 AM
My guess is that you switched the bandwidth and Q parameters as Bennett points out in his article. I have never A/B the two processors but the Ashly is a solid unit.


Sent from my iPhone using Tapatalk
Title: Re: DSP's sounding completely differing
Post by: TJ (Tom) Cornish on April 28, 2014, 08:30:27 AM
Hi all

I have read Bennetts article on DSP behavior and it seems that they sound and act very different.

My rig is a old Turbosound Floodlight with TSW-718 subs, so the settings are simple 24DB LR filters, delay and gain structure (no EQ). 

This weekend my trusty (very) old BSS FDS-388 started making a lot of noise in the left channel. Just before the show i swapped to a Ashly Protea 4.24c with the exact same settings. The result was a completely different sound. The tops sounded more crisp in the higs but a lot "colder" in the sound. The gain structure was way off, and i had to dampen to top end a lot and raise the sub a little to balance out things. The rig as a whole sounded a lot thinner with less sub and more aggressive highs. It turned out ok but the FDS-388 sounded way better on the Floods and TSW's. On other speakers i have always loved my Ashly.

Is it really impossible to switch DSP even with simple settings?
Without conversion and/or testing, yes.  You read the article, right?  You sound surprised that your findings agree.

People complain about the modern trend of a manufacturer bundling the amp/DSP with the speaker package - Nexo requires NXAMP, JBL strongly recommends ITechHD, EAW recommends UX8800/UX3600, etc.  Your experience highlights why this is being done.

I still find it hard to believe that some folks balk at manufacturer-provided tuning.  In my experience, the difference that high-quality manufacture-supplied processing makes is not subtle, compared to the Smaart-dude-du-jour.
Title: Re: DSP's sounding completely differing
Post by: Merlijn van Veen on April 28, 2014, 08:35:09 AM

Is it really impossible to switch DSP even with simple settings?

Unfortunately not with guarantee as you've experienced yourself. The only way to port settings with confidence is to measure the specified DSP with factory settings using e.g. a dual-channel analyzer like Smaart and reconstruct the response on the new DSP.
Title: Re: DSP's sounding completely differing
Post by: John Roberts {JR} on April 28, 2014, 09:38:56 AM
Unfortunately not with guarantee as you've experienced yourself. The only way to port settings with confidence is to measure the specified DSP with factory settings using e.g. a dual-channel analyzer like Smaart and reconstruct the response on the new DSP.

I gave up trying to get the AES to support definition and standards, or better yet one standard for Q in boost/cut EQ sections. For conventional crossover filters the Q definitions are adequate. 

There are a few other subtle issues that can cause errors between DSP platforms.

I find this unacceptable for the year 2014 and I don't even use them...

Back in the very old days (think vinyl playback or magnetic tape) they would often publish graphs and tables of dB vs frequency for important equalization curves.

That would work today. Hint hint.. At least until a standard emerges.

Until then buy powered speakers.  8) 8)

JR
Title: Re: DSP's sounding completely differing
Post by: Mike Pyle on April 28, 2014, 11:36:39 AM
I still find it hard to believe that some folks balk at manufacturer-provided tuning.  In my experience, the difference that high-quality manufacture-supplied processing makes is not subtle, compared to the Smaart-dude-du-jour.

COST is the complaint. Using proprietary tuning that requires proprietary hardware is disproportionately expensive, compared to the cost of equal or even better aftermarket dsp and amps.
Title: Re: DSP's sounding completely differing
Post by: Ivan Beaver on April 28, 2014, 11:40:21 AM
Unfortunately not with guarantee as you've experienced yourself. The only way to port settings with confidence is to measure the specified DSP with factory settings using e.g. a dual-channel analyzer like Smaart and reconstruct the response on the new DSP.
Agreed.

There are all sorts of differences-so the best way to is simply overlay the traces-using whatever (eq-freq-gain-delay etc) to get BOTH the amplitude AND phase traces the same
Title: Re: DSP's sounding completely differing
Post by: John Roberts {JR} on April 28, 2014, 11:49:24 AM
Just thinking out loud... Could a manufacturer publish a sound file of say shaped noise or sine wave sweeps, that are the inverse complement of the desired result. Then end users could use their measurement system du jour, that we ASSume are all accurate, or more accurate than their DSP interfaces, and tweak EQ/crossovers for flat response.

It might work....

JR
Title: Re: DSP's sounding completely differing
Post by: TJ (Tom) Cornish on April 28, 2014, 11:54:25 AM
COST is the complaint. Using proprietary tuning that requires proprietary hardware is disproportionately expensive, compared to the cost of equal or even better aftermarket dsp and amps.
I would argue that this is a shrinking problem.  On the small side of things, self-powered boxes make this moot.  On the larger side of things, the performance difference and rider necessity of having a complete "system" make it likely a bad business decision to not get the matching amps. 

For the remaining passive boxes, aftermarket DSP or amps aren't better if they can't be made to sound as good as the manufacturer's recommended - they're just cheaper.
Title: Re: DSP's sounding completely differing
Post by: Peter Hvedstrup on April 28, 2014, 02:48:10 PM
Hi all

Thanks for your replies.

I'm not shure a lot of you read that i actually stated that no EQ was used at all. Floodlight doesn't use EQ. That's why i'm surprised that a simple setup of 24db LR filters and delay can sound so different.



Title: Re: DSP's sounding completely differing
Post by: John Roberts {JR} on April 28, 2014, 03:00:27 PM
Hi all

Thanks for your replies.

