Reasonably priced good quality 6 or 8 channel USB DAC?

[JustIntonation]

Yeh, that's true. The rockwool is not flush with the walls but separated by an air gap (it's attached to the bracing, actually). We've used rockwool in the prototypes so far, wrapped inside cloth, but one thing I've been meaning to do is talk to an expert in this area to clarify whether the cloth is adequate for containing the rockwool particles when air moves through it. If not, we'll have to switch materials.. Is this something you know a bit about?

Aah yes I have experience with both Rockwool and fibreglass and generally you will not find porous materials that will contain fine Rockwool dust particles under those circumstances. But you can wrap it in thin plastic just as well or a non porous synthetic fabric, this will only affect high tones bass and low mid tones will go through just as well. Not sure exactly where the transition is for different fabrics / plastic wrapping is regarding frequency though. There are several people at Gearslutz who are real experts on this.
But in your application with a port I think most people would suggest using a different material alltogether instead of airtight wrapped Rockwool.

 

[andreasmaaan]

Aah yes I have experience with both Rockwool and fibreglass and generally you will not find porous materials that will contain fine Rockwool dust particles under those circumstances. But you can wrap it in thin plastic just as well or a non porous synthetic fabric, this will only affect high tones bass and low mid tones will go through just as well. Not sure exactly where the transition is for different fabrics / plastic wrapping is regarding frequency though. There are several people at Gearslutz who are real experts on this.
But in your application with a port I think most people would suggest using a different material alltogether instead of airtight wrapped Rockwool.

This is good to know, thanks

PS. @JustIntonation do you know of a non-porous synthetic fabric that would work? I'll of course do further research and testing, but the highest frequency necessary for the wool to work on is around 1250Hz, so the non-porous fabric may do the job. So a nudge in the right direction would be good.

Looks like there will need to be a re-think on the cloth Adequate for the prototypes but obviously not for production units...

 

[JustIntonation]

do you know of a non-porous synthetic fabric that would work? I'll of course do further research and testing, but the highest frequency necessary for the wool to work on is around 1250Hz, so the non-porous fabric may do the job. So a nudge in the right direction would be good.

No couldn't give you a fabric for this. I've used simple food wrapping plastic in addition to porous fabric myself in the past (and will do so again for the cloud I'm going to make) but I wouldn't trust this inside a speaker either (where vibrations are very extreme). I would suggest asking here: https://www.gearslutz.com/board/studio-building-acoustics/
Btw if you're already doing Rockwool in a porous fabric in a ported system now, do you notice very fine dust lining the port if you swipe your finger in the port? I wouldn't want this in my house.. All dust is bad ofcourse and Rockwool isn't carciogenic but it will cause scar tissue build up in the lungs over time.

 

[JustIntonation]

Btw I don't have experience in damping a ported box but perhaps something like Troels does makes most sense here? http://www.troelsgravesen.dk/cabinet-damping.htm
With perhaps a bit thicker "acoustilux" type material? (egg crate foam is nonsense btw)

 

[andreasmaaan]

Thanks again I haven't noticed any particles so far, but this is with quite thick cloth covering the wool. I'll also try the acoustilux and felt type material ultimately.

Totally agree re: egg crate foam.

EDIT: this is also all happening in the workshop where there are lots of particles of various kinds flying about, so it's possible that the rockwool particles are there, but just haven't been apparent for that reason.

 

[March Audio]

I used this and standard polyfill

https://www.parts-express.com/sonic...nd-damping-material-with-psa-18-x-24--260-535

seems to work

 

[JustIntonation]

Received a reply from Okto Research.
Asked if it was ok to share online and it was.

My original question was if it was possible for a custom build of the Okto DAC8 + Okto PSU + MiniDSP USB streamer in a case with a volume control pot on front for the ES9028 digital volume. Also asked if there was perhaps an alternative USB to i2s interface they would recommend over the USBStreamer. The reply was:
"
the unit you described seems very much like the one we are preparing to unveil in about a month . Apart from the USB input (not sure yet if USBStreamer or our own design) it will feature OLED front panel with rotary knob and remote control. We put a lot of effort into the front panel design. I would suggest you to wait for that if you are interested. The price will be roughly around 1000EUR.

