Bounced Mix not same as playback

Ok, I know this has been discussed before, but I could not find a conclusive answer to this.
After finishing mixing a project yesterday, for the first time in a while I had the same sensation I had experienced a few times in the past… the bounce mix sounds like sh*t vs the itb mix playback
So this time I have decided to check it for myself and this is what I did:
1.Bounced the mix 24 bit
2. Re-route the palyback into my digital input in my Lynx Hilo
3. Record the file at 24 bit
4. Imported bot files in Samplitude Prox3
5. Critical listening to both files. Blind tested and always chose the captured vs the bounced one
6. Did a null test
7. There were massive differences in the top end and mid range. (diff peaking as much as -24dbfs)

The bounced mix sounded less defined and with a more blurred stereo panorama.

I know other people in the past expressed the same concern. I just wanted to know if there is something I am missing here. I understand some plugs use random algos to emulate analogue behaviour, but this is not it. Anyone else experienced this?

Cubase 10.5 Pro on W10, lynx Hilo and Dutch&Dutch monitoring.

Thank you

forgot to mention, I have disabled the control room alltogether before doing the test.

Is there an email where you can contact the tech dpt at Steinberg? I couldn’t find one. Cheers

UPDATE, I have done another test without any plug in on the 2buss. And when re-did the test, it is a perfect null back in Samplitude.
As I very rarely use any type of processing on the 2buss, this could explain why I normlly have no issue with my bounces…
There is something that doesn’t add up, but I don’t think it is that relevant
I have found old posts mentioning similar findings…where people swore that the bounced mixes where truer to the original if not processed on the 2buss, or processed routing all channels to a submix and not the main 2buss… but they were pretty much all inconclusive.