There is a good point here that I don't think has been covered. As soon as you manipulate the signal at all you lose any dither information. The question becomes under what circumstances would re-dithering help given usually 24 bit processing off a 16 bit source...
I know my amp is the first to get into trouble. The arrays could handle 240 watt and stay within x-max. One day I will try a bigger one.
The EQ is to counter the drop in response that arrays with these size drivers have. I do use a room curve set to my preferred listening level with a declining trend toward the high frequency. Grown into that curve slowly by listening to a wide variety of music.
This is the impedance as measured (with partial conjugate network):
Amps love my impedance plot 😀...
Below is impedance, above is phase. The blip at 19 KHz is the soundcard of the cheap laptop used to measure it.
Last edited:
What is the calculation precision? If it's 32 or 64 bits (likely) then there's no real effect.
Well that is the question I am trying to formulate. If you are using up to 40dB of digital gain control then at what point would a redither help. I've not worked out where miniDSP does single precision (28bit) and double (56).
Sure! Impressive plots. Approximately 30° phase variation. You can be proud of all the work done.
Tends to agree to most papers and habits (on pro audio) on the subject. I start mine at 1.1khz and drop gently to past 20khz at approximately 1.5db/oct (-0.5db/ 1/3 octave). Tends to mimic natural 'air absorption'. It works well especially with dome tweeters.
Here is the 'target' hifi from Bruel & Kjaer:
http://i1217.photobucket.com/albums/dd381/mitchatola/bandk.jpg
Interesting to compare to PA sound system recommended: grey D&B, red E.V., blue 120hz 4pole LR (sub +15db).
http://hornplans.free.fr/courbe_ideale.gif
Once Fletcher & Munson into account, i find it kind of similar. 🙂
Last edited:
Before worrying about dithering i would about adc.
How do you plan to use your minidsp? With digital in or adc?
Both. Music off server in digitally. Music off turntable in via ADC and possibly RIAA done digitally now that Scott has done all the hard work on that and published the results and biquads.
Digital Riaa compensation curve have great potential if well executed (done some try with Prism converter offering the same option: sound great). You'll have to be carefull about input level at adc. You'll probably need some preamp here to fully use dynamic range of adc, or results could be not so great. Typically vinyl have 20db usable crest factor (depending on music style). Use high quality analog preamp if possible!
Once I have finished it the balanced amplifier for the cartridge shouldn't be too shabby at all. Ticks and pops are a concern but its only a 12 bit system so plenty of headroom to play wtih.
you would say 24bit? If you leave 6db/12db max headroom (-6dbfs/-12dbfs) you should be safe about surface noises. When using adc 24bit, it's good practice to leave 6db min headroom as you don't compromise dynamic to much and keep you away from nasty adc overload.
Note that 12bit sound is not so problematic after all! I listened to some 80's early 90's hip hop lately and most of them used Akai S900 samplers. It's this machine which give the 'crunch' on drum samples:12bit sound and restricted sampling freq.
Last edited:
Vinyl SNR is such that you only need 12 bits to capture it. The surface noise self dithers nicely. Ticks are wideband and can be very significantly above the program level. Taking both those into account backing a 24bit ADC off 40dB is not madness.
It's a rock being dragged through plastic after all!
It's a rock being dragged through plastic after all!
Number you give are a bit generous comparing to what i usually do and see. I don't agree about the 12bit to capture, i could agree that crest factor is in more or less this range but even if this is a rock dragged through plastic when well recorded mixed mastered there can be lots of details and low level information into it and having wide dynamic range may be advantageous .
But that's not the point and if you clean up your discs and use mint condition cartdrige you should not have so much ticks and not so loud.
By the way if you have some rarities on vinyl and want to hear what i'm talking about low level informations and such i could do some kind of 'restoration'/cleanup for a track. This is really impressive what can be recovered using availlable digital tools.
But that's not the point and if you clean up your discs and use mint condition cartdrige you should not have so much ticks and not so loud.