Apart from the USBStreamer, the only multichannel USB to I2S module we know about is the one from DIYINHK which we haven't tested.
"

I further asked if they have a linearity measurement of the DAC8 and how the DAC performs at higher sample rates. To which the reply was:
"
We didn't take a linearity measurement yet, but I'll remember to take it as soon I get to the audio analyzer again.

There was basically no performance difference between 44.1kHz an 96kHz sample rate. At 192k, I remember that increase of THD/THD+N observable - I've searched through the measurements we've taken and found one of a 10kHz tone at 192kHz sample rate showing -122.7dB total THD instead of -124 to -130dB achieved at lower rates. Again, I'll measure more at the next opportunity.
"

And they'll keep me informed.

 

[BYRTT]

Please see this file: https://we.tl/t-GiNDRUuPWy
It is part of a test I did. One file (original) is the original music, the other file is a simulated LR16 (96dB/oct) crossover at 2kHz simulated on-axis (so an allpass filter at this angle). The group delay around the crossover freq is 4 times the wavelength of 2kHz.
Of course this is a more extreme example than a LR4 filter. But I cannot tell any difference at all between the two files.
Can you? Note that two songs, one starting around 1:36 and the other around 2:36 have a deep soundstage. I cannot tell any difference in soundstage size or depth between the two files.
If you use ABX for Foobar2000 or something similar you can test this properly and switch back/forth etc.

Can tell a difference using a linear phase XOed 2 way speaker and hints me then my EQed headphone setup in its more direct wavefront relation should reveal same result, that said not shure my mother would notice if asked but for me there is a change in realism as if for example we say microphones recording distance original was say 1 meter then "co-sim original.wav" is sensed sound closer to that and its better resolution to extract realism and timing or line up of harmonics, where for "co-sim LR16 A0.wav" sound is senced that recording microphone is now pulled one or several more meters away and spoils ratio of resolution in a way so now it only sounds nice or fine but without the clues for realism into recording enviroment.

In this MS OneDrive link (https://1drv.ms/u/s!AqpXxWjyVD8ljyD3026mbIpvIkjp) find "co-sim original.wav" convoluted to "co-sim_LR4_at_160Hz_plus_LR4_at_1500Hz.wav".

Below is using Audacity to analyze those three files asking for spectrum into first 267 seconds of each track, then exported to REW to divide them to get their real differences in amplitude domain, green = "co-sim original.wav", red ="co-sim LR16 A0.wav", black = "co-sim_LR4_at_160Hz_plus_LR4_at_1500Hz.wav":

In general must say agree most of what arguments and experience pos and March Audio share here, also have a feeling from own experience that over time into future studies will change a bit on how phase sensitive humans can be.

Regarding phase and natural sound must say is a bit frustrated when surrounding nature enviroment is stated be of minimum phase pass band domain and the natural sound we hear, then how natural is it put one two or three more 4.th order phase shifts into audio domain and declare it perfect as lang as its of LR family, but lets see down the road when system is up running how real time performing filters verse some with more or less latency will be sensed.

Great thanks share alot of good info and concerns, about nuke the zero phase concern at nyquist can share it works pretty good run DSP up at 192kHz which at same give possibility to EQ and form tweeter roll off at HF to follow a more smooth textbook slope than inherent break up curve, subjective that feature is sensed up a bit on resolution plus 3d and rhythmic lineup of harmonics, drawback but alright compromise for my case is it cost one analog set of DAC/ADC interfacing because most track material is at lower rates and don't sound right forced to playback up at 192kHz.

 

[JustIntonation]

Can tell a difference using a linear phase XOed 2 way speaker and hints me then my EQed headphone setup in its more direct wavefront relation should reveal same result, that said not shure my mother would notice if asked but for me there is a change in realism as if for example we say microphones recording distance original was say 1 meter then "co-sim original.wav" is sensed sound closer to that and its better resolution to extract realism and timing or line up of harmonics, where for "co-sim LR16 A0.wav" sound is senced that recording microphone is now pulled one or several more meters away and spoils ratio of resolution in a way so now it only sounds nice or fine but without the clues for realism into recording enviroment.

