Measurements in JRiver? Do tell.
Thanks @J.D. I am also playing with the parametric filters in the 8Cs to get rid of a couple of room resonances at 33 and 57 hz. Itās very effective at tackling these types of issues and it does make a difference although the sound was great even before this. You do have to do a bunch of playing around with the Q factor to get it right it seems. Iām also going to let Thierry from Home Audio Fidelity build me a HAF convolution filter and then Iāll compare the two approaches.
If the Stereophile 48kHz DSP claim is true (and I have no reason to doubt their measurements), youāre better off using Roon to resample everything to 48kHz instead. Having two stages of resampling is worse than only one.
Completely agree @wklie. I read elsewhere that the dsp operates at 96K which is why Iām sampling to 96K. My dealer is verifying with d&d what they are doing with their dsp circuitry. I only want to resample once.
Good day Mr Kal !
The wine did indeed confuse some things.
What my friend and I did step by step:
- We used a program named Accurate. Apparently better then REW, ill try to find some information.
- Downloaded jRiver (free trial, I am going to use ROON in the future when the D&D8C get the support)
- Measured the room and speakers to see all the problems iĀ“ve got .
- Made a digital filter in Accurate. the filter takes care of phase, time compensation among other things (over my understanding, thats what friends are for )
- Went to jRiver and imported the digital filter to its Convolution Engine and voila , thats it.
Accurate made a neutral filter. Removing problems and time compensate. Thats step one, step to is to tweak the filter to my preferrings. In my case a little more dominant mid-range, but more of that when ROON support is here. For now, the PEQ works just fine.
Here is the software we used if you want to read
I also agree !
To be honest, i couldnĀ“t care less about the internal sample rate. Believe me when i say that these speakers takes the listening experience (domestic use) to another level.
But indeed, it would be much better to do the down sampling in ROON to match the sample rate of the speakersā¦ what ever the sampling rate is
Would be great to get some clarification from D&D.
Sounds great!
Yes, both me and my friend was surprised about how much the PEQ made of a difference. Usually, PEQ is great, but we did not expect it to do this small miracle.
The built in DSP does a great job. But owners that calibrate with PEQ unlocks another level of sound with the D&D8C.
Owners that makes a proper digital filter unlocks yet another level of sound. I donĀ“t have the knowledge of making 100% correct filters, thats why my friend who has done this thousands of times helps me.
Doing a faulty filter can really destroy the whole thing haha
When ROON comes i will special order a ROON Core or buy an Innuos again. I had the Innuos Statement before, but ROON support took to long so i had to return itā¦ Ill will also special order some cat cables , then make a Digital filter with all the components in the chain.
That makes more sense. Thanks.
That probably depends on what you were listening to previously or does it apply to every set up?
One thing I have leaned while measuring performance in REW for room correction, is that before creating any filters you should measure while changing the distance from front wall and side wall settings in the lanspeaker.com app. In my room, the speakers are about 37cm or so from the front and side wall so I had set the distance for both to 40cm which is the nearest value you can set. Today I compared the frequency response with the speakers set to 30cm, 40cm and 50cm and while 50cm was notably worse, 30cm was quite a bit better than 40cm and made a couple of likely room resonance issues mostly go away.
That is a great suggestion.
Of course not.
Only the setups Iāve had myself, priced many times the 8Cs.
Iāve been pushing my dealer (Vintage Kings in LA) to get an answer from D&D on what frequency the 8Cās DSP runs at to verify that 48khz is correct and below is the answer that came back:
āWe consider the rate at which the DSP runs as engineering information that is irrelevant to the user. We prefer not to talk about it. The reason is that many people believe that higher is better, which is not true. We chose the frequency that gave us the best audio performance in terms of signal to noise.ā
This is unfortunately an unhelpful answer. I do understand that they donāt want to be pegged by uneducated users that their speakers are only good at 48khz or below or other harmful conclusions that create customer confusion. BUT, their comment that āit is irrelevant to the userā is wrong IMHO.
When you must feed 96khz or below tracks via AES or 192khz or below tracks via ethernet to the 8Cs you will have to downsample anything higher before feeding it to the 8Cs or it will not play. As @wklie correctly noted above, resampling twice is bad (and frankly stupid) so what you want to do is as @wiklie suggests ā have Roon down sample everything above 48khz and upsample everything below 48khz to the 48khz 8C DSP frequency and have the 8Cs take it from there without further resampling.
I pushed back on Vintage Kings to push D&D harder for a helpful answer, weāll see if they are more forthcoming but Iām not going to hold my breath.
Given this stalemate, I will be taking the Stereophile educated guess of 48khz for the DSP frequency and having Roon resample everything to 48khz prior to sending to the 8Cs via AES.
If you havenāt seen them, the Stereophile measurements of the 8C are here: https://www.stereophile.com/content/dutch-dutch-8c-active-loudspeaker-system-measurements
I do want to note that this information roadblock has not dimmed my excitement for the amazing capability of the 8Cs, itās only getting in the way of ensuring Iāve set them up in an optimal fashion and can squeeze every last drop of performance out of themā¦
One last note. Probably obvious, but everything in this post is only relevant when trying to send digital to the 8C. If your sending analog then the 8Cs will of course resampleā¦
Those sound like the words of one of the designers @Martijn_Mensink. I shared a quote above, just a few days ago:
Yes it isā¦
Iāve tagged him here. If heās still not interested in discussing it, you wonāt get the right information from anybody betterā¦ in which case, best to go with the guesstimate from the Stereophile measurementsā¦
Thanks! I agreeā¦
For advanced users in-depth knowledge like this helps configuration.
However, I believe it is the norm in the industry to not talk about how internal DSP is done, especially when it might be negative for marketing. I can think of at least one Roon Ready stereo amplifier that features room correction, or typical AV receivers, are like this. For the stereo amplifier, I donāt think they have disclosed their internal DSP sampling rate either. I also believe every audio company has their own secret sauce, including us.
In our product line we have a low price Roon Ready digital amplifier (in addition to a class AB amplifier). I recall a customer asked how DSD was played. I think he was really not happy learning that DSD needs to be processed to be digitally amplified. Just sharing an example that transparency can be bad for marketing.
The DSP core of the 1452 runs at 294.912 kHz if I read factory specifications correctly.
Also, for a sample rate of 48 kHz the DSP can run up to 6144 instructions per sample
however at 96 kHz sample rate, max instructions per sample is down to 3072
No wonder D&D runs the internal processing at 48 kHz. One would presumably loose filter resolution (x-over and peq) at 96 kHz.
Still, informing advanced users about these facts would certainly be a good thing.
In regards to the question on DSP sample rates, according to Martin, the 8C designer from Dutch and Dutch:
āTheyāve received several questions about this and are going to release a document to respond to these questions.ā
Stay tunedā¦