By the way if you have some rarities on vinyl and want to hear what i'm talking about low level informations and such i could do some kind of 'restoration'/cleanup for a track. This is really impressive what can be recovered using availlable digital tools.
Ripping for restoration and real time playing through miniDSP are completely different animals though. When ripping you can push closer to the edge and also have more than one go at it.
At the end of the you are limited to around 70dB SNR by the cartridge intrinsic noise. Fact of life.
At the end of the you are limited to around 70dB SNR by the cartridge intrinsic noise. Fact of life.
Well I decided to go for the miniDSP U-DAC8 and JRiver. Bought the JRiver license this morning since my trial is expiring and just ordered the DAC a minute ago. We'll see how long it takes to get here from Hong Kong.
Now I just have to figure out what to do about amplifiers. I've got an old Harman Kardon that does 65W X 2 but it's not rated below 8 ohms. And I've got a Crown pro amp with tons of power into loads down to 2 ohms. I'd like to find 6 channels of at least 100W rated down to at least 4 ohms.
-Chris
Now I just have to figure out what to do about amplifiers. I've got an old Harman Kardon that does 65W X 2 but it's not rated below 8 ohms. And I've got a Crown pro amp with tons of power into loads down to 2 ohms. I'd like to find 6 channels of at least 100W rated down to at least 4 ohms.
-Chris
I'm a big fan of DSP from a DIY point of view. You buy 1 unit and can keep experimenting with different cabinets, drivers, crossover slopes and points without spending another dime. (or reading about and looking for some special flavor of caps...) With a press of the button you can switch from LR2 to LR4, flat frequency respons or sloping down, anything you can imagine and hear (or not) the difference in an instant.
On the topic of digital vs analogue gain and FIR vs IIR filters:
With the Najda DSP (which I'm using for a 4-way system) you have digital and analogue gain (you can set the max output from 1V to 6V in 0.5V increments per channel). Very handy if you are using different amps (I bought a couple of 10 to 15 year old ~80W amps for 50€ each second hand) You can also do any IIR filter either directly (common types) or via importing the a0 a1 b1 etc coefficients for any possible theoretical IIR filter or linear combination of filters.
You can also use FIR filters, Najda has enough processing power for 1023 taps per channel for 8 channels. 1023 taps does limit you to no steeper then 12dB per octave at ~100Hz. Software capable of decimation is being worked on and should completely alleviate the low frequency limitations.
More info here:
http://www.diyaudio.com/forums/digital-line-level/215379-dsp-xover-project-part-2-a.html
WAF-Audio
On the topic of digital vs analogue gain and FIR vs IIR filters:
With the Najda DSP (which I'm using for a 4-way system) you have digital and analogue gain (you can set the max output from 1V to 6V in 0.5V increments per channel). Very handy if you are using different amps (I bought a couple of 10 to 15 year old ~80W amps for 50€ each second hand) You can also do any IIR filter either directly (common types) or via importing the a0 a1 b1 etc coefficients for any possible theoretical IIR filter or linear combination of filters.
You can also use FIR filters, Najda has enough processing power for 1023 taps per channel for 8 channels. 1023 taps does limit you to no steeper then 12dB per octave at ~100Hz. Software capable of decimation is being worked on and should completely alleviate the low frequency limitations.
More info here:
http://www.diyaudio.com/forums/digital-line-level/215379-dsp-xover-project-part-2-a.html
WAF-Audio
Last edited:
Digital Volume Control
These discussions are always interesting. A lot of people freak out about "losing bits" wit digital volume control. At least this discussion has been more sane. IMO, if you are worried about digital attenuation and losing bits, you shouldn't be using a digital crossover. Why? It attenuates. 😉
There is the theoretical and the practical. I worried a lot about digital volume control when I first set it up. "I'll lose resolution!" But there is no need to worry if you get it right. Gain structure is key. When set up properly, you'll not be needing more than 10dB of volume control under normal listening conditions. I was surprised at how little was needed.