In this MS OneDrive link (https://1drv.ms/u/s!AqpXxWjyVD8ljyD3026mbIpvIkjp) find "co-sim original.wav" convoluted to "co-sim_LR4_at_160Hz_plus_LR4_at_1500Hz.wav".

Below is using Audacity to analyze those three files asking for spectrum into first 267 seconds of each track, then exported to REW to divide them to get their real differences in amplitude domain, green = "co-sim original.wav", red ="co-sim LR16 A0.wav", black = "co-sim_LR4_at_160Hz_plus_LR4_at_1500Hz.wav":
View attachment 17268

In general must say agree most of what arguments and experience pos and March Audio share here, also have a feeling from own experience that over time into future studies will change a bit on how phase sensitive humans can be.

Regarding phase and natural sound must say is a bit frustrated when surrounding nature enviroment is stated be of minimum phase pass band domain and the natural sound we hear, then how natural is it put one two or three more 4.th order phase shifts into audio domain and declare it perfect as lang as its of LR family, but lets see down the road when system is up running how real time performing filters verse some with more or less latency will be sensed.

Great thanks share alot of good info and concerns, about nuke the zero phase concern at nyquist can share it works pretty good run DSP up at 192kHz which at same give possibility to EQ and form tweeter roll off at HF to follow a more smooth textbook slope than inherent break up curve, subjective that feature is sensed up a bit on resolution plus 3d and rhythmic lineup of harmonics, drawback but alright compromise for my case is it cost one analog set of DAC/ADC interfacing because most track material is at lower rates and don't sound right forced to playback up at 192kHz.

Good first post

Can you tell me how to do the amplitude comparison you did in Audacity? I should have done this myself.
If I'm reading the graph correctly it seems to show that the curves of my LR16 filters are not correct (edit: or.. their phase is not correct minimal phase as the plugin promises, or.. perhaps it is due to for instance a one sample auto latency compensation error of the host since I was running two different plugins one for HP one for LP) I set them up in Fabfilter ProQ2 but apparently they don't produce a flat line in amplitude as they should. I checked them only by ear.. stupid me. If I'm reading the graph correctly there are errors up to almost 0.4dB in the result..

Thanks!!

 

[JustIntonation]

Ok I found the error. It was entirely my fault, I set up ProQ2 wrongly by taking a stupid shortcut and didn't create a correct Linkwitz-Riley
Sorry, and thanks for catching my mistake!

Here are the correct files and they now measure ruler flat other than the LR minimal phase group delay occuring at the crossover.
https://we.tl/t-0q8Ex8EDzz

If you compare them again to the original is there still any difference audible?
If so please do it with ABX comparator for instance for Foobar2000. https://www.foobar2000.org/components/view/foo_abx
I can't personally tell the difference, not even with the previous one with the 0.4dB amplitude dips error.. Perhaps after I've built my new speakers I can, will test again then. (and will test for minimal phase group delay with an impulse then, this audio file isn't really designed to be critical of group delay..)

 

[BYRTT]

My Audacity is v2.1.1 and in Danish language so will try translate the best i can to english : ) in top meny pick "analyzer" then "show spectrum" and from that window pick dialog "export". Not shure we can use curves for analyze any high precision result other as a broad indication , because some interpollation and non terminated tails at stopbands probably happen into process or what do i know, also into REW "Psychoacoustic" smoothing was used in that visual graph to help kill the grass look.

Have not visual analysed the new files but in a fast listening test result is subjective the same that original have far the best realism but admit this is total non controlled without using some ABX comparator and think in general that listening for phase only changes can be a bit hard on used energy trying to focus on differences where amplitude changes often is a bit easyer and sometimes a knight or day difference. On my replay system in the third track differences stands clear out on natural attack of bass guitar plus snare drum and symbols is more real but admit its hard tell a difference between the other two LR4 and LR16, its just that original is better travel into recording room enviroment especially in the long run over some time when one is relaxed and walk around it can be noticed.

