AAAAgghr only 16bit

[Crash]

I know your decision has been made but as an FYI, Version 5 is much cheaper than $300.00 these days as I recall. I am not sure that there would have been any literature that states a 16 bit fix since there was a patch available for V3 as you stated. In any case, best of luck on your march forward with whatever you go with and I look forward to more on this from you and the others.

[alfspanners]

Ok, I had been operating under the mistaken belief that I needed to upgrade either the cpu or memory if I had wanted to install version 5. When I read anyhorizon's post about just installing the v5 demo, I thought I should check again what the spec was - I think I must have misread it at the time -I certainly didn't have the optimum spec but my 166cpu and 32meg was just on the minimum so I downloaded it and installed it this morning. Took a while to test because I had to do several reboots, but I have enough info to form an opinion and, in since reinstalling version three I am able to show information for what actually changed in each step of the process.


Very important to note that the channel filters are OFF


firstly this is just V3 build 224 - no patch and with a bit of spoken word via an sm57

Digicheck version 3 build 224 left channel - channel 1 mic in spoken word

Version 3 pre patch digicheck spoken word.jpg (Array KiB) Viewed 1511 times

Bitter version 3 build 224 channel 1 mic in spoken word

Version 3 pre patch spoken word quiet.jpg (Array KiB) Viewed 1511 times

Clearly, according to bitter and digicheck this was only 16bit and when someone realised - mackie released the patch.


So now I installed the patch

Bitter and cubase stats Version 3 build 224 plus patch

Bitter and stats showing lower 8 bit anomaly.JPG (Array KiB) Viewed 1511 times

Digicheck mackie analogue trim comparisons

Digicheck mackie 8 bits switching.jpg (Array KiB) Viewed 1511 times

d8b line in 1hz sine

1hz sine mackie line in.jpg (Array KiB) Viewed 1511 times

Ok you've seen these three before. 8 new bits have now appeared to be at the bottom, though bitter is showing them to be odd and at low levels they flash on and off in unison in digicheck/ move in and out in unison in bitter and with the preamp trim at minimum - they disappear in digicheck.


cont

[anyhorizon]

Where do you get a 1 Hz sine wave from?

Peter

[alfspanners]

Now I installed V5

Version 5.1 preamp at minimum channel filter off - (nothing has been patched to the right channel)

version 5 point 1 preamp minimum.jpg (Array KiB) Viewed 1511 times

Version 5.1 preamp at min - Bitter

Version 5 point1 bitter analogue trim min.jpg (Array KiB) Viewed 1511 times

Version 5.1 digicheck - spoken word

version 5 point one spoken word.jpg (Array KiB) Viewed 1511 times

Version 5.1 spoken word Bitter

Version 5 point1 bitter spoken word.jpg (Array KiB) Viewed 1511 times

Version 5.1 -30db 1 hz sine fed to line in. 40db of gain added and zoomed to maximum vertical magnification

Version 5 point1 1hz sine recorded in d8b mic line.jpg (Array KiB) Viewed 1511 times

[alfspanners]

What I can say for certain is that our desk is only outputting our analogue conversions at 16bit, the extra 8 bits just dont add resolution.
The only other conclusion that I can draw is that V5.1 does not fix this bug on our desk.

Whether 5.1 delivers 24bit analogue conversions on other desks remains to be seen.

Clearly there was a change when the patch was added.

I propose that preamp at minimum on a proper 24bit system should look like this on digicheck

Digicheck on my RMe fireface with preamp to minimum - note also the lower noise floor

Digicheck Fireface trim comparisons.jpg (Array KiB) Viewed 1508 times

Just a reminder - the channel filters have to be switched off as they prevent the measurement of the lower bits.

I don't expect that we'll be able to agree on a test for all systems, but just to round it off

Mackie d8b version 5.1 mic in recorded as 24bit file measured in cubase stats.

Version 5 point1 audio file spoken word and yelling cubase stats.jpg (Array KiB) Viewed 1508 times

It was cubase statistics that started me off on this journey, whether it is accepted as a useful tool or not, It's results have concurred with the other tests that I have done.

[alfspanners]

One last point I've realised about bitscopes is that they are trying to show you an average of bit activity over a certain time period, because there are 44100 24bit words a second and no screen (mine is 60hz I think) or eye could perceive all of that information per second - so it gets averaged and different bitscopes will do this in different ways. Clearly RME digicheck is highly regarded, but I think bitter gives me more information by attempting to show how many times each bit is used in these intervals (the width of the line).

[anyhorizon]

Again, where are you generating a 1 Hz signal from? Truly! ... and what analog circuitry on this planet could accurately reproduce it in order that it could be converted to digital?

Peter

[alfspanners]

a 1 hz sine is easy to reproduce - it's just a slowly rising and falling voltage in the end. Cubase test tone creates it - I send it to the d8b digitally on a channel at -30db, send that channel out on an aux and patch it with a cable to a line in. I then adjust the line in so the signal peaks at -30db and record it. This is not to be confused with putting a 1hz signal to a speaker which is actually thereabouts a dc voltage - you can't hear it but the speaker moves in and out at 1 or 2 hz a DC voltage can fry the coil on a speaker , but they do exist. Why do we have to have high pass filters?

I am aware that the d8b aux outputs are only 16bit, but they are passed through a DAC which produces a smooth waveform (when I say smooth - well it's actually loaded with noise at the very low levels we're measuring). I replicated the effect with my fireface in an earlier post - converting the fireface file to 16bit then out via a cable and back in at 24bit resolution - just to be fair to the d8b - there was no comparison -the d8b clearly had larger stepping

[alfspanners]

In the end folks, there's no way I can actually make you believe me - you have to go and do the tests yourself if you wish to. Or if you think you can do them better or more reliably then why not propose alternative tests. I did what everyone asked, I downloaded digicheck and tested with that. As I said - the only things I can be sure of now is that the analogue ins on my desk are coming out as 16bit resolution audio and that version 5.1 doesn't fix it -it could just be that my desk is knacked - I don't know

[anyhorizon]

a 1 hz sine is easy to reproduce - it's just a slowly rising and falling voltage in the end. Cubase test tone creates it - I send it to the d8b digitally on a channel at -30db, send that channel out on an aux and patch it with a cable to a line in. I then adjust the line in so the signal peaks at -30db and record it. This is not to be confused with putting a 1hz signal to a speaker which is actually thereabouts a dc voltage - you can't hear it but the speaker moves in and out at 1 or 2 hz a DC voltage can fry the coil on a speaker , but they do exist. Why do we have to have high pass filters?

I am aware that the d8b aux outputs are only 16bit, but they are passed through a DAC which produces a smooth waveform (when I say smooth - well it's actually loaded with noise at the very low levels we're measuring). I replicated the effect with my fireface in an earlier post - converting the fireface file to 16bit then out via a cable and back in at 24bit resolution - just to be fair to the d8b - there was no comparison -the d8b clearly had larger stepping

Please don't lecture me on what a sine wave is. I put it to you that you haven't got a clue of what you're talking about or are trying to shove down our throats. You have just demonstrated your lack of knowledge. One thing is for sure, I'm glad you don't work for my organization. Get rid of your d8b and don't come back here. Please. I hear there's a new Behringer desk that would suit you down to the ground. Your trust in Cubase as a measuring tool is misguided at best. I can't believe we have come this far after having read your above quoted post.

Peter