DIY Audio DSP X/O DIY project

[speedysteve]

Only correction is for the RCAs at 1100Hz....

Cool - see other questions I've added above. Plus is that measured on axis / in listening chair?

 

[i_should_coco]

Lab subs - 100Hz, LR2 LP.

Vitavox AK151/Oracle - 100Hz LR2 HP, 400Hz LR4 LP.

RCA Mi-1428/Conical - 400Hz LR4 HP, 1500Hz LR4 LP.

Vitavox S2/Azura 550 - 1500Hz LR4 HP.

RAAL Lazy - 10kHz passive, 1st order.

A bit of delay on the S2, but need to play more.

 

[speedysteve]

The proof of the pudding is in the measuring

This morning I tried Najda on DSP X/O into an S2 (Red plot). Najda still in place but no DSP X/O, instead a 3uF cap on S2 (Green plot).

[/url]

So the cap is doing stuff DSP X/O is not / cannot. This is really discernible and my lovely vocals and upper dynamics are back.

Perhaps the DSP is true and the cap is not - don't care!

This is as I first heard Najda on the upper mids on 96KHz pass through. Tweeters back to higher passive X/O.

Now I have Najda bass, Najda X/O on mids (measures the same) and Najda (passthough) + cap X/O on upper mids.

I've still got to time align the whole lot yet.

 

[i_should_coco]

Of could try a different make of cap in case the parasitics are doing something. Or it could be voice coil inductance having an effect?

 

[ChrisPa]

That looks like a +5db shelf from 5k upwards

Can't remember, is there a shelf function built into najda?

 

[speedysteve]

That looks like a +5db shelf from 5k upwardsCan't remember, is there a shelf function built into najda?

Don't know. Can't see why there would be for the application Najda is intended - Not PA but hifi.

- - - Updated - - -

Of could try a different make of cap in case the parasitics are doing something. Or it could be voice coil inductance having an effect?

Yes could be inductance...

The DIYaudio thread advice is to cut in DSP never add. Anytime I have done both to comp drivers the results have been dire - worth trying...

 

[ChrisPa]

That looks like a +5db shelf from 5k upwardsCan't remember, is there a shelf function built into najda?

Don't know. Can't see why there would be for the application Najda is intended - Not PA but hifi.

DSP or any EQ is simply about EQing the system - there's no difference between PA or hifi with regards this requirement. It's all about whether the corresponding biquads have been preconfigured within the signal processing.

Anyway, according to the najda software page http://www.waf-audio.com/products.php?pos=1⊂=4〈=en it does have shelving (I thought I'd remembered that)

If you compare the red plot to the green plot in your graph, then by eye the green plot seems to +5dB from about 5k upwards in comparison to the red plot

 

[speedysteve]

DSP or any EQ is simply about EQing the system - there's no difference between PA or hifi with regards this requirement. It's all about whether the corresponding biquads have been preconfigured within the signal processing.Anyway, according to the najda software page http://www.waf-audio.com/products.php?pos=1⊂=4〈=en it does have shelving (I thought I'd remembered that)

If you compare the red plot to the green plot in your graph, then by eye the green plot seems to +5dB from about 5k upwards in comparison to the red plot

My comment was referring to what I assume 'they' do to PA X/O's etc to prevent tweeter damage at very high vols. PA boys don't aim above 15K or so as it sounds rubbish at those vols anyway.

I assumed shelving meant for the bass and treble mode - with that off should be ok. If this is the cause of my dB robbery with it disabled - it rather sucks!

I have been playing and testing and still prefer the caps/passives on the compression drivers to anything coming out of Najda X/O wise.

So far best combo is Najda X/O on Tapped and Mid Bass with sparing use of cutting only correction to tame room modes.

Mids, Upper mids and Tweets on passives fed though Najda with the X/O's off - no correction. X/O's for Tapped and Mid bass tuned to give best measurable response.

Sounds totally acceptable. Playing SPU/FR64s/SP-10 into it now - nice. Bass was good before - it is rock solid and as I want/like it now.

