It might seem that way on the surface, but to me AudioLens or Acourate are not really about hardware DSP vs software DSP.
Imo, they are really about how they actually generate the filters for either global overlay, or individual filters for drivers.
Because in the end...hardware DOES implement the filters.....whether the hardware is a PC with DACs and I/O interface, or a dedicated outboard unit...
...all filters, however they are generated, get implemented with some form of hardware.
What sucks imo, is when filter generating programs produce fillers in some proprietary format that demands either we use their processing software on a PC, or worse some proprietary hardware. But I'm speaking from a FIR point of view...my understanding is IIR DSP is more of a can of worms...
So maybe I need to soften my opening comment, if Audiolens and Acourate are also employing IRR filters...???
No, you can. You just don't get the one touch auto room tune.. however you can do it better manually. So a DIYer who understands the procedure has the advantage.
This is were Mitchba's convolver is clever: it is open to most popular format as player.
And you can implement IIR through Rephase... i don't see Acourate or Audiolense behemoth software skipping on it.
Let's ask @mitchba directly.
I agree with you Allen, choices made by humans will still have advantage over a routine. But it's not as simple as the push of a button, and not all want to spend time on this.
That said, only dsp can do such things as compensate for room mode compensation in time domain ( what Mitchba called 'low freq ER compensation' in Audiolense's review or his ebook). For things like this analog is useless.
Last edited:
I would regard the room correction functionality as just icing on the cake. The ability to deliver your preferred frequency response curve at your listening position AND with perfect time alignment / phase correction is just so cool.
Agreed. But i can do the same with my dsp too (not the room correction thing obvoiusly!- it's not an entry level dsp though). It just takes a bit more time and planing. 😉
I'm out of my depth when it comes to any kind of full-range global filter....😡
Yes, we can implement IIR through rePhase...so long as we remember remember it may be IIR, but it is limited to the frequency resolution of the FIR filter.
It is always choices made by humans, but the routine can do what you want it to do. It is not either/or, that's a bit misleading.
Like saying humans are better than hammers. Humans use hammers as tools.
Jan
I think you may need to review what you were taught at school about what is a valid scientific hypothesis, what is not and the role of invalid scientific hypotheses in science and technology. Listening can be a form of measurements as can subjective impressions and both have a valid role to play in the scientific process. Digital sheen doesn't.
By learning what a scientific hypothesis is a person has done due diligence. It is those that put forward "digital sheen" and similar that haven't or are being disingenuous in order to serve their interests.
When I started retaking an interest in home audio after dropping it for 25 years when audiophile nonsense first started being promoted in the mainstream rather than being ignored (it had always been around in the small ads) I assumed that everyone pushing it were rogues rather than fools. I steadily changed my mind as I came to see how the audiophile phenomenon worked and what audiophiles seemed to get from believing in fairy stories. The same way of reasoning is unquestionably causing significant harm in more important areas of life but it is hard to see much if any harm when it comes to luxury goods and listening to music.
If you like the idea of digital sheen then go for it. If it is only of significance in the midrange then good stuff. Don't look to science and engineering for support though because it won't be forthcoming unless you follow their rules. Look to audiophiles and their rules and it might be fun.
But... global filter full range it's not ( only!): it implement multichanel! And allow switchng between banks of multi chanel files...
Check it Mark, i'm sure you'll quickly see the potential it offer.
Yes but the routine is a 'generic' approach and as such if your loudspeakers or room are atypical one way or another it might not enter the generic approach a routine offer.
And even if AI driven routine lack subtility in my view.
But i'm not against it as long as you can have access to manual mode.
Too true. On the subject of analogue versus digital crossovers/processing, I would put good money on at least 95 out of 100 listeners not been able to identify which was which in blind tests. Subjective 'data' from 'experts' is no more useful than me telling somebody that red wine tastes better than white, or that a Porsche is better than a Ferrari. It is, however, what peddlers of audiophilia prey on with their emotive and unscientific verbiage. (There's clearly very little profit in room treatment as this critical part of the reproduction chain barely ever gets a mention...).