If it means anything then for info player is JRiver with volume leveling/adaptive volume/ISO226 equal loudness enabled so tracks are manual way analyzed before listening session.

Guess down the road when setting up those acoustic slopes it doesn't mean so much if one or another used software package has minor errors for slopes because real measurement will show and one then continue investigating how target and textbook stuff should look up inside what rate chain is running at and also have a eye on best calibration numbers for chain.

 

[JustIntonation]

If it means anything then for info player is JRiver with volume leveling/adaptive volume/ISO226 equal loudness enabled so tracks are manual way analyzed before listening session.

Aah.. well this does say something.
If you look at the files you'll see that the original has the tracks peak limited from mastering. This is very much undone on the LR4 and LR16 files since they are filtered and group delay occurs they have new peaks much higher than the original. If volume leveling/adaptive volume is applied then the "original" track and the LR16 LR4 tracks will no longer play at the exact same volume which surely makes a much bigger difference than any possible audibility of the group delay.

 

[BYRTT]

Here are JRivers analyzed number and think differences look abnormal minimal compared to for example how album to album differences look, sound wise on a dayly use think its good features to enable and especially a calibrated equal loundness filter is sensed give great psycoacoustic result.

 

[JustIntonation]

Here are JRivers analyzed number and think differences look abnormal minimal compared to for example how album to album differences look, sound wise on a dayly use think its good features to enable and especially a calibrated equal loundness filter is sensed give great psycoacoustic result.

View attachment 17305

Indeed, volume level (replay gain) doesn't matter as much as I expected.
Probably due to the last song which isn't compressed and is probably loudest as there is a big difference in peak level for the first few songs.

 

[JustIntonation]

For those interested, here an ideal crossover simulation comparison between the off-axis cancellation axis of LR4 (24db/oct) and LR16 (96dB/oct) at 1500Hz and 3300Hz.
All are minimal phase except for the "linphase" version I also included for LR16 3300Hz.
https://we.tl/t-xAnbvuzFBQ (280MB)

The LR16 versions are less bad, but since this is the cancellation axis and what's most important (at least in my case) is the on-axis further tests with an impulse response will possibly favor the LR4 due to less group delay (perhaps LR6 is an option as well with some minor benefits I won't discuss now).
Further, I'm not hearing much of a difference between the minimal phase and linear phase versions. But I'm not on a great soundsystem at the moment as I've sold off things to fund my speaker build. But the biggest difference seems to be latency for the linear phase version as far as the cancellation axis sound is concerned.

Now for what's the most important thing for me.
The sound difference between 1500Hz crossover and 3300Hz. This shows me 1500Hz is indeed much much better for a crossover frequency between mid and tweeter (if the tweeter can handle it of course).
Not only will it be much easier to get a very large vertical window without cancellation axis due to the lower frequency and more reasonable driver distance requirements for this. It also makes it easier to use a mid driver which doesn't fall off rapidly off-axis around the crossover freq (something not in the simulation but which also makes a crossover audible possibly well before the cancellation axis is reached).
And most importantly, for any remaining crossover effect the sound is just so much better for 1500Hz! It takes some longer listening (comparing LR4 1500Hz and 3300Hz) to the effect, and of course both crossover frequencies are bad at their cancellation axis, but that 3300Hz crossover phasing and ringing is exactly the sound I truly HATE. It messes with my understanding of the music, makes it small in a way and it has harshness. 1500Hz is bad too in its own ways and you still loose a lot at the cancellation axis but it doesn't touch the most important 2-5kHz range and is to me much more musical as a result. I wonder if this is only the equal loudness curve in effect or if there are other psychoacoustical things at play here for the differences? I suspect also the latter.

Btw, for bass to mid woofer the crossover is less critical as for a 150Hz freq with the drivers fairly close the cancellation axis is never reached (it's on a virtual plane well beyond the baffle) so the drivers behave like a dual concentric / coaxial arrangement.

