The DSP Rolling & How-To Thread

[castleofargh]

old is 15 years older than we are. it's a universal constant for the person thinking about it.

 

[gregorio]

[1] Why do you think that miniDSP implementation is better? Correcting "time smearing" is marketing bull****, you cannot directly control the duration of a bass resonance with active room correction (i.e. DSP). [2] The best such active (digital) RC can do it to reduce the amplitude (level) of a resonating peak. Of course, if you reduce the level at the frequency which resonates in your room, this resonating frequency will die faster, but it's indirect control. [3] If you want to directly control the resonance duration, you need not active RC, but passive RC (diffusion and absorption panels, bass traps, SBIR and RFZ panels, etc.). Most people are, of course, to lazy to make, buy and install such panels.

1. As I stated to start with, DSP RC does not "correct" a room, it does not cure room modes/standing waves/resonances/phase cancellations, it just somewhat improves them.
2. No, that is NOT the best digital RC can do! As I already stated, that's the best EQ can do but most/all of today's digital room correctors do not just use EQ, they also use timing/phase correction algorithms. While this has limited effect on bass (and other frequency) resonance in terms of duration, it can significantly help with the arrival time of transients at the listening position (LP). This introduces another problem, you can somewhat improve the situation at the LP but typically at the expense of other LPs, sometimes just a few inches away from that LP can make a significant difference and no one wants to stick their head in a fixed head-clamp to ensure their head is always in that one "corrected" LP. In theory, a solution to this problem would be to take multiple measurements in different LPs and then simply average them. In practise though, that solution is effectively a compromise between each of those measured LPs and is quite likely to make the situation worse or at least no better in some, most or even all of those different LPs. Some digital RC systems take a more sophisticated approach, employing quite complex algorithms to apply each individual, identified correction at an amount and in a way which is not too detrimental to any one LP but is somewhat beneficial to all. So, correcting the "time smearing" of transients is NOT marketing BS, although again, I would say "somewhat improving" rather than "correcting".
3. The best solution (although still not perfect), is three simultaneous approaches! 1. Room construction, IE. Building a room to start with which minimises acoustic issues, no/few parallel reflective surfaces, suspended floors, etc. 2. Room treatment, IE. Appropriate absorption and/or diffusion. 3. Signal processing/modification. For many decades the use of an EQ (typically a 31 band graphic EQ) has been completely standard practise in studios, to help lessen/even out the issues left after room construction and treatment have done their bit. Although the general rule of thumb has always been to use EQ to address no more than about 10% of acoustic issues.

These days, advanced digital RC is being used in more and more studios but only as a replacement for #3, NOT as a replacement for #1 or #2! As far as serious consumers/audiophiles are concerned, very few can do much to address #1 but most could at least do something to address #2 (to a level which won't upset the wife) but generally don't, instead seeming to prefer focusing their attention purely on the reproduction equipment rather than on the sound waves actually reaching their ears.

G

 

[WoodyLuvr]

If you play everything through a media server and it's just 2 channel, I don't see any advantage to not using a software EQ through the computer. If you have a multichannel or Atmos system, or if you have multiple sources in your system, the computer becomes less convenient.

Very well said.

And whether or not to install panels in a room depends on the materials used to make the room and the particular acoustics. In my case, the whole room is 1950s knotty pine. I can't cover that up with acoustic panels. It would look terrible. Every situation involves tradeoffs and compromises.

You could take some charcoals and turn those panels into works of art that would conversation pieces for years to come... there, fixed!

 

[WoodyLuvr]

Well, I am 42 years old, so I may be "old folks" myself to some young people!

old is 15 years older than we are. it's a universal constant for the person thinking about it.

Concur completely! Since when was early 40s old?! That be a young'in still in my book. Knows just enough to get themselves into trouble. Experience doesn't begin until 50s

 

[WoodyLuvr]

@Strangelove424 Trying to kick some life into this thread again. Don't forget to add these two crossfeeds I mentioned to you before: Foo VST xfeed and 112 dB Redline Monitor

 

[Strangelove424]

Sorry for the slump in progress. I was distracted, especially during the holidays, and neglected this thread a bit. I added xfeed and Redline 112db to the OP, and will be more attentive about updating.

 

[PleasantSounds]

It's been a while indeed - I don't even remember if we have mentioned resamplers at all? I don't see any in the 1st post...
I recently stumbled across a really great one: Resampler-V. It deploys two of the best resampling algorithms: SoX and SSRC, and provides visual representation of the parameters chosen. It will be hard to beat!

 

[castleofargh]

I thought there was a problem from the start with this topic, and it's becoming more and more obvious that it needs to be centered around something way more specific. otherwise you're pretty much asking @Strangelove424 to list all the plugins available in a DAW. well at the very least, all the VSTs to use in foobar or a VST host, which adds the possibility for plugins incompatible with foobar to be used anyway(GUI reasons, or VST3 and 64bit versions of the plugin), meaning hundreds of plugins.

 

[Strangelove424]

I don't really see it as a problem, as many different ideas have come up due to the lack of specificity. It's true, listing every available plugin is impossible, but it's nice to have a wide selection available to browse.

