Genelec Community Forum has been archived

Thank you for all the years of Community discussions and activity!

The time has come for us to retire the community forums, but we will keep everything available for reading. There is a lot of valuable content written over the years, and you'll be able to access all of that. However, no new posts can be written, or old posts modified.

If you have questions, we recommend you use the Support function on the bottom right corner to contact our Customer Support directly.

Alternatively, if you wish to engage in a community discussion with other people, there are many active forums available. There is also an active, fan managed GENELEC COMMUNITY in Facebook, and many Genelec employees are active in the discussions there.

We are sorry for any possible inconvenience this may cause you, but we hope to hear from you through the other channels mentioned above.

Genelec Support

Message Boards

AES/EBU connection and levels

Steve G, modified 1 Year ago.

AES/EBU connection and levels

Youngling Posts: 1 Join Date: 7/21/21 Recent Posts

Hi,

I have recently set up my home music studio with a pair of 8340's via AES/EBU and used the GLM to calibrate. But, I was wondering the best way to make sure the quantisation and bit resolution is kept in tact. If the audio leaves my RME sound card via AES at 0dbfs and I use the GLM software slider (or the 9310A volume knob) to manage the overall speaker levels, will I be lowering the bits. Or, should I trim the volume from the rotary volume level on the back of the monitors to get it in the ballpark first?

Regards, Steve 

lds, modified 1 Year ago.

RE: AES/EBU connection and levels

Youngling Posts: 2 Join Date: 11/3/19 Recent Posts

Quantisation resolution? Quantisation is always constant through digital audio. The only way to lower the bits is by truncation or dither - taking a 24 bit signal and making it 16 bit. I would just set the levels on the speakers to unity, then use GLM. Quantisation noise is always constant. It is always expressed as half the value of the least significant bit - this is because of the rounding error. Below 0.5 gets rounded down. Above 0.5 gets rounded up. 

Bit depth is nothing more than a base two counting system. It translates perfectly to decimal. Instead of sound, consider it like counting chocolate bars using one of those rotary counting 'clickers' that security guards might use to count the number of people entering a nightclub. The system can't represent half a chocolate bar. But, that doesn't change the resolution of the system based on the number of chocolate bars we do count. 0.5 chocolate bars. 5.5 chocolate bars. 100.5 chocolate bars. They all experience the same amount of rounding error... the same amount of quantisation. 

What is effected is signal to noise ratio, but that equally applies to the analogue realm too.