I use DSP entirely as it's cheap, quick and easy to use, and produces results very pleasing to my ear. I grew out of listening to the system decades' ago and now only listen to the music. Happy days!
I don't think this is a reasonable way to express what is going on. Audiophiles place significant value on the stories and intangibles that surround the particular luxury goods they are attracted to. Many are dismissive of the relevant engineering and science or when they are not tend to view it more like a broad opinion rather than something narrow and specific that is unavoidably true. They place little value in it although, obviously, they would like it to endorse whatever they are attracted to. This thread is an example of this.
Those marketing goods with significant audiophile appeal are usually well aware of what they are selling and what their audiophile customers value and act to add this value for them. Most customers of strongly audiophile hardware are happy with their purchases and go on to buy more audiophile hardware with which they are also happy. Not all but most. Given it does no harm (i.e. it is bought with the disposable income of the relatively rich) it is hard to see it as anything other than fine with all parties involved being happy. OK it looks a bit silly and poor value to those with technical knowledge and an awareness of what is going on but there are other manufacturers addressing what they value. So that's fine as well.
I am afraid that many of the salesmen of room treatment devices to audiophiles have a pretty low standing among those with a formal acoustical engineering background. Then again some people are happy with their purchases though I see a fairly significant proportion of posts from those that have not had their expectations met. It's an interesting area in having one foot in the closed audiophile world and one foot in the open acoustical engineering world.
I discovered to my chagrin a year or so ago that my room could have been a lot better with some treatment, most if not all of the treatment needed is now in place. I will continue to experiment.
I wish I could just listen to the music, but the system is an integral part of the experience for me, and a moving target. (Moving much slower these days - DSP made that possible.)
Thx Krivium,
yes, i've looked at Mitch's Hang Loose Convolver...very nice piece of work. Wish I had it back when I was trying JRiver for multi-channel convolution.
Correct me as needed, but as I remember it is essentially a filter switching program, allowing seamless changes between filter banks. And that it doesn't generate any FIR files itself, but using files made elsewhere. Files made with Acourate, Audiolens, rePhase, FirDesigner, maybe ever REW now, etc.
In my mind, i see maybe three main items to think about with DSP convolution implementations:
- generating the FIR files
- convolvolution processing and routing signal flow to DACs, which includes switching capabilities like in HLC
- and then the DACs and amp channels
So for me, whether it is a dedicated DSP hardware or a PC really doesn't matter, when it comes to how the filters were made.
The first step, filter generation, is the same for either...some FIR file software.
Whether the second step is a PC or dedicated processor. Choosing just comes down to cost, flexibility, and capability imo. And incorporates hardware either way.
The third step is pure hardware i think, whether the DACs are outboard from a PC, or embedded in a dedicated processor.
Just trying to explain my view that programs like Acourate or Audiolens are not really about hardware DSP vs software DSP.
I mean, if my understanding is correct, I can use Acourate or AudioLens generated files in my Q-Sys hardware.
Are you sure about that? I haven't seen such deviations when measuring the output of the 2x4HD.
Here is the a 50 Hz LR4 LPF from the output of the miniDSP 2x4HD compared to an ideal filter.
And here are the responses divided by each other in REW. I don't see any deviation from ideal.