Next round of measuring and playing will be time alignment for all channels in Najda to see how that affects things - Could be positive from time aligning POV or mush things up from DSP having to work harder POV? Lets see.

 

[ChrisPa]

Time aligning shouldn't really give the dsp any work, it's more a matter of adequate memory.

It should just poke calculated data away in one location and read it out from another (delayed)

 

[speedysteve]

Time aligning shouldn't really give the dsp any work, it's more a matter of adequate memory.It should just poke calculated data away in one location and read it out from another (delayed)

Yeah, does NOT push the processor load up. Not much in it but you can hear the difference, and I prefer it. The effect is on coherence, hard to describe, calmer also. Same as I have heard with passive time alignment. Leaving it alone now for a long while... famous last words.

 

[speedysteve]

Wife wanted some music streamed via Spotify - sounded pretty dire then ha! I EQ'd it to sound much more punchy and dynamic.

Not so bad now. Still way off full fat or vinyl but acceptable.

 

[speedysteve]

Re-awaken this one...

I am play much more computer based audio having ripped all of me CD's and a few records that you can't easily buy or get on CD at all...

So thoughts are turning to I2S.

I want something cheap that sounds good and is mounted in the Najda box (important to keep the wires to the I2S input short I understand.

This is tempting

http://www.ebay.co.uk/itm/New-USB-TO-I2C-IIC-UART-TTL-Adapter-Converter-/191248965233?pt=UK_BOI_Electrical_Components_Supplies_ET&hash=item2c87539e71

Nick of (WAF) talks of Tenor or Amanero based and has tested against them.

http://amanero.com/

http://www.ebay.co.uk/itm/LJM-Audio-TE7022-USB-To-SPDIF-I2S-processor-board-TE7022-Support-24-96KHZ-/111097047820?pt=UK_Sound_Vision_Amplifier_Parts&hash=item19dde66b0c

Or are there others with better bang for buck?

 

[speedysteve]

DIYAudio Najda thread to the rescue - Cheque book almost out...

 

[vinylnvalves]

That reminds me i need to install the Amanero USB I2S card into my Nadja only had the card 6 months! (maybe i should rename myself as slowsteve!)- Maybe a job for a rainy afternoon, it is the soldering the pins onto the board, disassembling the Nadja which has put me off.

 

[speedysteve]

Yes, SlowSteve

From memory it is the holes that are on the expansion port part of the board.

The I2S blurb in the manual makes it where but not so clear as to what pin / out to what pin of the the USB->I2S board / card. They all seem to have similar but sightly differing names for the pins. This is what has put me off doing it so far...

A chap on DIYaudio got all helpful and was going to email me schematic/pin naming for the board I was thinking of getting (WaveIO board) - But then he went all quiet

Sometimes, life gets in the way

That reminds me i need to install the Amanero USB I2S card into my Nadja only had the card 6 months! (maybe i should rename myself as slowsteve!)- Maybe a job for a rainy afternoon, it is the soldering the pins onto the board, disassembling the Nadja which has put me off.

 

[vinylnvalves]

For me one who owns one of the older red Nadja boards - its the removal and the soldering of the 2.5mm pins on the expansion which is the pain, if i had one of the latest black boards it would have been really easy. There aren't that many pins to connect up, even less if like me you have a dedicated 3.3v supply for the I2S.

Have you seen my post on the other forum - regarding issues with the Nadja and SUT's.

 

[speedysteve]

I don't mind fiddle soldering. I'm getting there on which pins to connect where and have a WaveIO on order following recommendation / help.

I've answered you on said other forum but might be wrong...

For me one who owns one of the older red Nadja boards - its the removal and the soldering of the 2.5mm pins on the expansion which is the pain, if i had one of the latest black boards it would have been really easy. There aren't that many pins to connect up, even less if like me you have a dedicated 3.3v supply for the I2S.Have you seen my post on the other forum - regarding issues with the Nadja and SUT's.

 

[speedysteve]

Got an email from the WaveIO maker. It is being manually assembled, will take 3 weeks (as per info on the website), and will be tested before shipping.

http://luckit.biz/

I now have connection diagrams from a chap in France who uses one (recommended it), so connecting up should be a doddle.