I'm not shure a lot of you read that i actually stated that no EQ was used at all. Floodlight doesn't use EQ. That's why i'm surprised that a simple setup of 24db LR filters and delay can sound so different.
They shouldn't

JR
Title: Re: DSP's sounding completely differing
Post by: Bob Leonard on April 28, 2014, 06:25:48 PM
Price is the deciding factor, IMO. I have used practically most of the DSPs that have graced the face of the earth, and remember specifically the surprise when moving from an Ashly to a dbx 260. Night and day using the JBL presets. Time marches on and newer presets prevail, still better, then oops, I bought a 480. The DSP was a joy to work with, however, JBL no longer had recent presets for the now discontinued DSP and I was left on my own. Transferring the 260's settings to the 480 was a giant pita, and after the transfer an additional 2 weeks of work with calls to both JBL and dbx resulted in a "usable" system.

Time marches on again, and my obsession with quality sound led to the purchase of a dbx 4800. Not for the faint of heart or weak of wallet, but by far one of the best DSPs on the planet. The sound is incredible, the support is incredible, the options are second to none. The difference between the 260 and the 4800??? Night and day, the 4800 walks away from the 260. The difference in manufacturer supplied presets?? Again, night and day with the 4800 addressing option the 260 doesn't have. So, price makes a difference, manufacturer makes a difference, and the presets make up the rest.
Title: Re: DSP's sounding completely differing
Post by: John Roberts {JR} on April 28, 2014, 07:43:17 PM
So, price makes a difference, manufacturer makes a difference, and the presets make up the rest.

From my understanding of the technology price should not make a significant difference wrt accuracy of presets.

Manufacturers apparently use different EQ topology, just like all the analog GEQs out there are different.

Manufacturer presets really need to be looked at in the context of the EQ design used to capture the preset. Imagine trying to apply a +4 dB boost at 2kHz using an analog 31 band GEQ. Your +4 dB from a Rane GEQ will probably be different than a DBX GEQ, or a Peavey GEQ. Making them digital makes them repeatable, but not the same. It's the design decisions that makes them different

JR
Title: Re: DSP's sounding completely differing
Post by: Jonathan Betts on April 28, 2014, 07:51:13 PM
I wish Harman would supply Danley Sound Labs with some presets for the original Crown ITech and HD amplifiers. Basically shooting in the dark here with tunings given to me by the engineers at Danley from a different processor.

For my SRX system I have always used the 4800 tunings as there are none for the ITech.
Title: Re: DSP's sounding completely differing
Post by: jason misterka on April 28, 2014, 08:22:31 PM
Hi all

Thanks for your replies.

I'm not shure a lot of you read that i actually stated that no EQ was used at all. Floodlight doesn't use EQ. That's why i'm surprised that a simple setup of 24db LR filters and delay can sound so different.

Peter-

I had a similar experience as you when moving from BSS Omnidrive to XTA-made TurboSound LMS-D6 processors on our Floodlights.  It was a surprisingly big difference on a preset that included NO EQ filters.  My guess at the time was simply different quality AD and DA chips.

Most of the difference was in the tightness of the subwoofer frequencies and also how the XTA sounded above 8k or so.

Interestingly, the TurboSound factory presets ARE different between the two processors.

Jason
Title: Re: DSP's sounding completely differing
Post by: David Sturzenbecher on April 28, 2014, 10:32:02 PM
I wish Harman would supply Danley Sound Labs with some presets for the original Crown ITech and HD amplifiers. Basically shooting in the dark here with tunings given to me by the engineers at Danley from a different processor.

For my SRX system I have always used the 4800 tunings as there are none for the ITech.

Most Danley speakers don't need tunings. But I am pretty sure Danley has an Itech in house, or at least did.  Last I heard they were dialing in limiter settings for a bunch of products but that was it. If the Danley guys can give you a transfer function of the curve in their Danley specific processor, you would be able to recreate those in the Itechs if you have something like a VIBox.
Title: Re: DSP's sounding completely differing
Post by: Tim McCulloch on April 28, 2014, 10:34:25 PM
I wish Harman would supply Danley Sound Labs with some presets for the original Crown ITech and HD amplifiers. Basically shooting in the dark here with tunings given to me by the engineers at Danley from a different processor.

For my SRX system I have always used the 4800 tunings as there are none for the ITech.

The original ITech processing was based on the dbx 480/4800.  You're using the best published numbers for your SRX.
Title: Re: DSP's sounding completely differing
Post by: Peter Hvedstrup on April 29, 2014, 03:46:52 AM
Peter-

I had a similar experience as you when moving from BSS Omnidrive to XTA-made TurboSound LMS-D6 processors on our Floodlights.  It was a surprisingly big difference on a preset that included NO EQ filters.  My guess at the time was simply different quality AD and DA chips.

Most of the difference was in the tightness of the subwoofer frequencies and also how the XTA sounded above 8k or so.

Interestingly, the TurboSound factory presets ARE different between the two processors.

Jason


Hi Jason

It sounds like you experienced something very similar to what i did. Clearer higs and a tighter sub. The added tightness i the sub makes it seem like it isn't playing as loud and low. 

On another system i did a switch from my Ashly to an XTA DP-226. Here i didn't notice any significant difference, maybe since they are both from a never generation than the FDS-388 Omnidrive.

As long as manufactures don't have anything to win in having a common standard for how a DSP should act there will be issues for people like me who can't afford the manufacturer recomended (branded) gear.