 

[andreasmaaan]

@JustIntonation FWIW this is the next material I'll experiment with, which is made of recycled cotton and which appears to have reasonably good absorption in the desired frequency range, especially given I'll be able to line one 50mm sheet on either side of the bracing to give a 100mm total width with a small internal air (and wood) gap. This should significantly improve bass absorption compared to a single 50mm sheet (I speculate/hope).

 

[eugenius]

The amp power for the mids seems reasonable.
http://www.sbacoustics.com/index.ph...7nbac35-4-br-font-color-c70039-shipping-font/
60W rated power, but I don't want to push them to their limit anyhow.

Since you paid a fortune for the Bliesma tweeter to enable a lower crossover frequency on the tweeter, I suggest increasing your midwoofer size and quality to get better low mid dynamics and lower distortion.

From SB Acoustics one of these MW19P-8 drivers looks ideal for a two way.
http://www.sbacoustics.com/index.php/products/midwoofers/satori/7-satori-mw19p-8/
Too bad they don't have a midrange version like they do for the 6" midwoofer.

For a three way mid, I suggest looking at the Acoustic Elegance TD8M:
http://aespeakers.com/shop/td/td8m/

If you want to be unreasonable about enclosure quality, I suggest building your bass enclosure in a rigid B&W Matrix-like structure and damping your midrange enclosure with Hawaphon.
https://www.korff.ch/en/hawaphon-eng/

IF you want to be really autistic, I suggest lining your listening room walls with Hawaphon as well, especially if you live in the city.

 

[eugenius]

Received a reply from Okto Research.
Asked if it was ok to share online and it was.

My original question was if it was possible for a custom build of the Okto DAC8 + Okto PSU + MiniDSP USB streamer in a case with a volume control pot on front for the ES9028 digital volume. Also asked if there was perhaps an alternative USB to i2s interface they would recommend over the USBStreamer. The reply was:
"
the unit you described seems very much like the one we are preparing to unveil in about a month . Apart from the USB input (not sure yet if USBStreamer or our own design) it will feature OLED front panel with rotary knob and remote control. We put a lot of effort into the front panel design. I would suggest you to wait for that if you are interested. The price will be roughly around 1000EUR.

Apart from the USBStreamer, the only multichannel USB to I2S module we know about is the one from DIYINHK which we haven't tested.
"

I further asked if they have a linearity measurement of the DAC8 and how the DAC performs at higher sample rates. To which the reply was:
"
We didn't take a linearity measurement yet, but I'll remember to take it as soon I get to the audio analyzer again.

There was basically no performance difference between 44.1kHz an 96kHz sample rate. At 192k, I remember that increase of THD/THD+N observable - I've searched through the measurements we've taken and found one of a 10kHz tone at 192kHz sample rate showing -122.7dB total THD instead of -124 to -130dB achieved at lower rates. Again, I'll measure more at the next opportunity.
"

And they'll keep me informed.

Did they mention if that box has AES inputs as well? Or at least a DAC only version with AES inputs?

 

[JustIntonation]

Since you paid a fortune for the Bliesma tweeter to enable a lower crossover frequency on the tweeter, I suggest increasing your midwoofer size and quality to get better low mid dynamics and lower distortion.

From SB Acoustics one of these MW19P-8 drivers looks ideal for a two way.
http://www.sbacoustics.com/index.php/products/midwoofers/satori/7-satori-mw19p-8/
Too bad they don't have a midrange version like they do for the 6" midwoofer.

For a three way mid, I suggest looking at the Acoustic Elegance TD8M:
http://aespeakers.com/shop/td/td8m/

If you want to be unreasonable about enclosure quality, I suggest building your bass enclosure in a rigid B&W Matrix-like structure and damping your midrange enclosure with Hawaphon.
https://www.korff.ch/en/hawaphon-eng/

IF you want to be really autistic, I suggest lining your listening room walls with Hawaphon as well, especially if you live in the city.