As for signal degradation, there is less than with an analog volume control. In a player like JRiver, the volume control is done in 64 bit, then reduced to 32 or 24 for your DAC. That is a tremendously fine scale and way more dynamic range than you'd ever need. Just get your gain structure right.
I have done extensive listening tests (yes, my system is better than yours) and also a great deal of signal analysis looking at noise, THD IMD, harmonic structure, etc. The only thing digital attenuation does is reduce the S/N ratio. That's all, Basta. Any analog control does that and more. That's the practical side of it. Real life implementation.
From my extensive testing and listening I say: If your gain structure is properly set so that you are using no more than 12dB of attenuation, digital volume control is the way to go. It only attenuates, nothing else. Stop worrying about it.
These discussions are always interesting. A lot of people freak out about "losing bits" wit digital volume control. At least this discussion has been more sane. IMO, if you are worried about digital attenuation and losing bits, you shouldn't be using a digital crossover. Why? It attenuates. 😉
There is the theoretical and the practical. I worried a lot about digital volume control when I first set it up. "I'll lose resolution!" But there is no need to worry if you get it right. Gain structure is key. When set up properly, you'll not be needing more than 10dB of volume control under normal listening conditions. I was surprised at how little was needed.
As for signal degradation, there is less than with an analog volume control. In a player like JRiver, the volume control is done in 64 bit, then reduced to 32 or 24 for your DAC. That is a tremendously fine scale and way more dynamic range than you'd ever need. Just get your gain structure right.
I have done extensive listening tests (yes, my system is better than yours) and also a great deal of signal analysis looking at noise, THD IMD, harmonic structure, etc. The only thing digital attenuation does is reduce the S/N ratio. That's all, Basta. Any analog control does that and more. That's the practical side of it. Real life implementation.
From my extensive testing and listening I say: If your gain structure is properly set so that you are using no more than 12dB of attenuation, digital volume control is the way to go. It only attenuates, nothing else. Stop worrying about it.
Yeah but mine scored me my wife 😉
Very good points though, well put. And all things that need to be considered. And given how many people get it wrong, a good time to restate them!
Those lost bits are generally in the noise floor anyway. And how is it different that listening to a piece of music that in and of itself has sections that are 20 or 30dB quieter than other sections? Those quiet passages have a lot of lost bits by comparison, no?
I think that if someone else than you made this statement Pano, i could leave definitely this forum. 🙂
You made a nice summary of situation. But at least one point need to be adressed imho.
What is normal listening condition? I tend to agree with you about range you give in domestic environment, given you don't have close neighboor(or child) or a dedicated place for listening.
but for me this is not the case: i'm not lucky to have a dedicated room, have neighboor, a small child so night listening session can be problematic (even with pads and digital volume control) and add to this facts that i sometimes use my system for work i often need to vary levels in a much greater amount than 10db (-20 to -35 db to check of work done) and in this case i'm absolutely in needs too that the character of the sound of the dac don't change: if analog part of converter clip (because of a dynamic treatment done upstream digitally for example) i must have the ability to hear it even at lower level.
The only solution is a passive attenuation.
I don't freak out loosing resolution with my digital level control (i couldn't perform any fade in/out otherwise, nor gain riding) or my dacs (they are high ends pro units and once clocked really up par to world class range), i freak out that once the level is attenuated digitally the general character of the sound change (for the better or the worst). For me this is part of the high fidelity needed to work. And because this is my home system too i made the choice to use this solution at home.
And for convenience i'll upgrade to something using constant input z switched R2R ladder in the futur giving more flexibility/usability.
I understand others won't care about that especially for home use (or think that the trade off of using some resistors to attenuate is a nono), but for me this is an issue (that compromise the level of quality of the whole chain) and the way i've choosen to adress it. But i'm a particular case i admit.
Last edited:
- Status
- Not open for further replies.