Michael
Hi all,
the discussion has been conducted many times in different audio forums. One of the main cost saving design flaws i see with DSP solutions is here:
unfortunately i did not find the second picture with a decent english translation; the main idea is to reduce the volume for the power amplifiers after the digital to analog conversion and not already in the DSP output stage before the digital to analog conversion
the Audaphon engineer approach has been abandoned some years ago for cost reasons
the easy to test version of this is to use an old AVR with 6.1 CH or 7.1 CH input for the activation of a speaker test let's say with a quite simple DSP board like the 2X4 HD of miniDSP and to try to find out if there is an audible difference between the DSP driven volume control and the analog driven volume control of the AVR used as multi channel power amplifier with analog volume control
especially with high efficiency speaker you will reduce the volume for a standard listening SPL level about more than 6o dB; it is an easy calculation what remain from your 24 bit DAC resolution when the DSP do the volume control digitally, it might come to audible quantization distortion at low listening levels
in a public address insonation situation this will be no issue, you want to have a good sound only at SPL levels only slightly below max SPL
there are many more aspects to consider, but this is an example where the digital processing could make some audible effects
many threads can be found about the topic in the Audio Science Review forum, but you need a lot of time to get through this jungle of different opinons
my two cents worth for this topic, hope it helps
the discussion has been conducted many times in different audio forums. One of the main cost saving design flaws i see with DSP solutions is here:
unfortunately i did not find the second picture with a decent english translation; the main idea is to reduce the volume for the power amplifiers after the digital to analog conversion and not already in the DSP output stage before the digital to analog conversion
the Audaphon engineer approach has been abandoned some years ago for cost reasons
the easy to test version of this is to use an old AVR with 6.1 CH or 7.1 CH input for the activation of a speaker test let's say with a quite simple DSP board like the 2X4 HD of miniDSP and to try to find out if there is an audible difference between the DSP driven volume control and the analog driven volume control of the AVR used as multi channel power amplifier with analog volume control
especially with high efficiency speaker you will reduce the volume for a standard listening SPL level about more than 6o dB; it is an easy calculation what remain from your 24 bit DAC resolution when the DSP do the volume control digitally, it might come to audible quantization distortion at low listening levels
in a public address insonation situation this will be no issue, you want to have a good sound only at SPL levels only slightly below max SPL
there are many more aspects to consider, but this is an example where the digital processing could make some audible effects
many threads can be found about the topic in the Audio Science Review forum, but you need a lot of time to get through this jungle of different opinons
my two cents worth for this topic, hope it helps
Last edited:
I want to make it clear that I am not married to the idea of digital sheen, and a lot of what people have said here indicate that it's probably nonsense. That said, while the scientific method can be applied to most issues, I don't think it's realistic to expect everyone to use it constantly. I heard people talking about this inscrutable phenomena, so I asked about it here on DIYA, and quickly got schooled. Lesson learned! Thanks for your responses Andy
That's misleading. The routine is set up to collect information about the room and your system and make it so that the reproduction is flat at the listening position(s) or whatever you want the target curve to be
(Bob Katz for instance recommends flat to 1kHz then roll of to -6dB at 20kHz. But it's your choice).
You are in the driver seat, and if you f*ck up, yes, results are abysmal.
Jan
The scientific method it is rather limited in what it can be applied to which is why it has nothing to say directly about digital sheen or many other audiophile notions (but the fact it has nothing to say directly says something important indirectly!). Most people obviously lack the relevant engineering and scientific knowledge to reason like a scientist or engineer but they have the capacity to understand what the scientific method actually is, what it can and cannot address and what it means when it can't be applied to things that may appear addressable like digital sheen.
That result looks perfect. Is that the latest firmware version? I am running the old firmware. Thanks for posting that. I knew I would regret not showing a specific example in my post. I usually record all of my test results in documents, but I have not been able to find anything in my documents folder. I often use the advanced tab and load the filter coefficients in from the miniDSP spreadsheet. It is possible some error in my cut and paste of coefficients was the source of the errors I was seeing. I will have to repeat my measurements. I don't have access to my equipment this week, so it will be a few days before I can do it.
It is firmware 1.5 which was released in February with the device console.
I've beta tested firmware 1.6 and it makes some improvements to the ASRC gain structure and minimizes some of the rising noise floor issues with lower frequency filters. Unfortunately, I tested it a few months ago and it is a shame they haven't released it. Some measurement/discussion of firmware 1.6 can be found here -> https://www.audiosciencereview.com/...nts-and-rising-noise-floor.42383/post-1530158.
Michael