Thanks for the suggestions.
But as far as I can see now the SB Acoustics SB17NAC35-4 (and SB17BNAC35-4) midwoofer seems to be of top quality in a 3-way?
I'm thinking I can take it to 120 to 150Hz LR4.
The reasons I chose it are because it has a very high cone breakup of around 10kHz well outside the most sensitive area of our ears so any residual distortion falling on it are not going to be audible, also the ringing isn't that bad. Compared to for instance a Seas W18EX001 which breaks up around 5kHz which is much more audible and closer to the crossover freq of ~1500Hz and it rings much more severe I think amplified HD will be more audible there.
As far as I can tell the good aluminium coned midwoofers are a fair bit tighter in time / CSD than the best paper cones who always have very minor breakup in their passband range visible in CSD's. Also the harmonic distortion of the SB17NAC35-4 is the best in the entire range of SB Acoustics. The Satori range does not give better performance there. There's no getting lower distortion anywhere for 200Hz up, I looked very carefully at this also at much more expensive 8" drivers (and even some 10" drivers).
As for dynamics in the low mid range. Yes it can be a bit better in the 120-200Hz range with some 8" or 10" drivers but I'm actually not going to use them very loud as I'll be using them nearfield. Also if I were to do a 150Hz crossover it's down 6dB at 150Hz and almost 3dB I think at 200Hz further helping. In any case for my use the distortion is still very low in this area and I expect no compression at my volume needs, and if need be when with a 150Hz crossover I can play the mid driver over the entire range I use it at almost 120dB SPL volume at 1m (both speakers at once and in linear coil travel of the drivers, even higher SPL for maximum coil travel) that's louder than I need but it indicates the reserve I have even in the lower mids. So all in all it seems like the best driver for my project despite its low cost.
Here some measurements of it and the related (but slightly less well measuring) 8-ohm version:
http://www.audioexcite.com/?page_id=5833
http://www.troelsgravesen.dk/SBAcoustics-61-NAC.htm
https://hificompass.com/en/speakers/measurements/sbacoustics/sb-acoustics-sb17nbac35-8
I also have 2 measurement sets of it in hobby hifi magazines which I can't show here but are equally well.
A two way with MW19P-8 as done by hificompass https://hificompass.com/en/projects/2-way-systems/pharaoh measures worse, my speakers should be better in many ways also in the mids. (also due to better baffle, better DSP crossover, better dead cabinet and due to it being closed 3-way also better in the bass). (oh and hificompass even reported problems themselves in audibility of mid-tweeter differences and trying different crossovers to lessen this)
edit: oh yes and one more reason to use a 6" driver is that the mid and tweeter can be closer physically giving a near coaxial/point source radiation / no lobing with the low crossover freq. This is less so with an 8" driver. Also the SB17NAC35-4 has almost no freq drop around 1500Hz off axis while many other 6" drivers and all 8" and larger drivers have significant off axis dropoff at that freq further degrading off-axis performance. I think overall the SB17NAC35-4 is really the perfect match for the T34B-4 with 1500Hz LR4 DSP crossover. There shouldn't really be an audible error in any area I can see with my knowledge so far.

The Hawaphon hmm I'm not that impressed actually.. I believe 5mm leadbitumen or a mass plate gives about the same result? I'll be using 19mm MDF + 19mm birch plywood + bracing and then 5mm leadbitumen on both the inside and outside of the speakers (won't be the prettiest but I don't care at all about that and decided to do the inside stuffing with fibreglass (cheaper than wool with at least equal results but for the fibreglass I know the exact gass flow resistance value and don't know it for wool)
And about my room not too worried about the neighbours. Only have one neighbour who I could possibly bother but it's hard to do so, live in an old city building with massive walls and the neighbour is gone during the day. Also my walls are isolated and I'll be adding 27cm to 45cm thick Knauf Acoustifit to the walls for creating my near anechoic space (only bass won't be anechoic but where not absorbed it can either escape into a larger space or when it is reflected very close to the speakers it will be in phase)

But thanks for the suggestions again though! This was about the last moment for changing my mind on things, starting building the studio next week and the speakers right after

 

[JustIntonation]

Did they mention if that box has AES inputs as well? Or at least a DAC only version with AES inputs?

No, no mention of this.
But the designer seemed to be open to custom modifications, perhaps a box with AES inputs can be made?