Bits-are-bits-brigade and why they are wrong

[maxflinn]

I'll check back later. Any chance somebody could resize that troll image.

Done.



Just saying guys - this thread has been very civilised so far. Lets keep it that way. We've had a few posts on either side of the argument that indeed have been worth reading. If this descends into making the same argument again and again it'll be closed and locked. I've seen so many threads start like this and then go bad - its pretty much inevitable.

Rude posts, name calling, circular arguments, trolling etc will mean a ban. No post so far has crossed this line IMHO, but some are not far off. Please think before you post and please don't make me ban a member here.

Fran


Fran

Cheers.

[jkeny]

I know this post is off topic but it follows on from my last post & relates to some stuff I was mentioning to people at tonight's bits-are-bits get-together (thanks Tony). Mods, if you wish you can move these posts to a separate thread.

BTW, to summarise JNeutron's position on speaker cables: http://www.audiocircle.com/index.php?to ... msg1176464
The optimal cable is one with a characteristic impedance at 1kHz (?) that is equal to the nominal impedance of your speakers. That will deliver maximum energy of the signal & no reflections but more importantly there will be no difference in delivery time between the different frequencies. This is the well known transmission line model of RF engineers but applied to audio frequencies rather than RF as is usual.

So what if the different frequencies have different time delays? Well we localise the location of an image based on time differences (inter-aural-time-difference - ITD) & volume differences (inter-aural-intensity-difference? - IID). These can be very small - I've seen mention 5usec down to 1.5uSec as the smallest time delay between channels that is discernible & .05 dB volume difference. So anything that affects the timing of the signal at this level or greater can have an audible effect on the imaging.

Key points:

We lateralize via ITD and IID. ITD is demonstrable to the level of 1.5 uSec difference.

Any effect which cannot alter the signals at that level or above can be discounted. 20 nSec for example (a ten foot cable with a matched load at the end), totally below the radar and not worth a second look. 20 uSec exceeds threshold so warrants further investigation.

Extreme mismatches between a transmission line and the load at the end does not achieve final value quickly, it can require significant settling time, up to hundreds and hundreds of the transit time of the cable..When this settling time is longer than 1.5 usec, it warrants further investigation. This is actually the case with zip and 4 or 8 ohm loads.

Speaker loads vary tremendously. The higher it's impedance, the faster the system will settle. If the speaker actually makes it to about 100 ohms at some frequency, that frequency content will settle on the line at about 20 nSec. Frequencies where the impedance is very low will have that frequency content delayed based on the mismatch. If the mismatch causes delays above ITD threshold, that warrants further investigation.

Since we localize based on ITD and IID, and due to the fact that some of the content will be delayed differently because of the wildy varying speaker impedance vs fequency, we have to worry about image disruption.

In a mono world, these frequency based shifts due to symmetry, will remain centerstage.

In a stereo world, these frequency based shifts can move off axis images away from their intended location with respect to a central mono image, as the mono image retains angular stability. Key to this is the fact that you are trying to locate off axis images with respect to on axis images...humans are very good at differential location discernment.. Greisinger mentioned 3 degrees of horizontal angular discernment in his talk.

This discussion is essentially focussed on the 500 to 3 or 5Khz range of course.

Whether this is audible will depend on a lot of variables - not least multi-driver Vs single driver speaker, crossover & driver time alignments, nominal impedance of drivers, etc.

So a practical example based on my own speakers which are Jordan JX92S single drive DIY speakers - a stated nominal impedance of 8 ohms but a measured impedance above 500Hz of 5ohm. I use cat-5e for speaker cable which is 100ohm impedance per twisted pair. To bring this down to 5ohms I need to parallel 5 cables with all stripes tied together & all solids tied together. Each twisted pair is 100 ohms. Assuming you connect all the stripes together as one leg , and all the solids together as another leg, the impedance will be 100/n, with n being the number of twisted pairs. One cat5e will have n=4, Z (impedance) = 25ohm. 8 runs of cat5e is 32 pairs, so z = 100/32 or about 3ohms. 6 runs is 24 pairs, 100/24 =4ohms. 5 runs is 20 pairs, 100/20 = 5ohms. 4 runs is 16 pair, 100/16 =6.25ohms.

I'll be trying this soon. A way of testing if this has an effect is to use a single cat-5e wire on one speaker & the paralleled wire on the other speaker - send a mono signal to them & listen to check for a shift in what should be a central image

[jkeny]

Actually here's the test explained in better detail by JN - this is for zip wires

1. make 8 identical lengths of #16 zip.
2. Parallel 4 on the right speaker, 4 on the left. Tie them all to one amp channel to force actual mono. You now have two cables of characteristic impedance of about 25 ohms as long as you do not tie them together.
3. Listen to your system for soundstage.. verify that every sound you hear is coming from the exact same central point in space while in the sweetspot.
4. Once satisfied that both speakers are producing the exact same thing (speaker symmetry), take the 4 zips of one channel, split them apart at the insulation and bundle all the positives together and all the negatives together, space all positives at least 6 inches from the negatives. You have just maxed the cable impedance while keeping the resistance the exact same. (edit: I mention 4 per channel, but you could use 8 if you want, get the z closer to the load..)

5. Listen to the soundstage. If you detect that some of the program content has shifted off axis with respect to any other content, then the change you have made (you have modified the cable's L and C) has made an audible difference.

This test is VERY sensitive to cable changes, as it uses your hearing to discern relative positioning of image content. It is a more sensitive test than trying different cables with a stereo program. If you do not hear any difference in imaging as a result of this test, no cables in the L/C range tried will make any difference to your system.

[tony]

One of these days I will read those tests!. Don't think I want to go on with this thread the threat of the chop has stunted the enthusiasm!
From the session last night with 2 software players and TT the group could identify differences that neither of the system owners realised were configured differently. This lead to different sound levels and a two different presentations of the same track. On realising that we had both configured Jplay differently we adjusted the settings to match and the sound from both set ups were more or less identical .This would support the bits are bits brigade. But when we introduced a set of supertweeters which some people are not overly enamoured with differences in the hi frequencies could be clearly identified again.

The conclusion I would make is that depending on your components/software/room/cables you are going to hear differences. But not all the time.On some occasions those that believe all CA sounds the same would be able to jump in and say told you so. But if they stayed to hear other stuff they would see that the opinion doesn't hold true all the time.

BTW thanks to Pearse for bringing a really nice Turntable and some of the lads from bringing along some vinyl.The crackling when playing would still be an issue for me but there is no doubting the magic of a really nice TT set up. CA seems to say the least at little cold in comparison.