Regarding resampling DSP, I have never heard a difference after resampling. I know resampling is important for editing or distribution purposes, but I can't see it making an effect on the consumer listening end. I was hoping to stay focused on DSPs that make a very clearly defined difference to the sound, not just the format or rate.

 

[castleofargh]

I don't really see it as a problem, as many different ideas have come up due to the lack of specificity. It's true, listing every available plugin is impossible, but it's nice to have a wide selection available to browse.

Regarding resampling DSP, I have never heard a difference after resampling. I know resampling is important for editing or distribution purposes, but I can't see it making an effect on the consumer listening end. I was hoping to stay focused on DSPs that make a very clearly defined difference to the sound, not just the format or rate.

you're the captain of the thread. if you're ok with too much work, who am I to complain ^_^.
about resampling, it will impact some gears. most antiquated NOS DACs are likely to suck at 44.1. they will roll off too much in the high freqs or be an aliasing fest. so for those it's probably a good idea to oversample. even if it also means admitting that adhering to the NOS DAC principle was a mistake.
then some have reported changes even on other devices, but I wouldn't make any claim about audibility there, as the feedback usually comes from sighted experience.
and some just do it because they assume they have a method that is objectively superior to the one used in the DAC, which is probably true with the right app. but as the chip works at really high sample rate, you will not have the opportunity to directly send that rate to the DAC(unless you're converting to DSD for the appropriate kind of DAC). meaning we do one oversampling, and then the DAC does another one anyway. how beneficial that can be is something I honestly don't know.
I also vaguely remember reading something about situations where the chip deals with 44.1 or 48khz but kind of sucks with the other one. but I can't remember if it was solid stuff or audiophile talk. maybe some stuff about a very crappy clock? or some specific anti jitter upsampling thingy? I don't remember. does anybody know if it is something that exists/existed? (secretly summoning @KeithEmo for information).

but I would assume that the main use of a resampler in foobar is to have a fixed rate no matter the file played so that you can output something consistent through bit perfect, or some specific virtual cables, or maybe even something related with convolution if the filter is only available at one sample rate. so it's not completely useless.

 

[WoodyLuvr]

I don't really see it as a problem, as many different ideas have come up due to the lack of specificity. It's true, listing every available plugin is impossible, but it's nice to have a wide selection available to browse.

Concur... I view this thread as an initial "Jump-in-There" + "How-To" with some recommended samples to try out.

 

[Strangelove424]

you're the captain of the thread. if you're ok with too much work, who am I to complain ^_^.
about resampling, it will impact some gears. most antiquated NOS DACs are likely to suck at 44.1. they will roll off too much in the high freqs or be an aliasing fest. so for those it's probably a good idea to oversample. even if it also means admitting that adhering to the NOS DAC principle was a mistake.
then some have reported changes even on other devices, but I wouldn't make any claim about audibility there, as the feedback usually comes from sighted experience.
and some just do it because they assume they have a method that is objectively superior to the one used in the DAC, which is probably true with the right app. but as the chip works at really high sample rate, you will not have the opportunity to directly send that rate to the DAC(unless you're converting to DSD for the appropriate kind of DAC). meaning we do one oversampling, and then the DAC does another one anyway. how beneficial that can be is something I honestly don't know.
I also vaguely remember reading something about situations where the chip deals with 44.1 or 48khz but kind of sucks with the other one. but I can't remember if it was solid stuff or audiophile talk. maybe some stuff about a very crappy clock? or some specific anti jitter upsampling thingy? I don't remember. does anybody know if it is something that exists/existed? (secretly summoning @KeithEmo for information).

but I would assume that the main use of a resampler in foobar is to have a fixed rate no matter the file played so that you can output something consistent through bit perfect, or some specific virtual cables, or maybe even something related with convolution if the filter is only available at one sample rate. so it's not completely useless.

I don't mind the work if people are patient about delays. I might not always be able to update right away, as I have day to day responsibilities like everyone else, but I will always get around to updating eventually.

Almost every DAC performs differently between different sample rates. It could be related to the software driver, or hardware as well depending on design. The question is if those difference are large enough to be audible, and I'm not convinced they would be aside from really crummy gear, which is very hard to come by nowadays.

 

[PleasantSounds]

[..]
Almost every DAC performs differently between different sample rates. It could be related to the software driver, or hardware as well depending on design. The question is if those difference are large enough to be audible, and I'm not convinced they would be aside from really crummy gear, which is very hard to come by nowadays.

Note that resamplers can be used not only to increase sampling rate but also to lower it. I have some 24/192 and 24/96 albums which perform equally well at 24/48, so to reduce the storage footprint on my server and the network bandwidth my working copies are usually either 16/44.1 or 24/48.
In some software DSP filters (including resamplers) can be used in the conversion path.

Regards

 

[Strangelove424]

It did not occur to me you were using it as a compression for re-encoding files. Though I think that’s an interesting application, it’s aimed at data formatting and storage. I’d really like to keep the thread focused on real time DSPs that alter the sound.

 

[Strangelove424]

Update: Have combed through thread, and believe I captured all the suggested DSPs so far. Added Reverb/Convolution section to OP, and thanks to ironmine for those suggestions.

I've been looking for some free surround sound mixing DSP, hopefully something in VST form. I know I have the Foobar one listed, but I was hoping to find a free VST with this kind interface (this particular one sells for a few hundred)...